Introduction to WiMAX Presentation

Introduction to WiMAX

Technology

Khalid Sheikh

January 2009

Introduction to WiMAX Technology Page 2

Presentation Overview

• Introduction

• Overview of other wireless technologies

• Radio Architectures

• OFDM Basic

• OFDMA Basic

• WiMAX Architecture

– PHY, MAC, Control & Management

• System Performance

• WiMAX Validation

• WiMAX Interoperability

• WiMAX Network

• Summary

• Q & A


Wireless Systems

• Wireless communication has been around for over 100 years

– Pioneered by Thomas Edison who did not take serious interest in this technology and sold his patent to Marconi for a single song

– Most non-broadcast typed applications geared to P2P & PMP

– Generally standalone operation based on proprietary architecture

• Incompatible over the air interface

• Many lacks interoperability with other equipment

• Most failed to provide high data rate and mobility

• Some commonly used recent LAN applications include

– Infra-Red, IR

– Bluetooth (WPAN)

– Mobile Phone

– WiFi (LAN)

– WiMAX (MAN)


A Case, why we need it!

• Wide use of internet resulting in an increased demand for convenient internet access and high speed data access.

• Increase demanded by new applications such as streaming video, on-line gaming, on-demand movie distribution, VoIP, video teleconferencing, telemedicine, serveillance & monitoring, etc.

• Fixed wireless offers several advantages over traditional wired solutions. These advantages include lower entry and deployment cost; faster and easier deployment and revenue realization; ability to build out the network as needed; lower operational costs for network maintenance, management, and operation; and independence from the incumbent carriers


WPAN, Bluetooth

• Wireless personal area networks based on IEEE802.15.1

• FHSS operation using TDD

• Data rate

– Sync., connection oriented, 64 kbps

– Async., 433.9 kbps symmetric

– Async., 723.2 / 57.6 kbps asymmetric, 1 Mbps aggregate bit rate

– Ver2 to increase up to 3 Mbps

• Three PO classes

– Class 1, 1 to 100 mW, for 100 m range

– Class 2, 0.25 to 2.5 mW, for 10 m range

– Class 3, up to 1 mW, for 1 m range

• Shares data among up to 8 Bluetooth enabled devices

• One of 79 channels in 2.402-2.480 GHz ISM-band

• 100 bytes long packet length

• BPSK for Ver1, DQPSK & 8-DPSK for Ver2

• GFSK with mod index of h = 0.28

• Limited QoS

• Guarantees, ARQ/FEC

• Connection setup time

– Depends on power mode

– 2.56s max, 0.64s average

• Provides high level protocol support

• Security

– Challenge/response (SAFER+) hopping sequence


Bluetooth Protocol Stack

• Middleware Layer to support:

– L2CAP, logical link control and adaptation protocol

– SDP, service discovery protocol

Required Not part of Bluetooth standardOptional

Baseband

Radio

Link Manager

Audio

HCI, Host Control Interface

L2CAP

RF Comm

OBEX

vCard

BNEP

Networking

Apps

AT

Phone AppsManagement AppsAudio

Apps

TCSSDPControl

Application

Layer

Transport

Layer

Middleware

Layer


WiFi, Wireless Fidelity

• Low cost & very popular due to standardized architecture & mutual IOP

• Short distance coverage, ≤ 100 m in open area

• No QoS offered, best effort only

• Operation in License Exempt Band

• IEEE802.11a/b/g/n, Standard first adopted in 1997

• Delivers services previously found in wired networks

• Relatively high throughput

• Highly reliable against interference by applying fragmentation technique

• Continuous connection

– Every station reacts to every frame it receives

– Requires participation of all stations

• Low power operation to prolong battery life

• Authentication services

• Architecture includes IR, FHSS, DSSS, OFDM


IR, Infra-Red

• Uses a near visible light as a transmission media

• Typically Line of Sight operation or reflected from object

• Restricted to indoor applications

• Can not pass through walls

• Data rate 1-2 Mbps

– Uses 16-PPM modulation for 1 Mbps

– Uses 4-PPM for 2 Mbps

• PPM is a modulation technique that keeps the amplitude and pulse width

constant and varies the position of the pulse in time. Each position

represents a different symbol in time.

• Operates at base band

• Inexpensive system based on IEEE802.15.1

• Short range, ≤ 1m

IR PMD XCVR

PSDU Symbol

Mapping

16-PPM

4-PPM

Modulator

PSDU Symbol

De-mapping

16-PPM

4-PPM

Demodulator

LED

Driver

LED

Detector


FHSS, Frequency Hop Spread Spectrum

• Data rate 1-2 Mbps, TDD (2-level GFSK for 1 Mbps, 4-level GFSK for 2 Mbps)

• A set of hop sequence defined in 802.11

– Channels are evenly spaced across a span of 83.5 MHz

– 78 (75 min) frequencies, each occupying 1 MHz BW channel (total of 79 Ch.)

– Hopping at least every 400 ms, then resync before resuming data transmission

– Predetermined pseudo random pattern

• Used in licensed exempt 2.4 GHz band (2.4 to 2.4835 GHz)

• Hopping sequence: 3 set of 26 channels (min hop distance of 6 Ch.)

• Interference tolerant

• Echo resistance

• Simpler to install than the DSSS

• Less expensive than the DSSS

• More vendors/selections than the DSSS (not true anymore)

• No spreading gain, No SNR improvement


FHSS, Frequency Hop Spread Spectrum

DSSS Packet

Fre

qu

en

cy S

lots

Time0

54321 76

20

40

60

80

1 MHz

22 MHzDSSS PacketDSSS Packet

FHSS Packets

Tim

e

Freq

(GHz)

1

2.442.432.422.412.40 2.45

2

3

4

5

Hopping Pattern: C A B E D

A

E

D

C

B


DSSS, Discrete Sequence Spread Spectrum

• Regulated per IEEE802.11b

• 1 or 2 Mbps using 11-bit Barker spreading code

– Spreading yields processing gain at receiver

– Requires channel linearity over 11 MHz

• Operates in license exempt 2.4 GHz ISM band (2.4 to 2.4835 GHz)

– 3 non-interfering 25 MHz apart channels

– Extremely crowded band


DSSS, (1)

• Fairly effective at low data rate

• Bandwidth requirement becomes too large at higher data rate

– Not practical to implement due to increased cost, power, size & technical

difficulties

• Processing gain = 10*Log (chip rate / bit rate) = 10.4 dB for 11 chips

• Feasible to achieve negative SNR at lower modulation in equivalent BW

Original Data

Barker Sequence

Spread Data

1 bit

1 chip

SNR

Noisefloor

Interfere

Narrowband Signal


DSSS, (2)

• Frequency, BW & PO are regulated worldwide

– PO of 100 mW nominal

• Data rate up to 11 Mbps

– operates at 11 Mbps and falls back to 1/2 Mbps as the legacy 802.11

– DBPSK & DQPSK for 1&2 Mbps, CCK for 5.5 & 11 Mbps with

enhanced 802.11b

• Interference tolerant

• Upgradeable to higher speed while operating in 2.4 GHz


DSSS Spectrum

• Regulated spectral mask

– Signal occupies in about 20 MHz BW regardless of data rate (1, 2, 5.5, or

11 Mbps)

– Spectral shape of the channel represents sin(x)/x function

– Spectral products to be filtered to -30 dBr from central frequency and all other

products to be filtered to -50 dBr

2.400 GHz 2.483 GHz2.412 GHz

(channel 1)

2.462 GHz

(channel 11)

2.437 GHz

(channel 6)

25 MHz25 MHz

Minimum

Channel spacing between

center frequency

0 dBr

-30 dBr

-50 dBr

fc + 11MHzfc + 22MHz

fcfc - 22MHz

fc - 11MHz

Transmit Channel Shape


DSSS & FHSS Implementations


DSSS vs. FHSS, Summary

• DSSS

– Short latency time

– Constant proc gain = a better SNR

– Quick lock-in as radios synchronize

– No dwell time

– No re-sync with other radio

necessary

– Short indoor range

– Long outdoor range (40 km in LoS)

– Greater overall data throughput

– Noise immunity (high)

– Multipath immunity (good)

• FHSS

– Long latency time

– No processing gain

– Slow lock-in, must search a channel

– 400 ms dwell time

– Must re-sync with other radio after

every hop

– Short indoor range

– Short outdoor range (10 km in LoS)

– Lower overall data throughput

– Noise immunity (low)

– Multipath immunity (none)


802.11a, (1)

• LAN standard revised and released in 1999

• Multicarrier OFDM system

– Similar to ETSI Hiperlan-2, main difference resides in the convolution encoding

method

• Operates in 5 GHz UNII license-exempt band

– Three 100 MHz bands in ANSI operation

• PO restricted per operating band

– 40 mW in 5.15-5.25 MHz, 200 mW in 5.25-5.35 MHz & 800 mW in 5.725-

5.825 MHz

– Antenna gain restricted to 6 dBi max

• The centers of the outmost channels shall be at a distance 30 MHz from

band‟s edges for the lower and middle bands, 20 MHz for the upper bands

• Channel frequency numbers are defined by 802.11a

• Minimum sensitivity -82 to -65 dBm depending on the chosen data rate & Mod

• PER rate to be less than 10% at a physical sub-layer service data units of length

1000 bytes


802.11a, (2)

• BPSK to 64 QAM modulation

• Different FEC rates, double encoding (inner & outer) & block interleaving

• Raw data rate up to 54 Mbps

• Multi-path fade tolerant

• Intended for short range, ≤ 100 m

• Best effort service (no QoS)

• Sync using a fixed training sequence lasting less than 16 us

• Data rate, FEC & Mod throttle up / down based on path conditions

• Encrypted security with enable/disable option


802.11a, Receiver Performance Requirements

• Required minimum threshold & tolerable interference level

• NF dependency per frequency band

• FEC & Mod dependency

Data rate

(Mbps)

Sensitivity

(dBm), 1e- 6

Adjacent Channel

Rejection (dB), 1dB

Alternate Adjacent

Channel Rej. (dB), 1dB

6 -82 16 32

9 -81 15 31

12 -79 13 29

18 -77 11 27

24 -74 8 24

36 -70 4 20

48 -66 0 16

54 -65 -1 15


802.11a, Example Symbol

• OFDM using 64-points IFFT/FFT– Data subcarriers: 48

– Pilot subcarriers: 4

– DC subcarrier: 1

– Null subcarriers: (6+5)

• 20 MHz BW– Subcarrier spacing: 20M/64 = 312.5 kHz

– Per channel spacing: 3.2 us

– Guard band: 0.8 us

– Symbol spacing: 3.2 + 0.8 = 4 us

– Symbol rate: 250 kHz

– Bit rate at 64QAM-3/4: Symbol * FEC * Active-Subcarriers * Symbol-rate = 6 * (3/4) * 48 * 250k = 54 Mbps


802.11a, Example Frame

• Frame format: PSDU PHY sub-layer service data unit, MPDU MAC protocol data

unit, PLCP physical layer convergence procedure, PPDU preamble and header to

create the PLCP protocol data unit.

• Preamble (header) is always BPSK-1/2 for ruggedness

PLCP Header

Rate

4 bits

Reserved

1 bit

Length

12 bits

Parity

1 bit

Tail

6 bits

Service

16 bitsPSDU

Tail

6 bits

Pad

bits

PPDU Frame Format

Signal

One OFDM symbol

Data

Variable number of OFDM symb

PLCP Preamble

12 symbols

Coded OFDM

(rate is indicated in signal)


802.11g

• Covers both 802.11a and 802.11b standards

• Same MAC layers for all variants (802.11b, a & g)

• Adds 802.11a equivalent operation in 2.4 GHz band

–Higher data rate (54 Mbps) than the 802.11b

–OFDM based transmission


802.11n

• Uses either 20 or 40 MHz channel

• Transmit diversity with multiple data streams

• Increased use of MIMO operation (SM, STBC, Tx Beam

Forming)

• Increased throughput to 100 Mbps per stream (600 Mbps

with all options)

• Enhanced QoS & FEC

• Selectable CP delays (400n or 800n)

• Backward compatibility for 802.11a/b/g

• Increased data sub-carriers (48 to 52 in 20 MHz, double

for 40 MHz)


LTE – High Level Requirements

• Standardization effort for LTE was launched in Nov 2004

– Expected to complete in Oct,09

• Peak data rate

– 100 Mbps in 20 M BW in the DL (with 2x2 MIMO)

– 50 Mbps in 20 M BW in the UL (without MIMO)

• Control plane latency

– Transition time from idle to active state ≤ 100ms

– Transition time from dormant to active state ≤ 50ms

• User plane latency

– Measured from UE to edge of RAN (one way)

– Shall be less than 5 ms for single user for small IP packet

• Control Plane Capacity

– At least 200 active voice calls / cell / 5 MHz

• Mobility

– Optimum performance at low speeds – from 0 to 15 km/hr

– High performance at higher speeds – from 15 to 120 km/hr

• Spectrum flexibility

– 1.25, 1.6, 2.5, 5, 10,15 and 20 MHz

• All IP network

– All services in the packet switched domain

– No circuit switched domains


WLAN Standards, Perspective


Super Heterodyne Architecture


Zero-IF Architecture


Radio Architectures

• Direct Conversion

– Advantages

• No off-chip IF filter

• Single synthesizer

• Cheap

• Low power consumption

• No image signal

– Disadvantages

• LO leakage

• LO pulling range

• High freq low PN requirements

• I/Q mismatch

• Quadrature LORF

• DC Offset

• Super-heterodyne

– Advantages

• Low LO leakage

• Wide LO pulling range

• No quadrature LO

• Design flexibility

• Superior I & Q matching at IF

• High performance

– Disadvantages

• Off-chip IF filter

• Two synthesizers

• Low integration

• High IF-RF separation to avoid

Image signal


SDR, Software Defined Radio

• Advantages (with direct

conversion)

– S/W configurable in the field

for specific conditions

– S/W upgrade in the field

– Reduced parts count

– Reduced die/board space

– Lower power consumption

– Lower over all cost

– Simpler assembly

– Single integrated synthesizer

– Almost spurious free

• Reduced RF filtering

• Disadvantages

– DC offset issue due to RF to LO isolation

or 2nd order non-linearity

• Appropriate mixer design ≥ high IP2

• 2nd order nonlinearity in the LNA

generates low frequency beat

– Static (slow varying) DC offset due to LO

to RF isolation

• DC offset cancellation at each received

packet

– LO pulling

• VCO at multiple or sub-multiple of LO

• Fast synthesizer response

– LO emission

• LO leakage to mixer input or antenna


Disadvantages of Direct Conversion,

(conti’d)

• Gain control to move to RF section

• Higher cost to pay for lower phase noise higher frequency LO

• No possibility to reduce in-band noise & spurious generated

by signal chain

• Suffers LO and side band leakage when DAC Synth a low IF

• Flicker noise

– Low frequency noise in all active devices

– WLAN has a large modulation bandwidth, no energy at low frequency

• Phase and gain mismatch the IQ-symbol

– Difficult to correct imperfect quadrature error


LAN, 3G & others vs. MAN, (1)

• MAN– BS connected to public networks

– BS serves subscriber stations

• BS and stationary/mobile SS

• SS typically serves a building (business or residence) & mobile

• Provide SS with first mile access to public networks

– Multiple services (voice, data & multimedia) with different QoS priority simultaneously

– Robust security

– Many more users

– Much higher data rate

– Much longer distance

– Selectable bandwidth (1.25- 20 MHz) & data rate

– Adaptive modulation & coding

– Advanced antenna techniques

– TDD & FDD, symmetric & asymmetric rate

– Spectrally efficient

– Lower cost than 3G solution

– Link layer retransmission

– OFDMA for mobility, freq & multi-user diversity

– Support for fixed SC, MC & mobility

– IP based architecture



• LAN

–Already covered under 802.11



• 3G:

– Fixed operating bandwidth

– Very difficult & expensive to spread higher data rate using CDMA

– Established infrastructures in process of upgrading

– GSM using UMTS and or HSDPA

• DL only, 14.4 Mbps in 5M BW using 15 codes (specified but a challenging task)

• DL only, 3.6 Mbps in 5M BW using 5 codes but typically averages about 250 kbps

• DL only, 7.2 Mbps in 5M BW using 10 codes but typically averages about 750 kbps

• UL specified to support 384 kbps but typically averages about 40 to 100 kbps

– CDMA using 1x EV-DO

• Rev A DL, specified to support 2.4 Mbps in 1.25 M BW but typically 100 to 300 kbps

• Rev A UL, specified to support 1.8 Mbps in 1.25 M BW

• Rev A to provide low latency of 30 ms, VoIP, video, QoS, fast handoffs and broadcast

applications

• Rev B is specified to support 73 Mbps DL, 27 Mbps UL in 20 M BW



• Others

– IEEE820.20

• Standard under development

• For mobility up to 250 kmph

• 3.5 G band

• 4 Mbps DL, 1.2 Mbps UL

– IEEE820.22

• Standard under development

• Broadband access targeted for far reaching rural area

• Using unused VHF & UHF bands


WiFi vs. WiMAX

Parameter 802.11 WiMAX

Licensed Band Operation No Yes

AGC Range (dB), 64QAM N/A 50

OFDMA No, yes with 803.2a Yes

Advanced antenna None Standard supports advanced antenna technique

Mesh antenna Can introduce mesh topology, but not supported by Std supports mesh network topology

Power Output (dBm) Restricted in unlicensed bands ≥43, per local regulatory requirements

Range

Optimization centers around PHY and MAC layer for

100m range.

Range can be extended by cranking up the power - but

PHY and MAC designed with multi-km (40) range in

mind.

Standard MAC

NF, dB 10 7

SNDR (dBc), TX ≤31 ≤31

Alternate channel Rej N/A 30

Channel BW (MHz) 10, 20

Channel bandwidths can be chosen by operator (e.g.

for sectorization). 1.25 MHz to 20 MHz width channels.

MAC designed for scalability independent of channel

BW

Maximum bps/Hz ~2.7 ~4.5

User handling No "near-far" compensation

Designed to handle many users spread out over

distance

Multipath

Optimized for indoor non-line-of-sight (NLoS)

performance

Designed to tolerate greater multipath delay spread

(signal reflection). Optimized for outdoor NLoS.

MAC capability MAC designed to support 10‟s of users MAC designed to support thousands of users

MAC operation

Contention - based MAC (CSMA/CA) ≥ no guaranteed

QoS Grant-request MAC

Latency

Standard cannot currently gaurantee latency for Voice,

Video Designed to support Voice and Video from ground up

Services

Standard does not allow for differentiated levels of

service on a per user basis

Supports differentiated service levels: e.g. T1 for

business customers; best effort for residential

Transmission TDD only - asymmetric TDD/ FDD/HFDD – symmetric or asymmetric

QoS

No QoS today. 802.11e (proposed) QoS is prioritization

only Centrally enforced QoS

FEC Convolution RS & CC

Security

Existing standards is WEP. 802.11i in process of

addressing security RSA (1024 bits)

Encryption Optional RC4 Mandatory, triple-DES (128 bits)


Technology Throughput Comparison


Spectral Efficiency Comparison


Point to Multipoint WiMAX Applications


WiMAX, (1)

• Stands for Worldwide Interoperability for Microwave Access

• A cost effective alternative to wireline services especially in the developing countries where no existing wireline services available

• Operation in licensed & licensed exempt bands using 1.25-20 MHz BW

• QoS, advanced security & higher throughput than WiFi

• Supports QoS, VoIP, video distribution, on-line gaming & real time video conferencing

• Standards and interoperability is the key to its success

• GPS and IEEE 1588 over IP to synchronize the network from a master clock

– GPS may be more expensive and difficult to access open sky if in the basement

– IEEE 1588 requires a master source access in the network

– WiMAX network is entirely IP and there is no option of recovering timing signal (without embedded mechanism) as there is with the TDM application

• Dynamic frequency selection for operation in unlicensed bands

• Longer wavelength makes multipath more significant

– LOS not feasible in residential applications

– There may be cost associated with outdoor mounted antenna

• Uses a very versatile configurable modulation schemes that adds complexity


WiMAX, (2)

• Why we need it? Demand created by internet & mobile usage. Broadband access to residential, SOHO, SME, backhauling hotspot, long wait time for increased T1 services. Lack of land lines in some countries. Mobility to offer in two stages (portable-nomadic and seamless mobility)

– Higher data rate

– Multiple levels of guaranteed QoS

– Stationary & Mobility

– Multipath tolerant by using multiple lower frequency carriers

– Switchable mode

– Concatenated FEC

– Low latency

– Security

– Ease of installation

– Lower cost deployment and operating solution

– Large system gain (about 150 dB) and coverage range


WiMAX, (3)

• Mobility under 120 km/hr (target applications are handset,

laptop). Standard released in Nov‟05.

• Support for both LOS & NLOS

• It is not mandated by standard but TDD is most likely

mode of choice for mobile applications because it divides

the entire frequency spectrum into upstream and down

stream time slots (more efficient use of limited frequency)

• Uses all IP backbone

• OFDMA-PHY with sub-channelization allows time &

frequency resources to be dynamically allocated among

multiple users across DL & UL subframe


WiMAX, (4)

• Broadcast and Multicast support, low latency ≤ 100 ms, low to zero packet loss during handovers at speed 120 km/Hr or higher

• Simultaneous support of real time multimedia and isochronous applications like VoIP

• Simple self installed user station (SS/MS)

– Automated management of IP connection with session persistence

– Automatic reestablishment following transitions between access points

• Likely applications: single carrier for back haul, OFDM for fixed access in up to 28 MHz BW, scalable OFDMA is most versatile and preferred for mobile operation in 1.25 to 20 MHz BW

• Frequency inaccuracy of 1e-6 max for FDD and TDD

• Time accuracy: N/A for FDD but 5-25 us for TDD


Other Systems

• Non-standard i-Burst from ArrayComm and Flash-

OFDM from Qualcomm

• 3G (UMTS, HSDPA by GSM operators, EV-DO by

CDMA), WiFi (higher data rate than 3G due to 20 M

BW, using inefficient CSMA protocol-”carrier sense

multiple access”)

• WiFi standard 802.11n is being enhanced to support

100 Mbps, better QoS, transmit diversity and other

enhanced techniques


WiMAX Standards, Timeline

• Evolving standard, work started under 802.16 in 1999

• 802.16a, Jan‟03

• 802.16d, July’04, replaced 802.16, a & c. 895 pages

• 802.16e, MAC function to support higher layer handoff in under 6 GHz band, Dec’05. 864 pages

• 802.16f, fixed WiMAX management information base. Added multi-hop functionality

• 802.16g, management procedures and interfaces for fixed and mobile 802.16 systems. Addresses efficient handover and further improves the QoS support.

• 802.16h, mechanisms, policies and MAC enhancements for coexistence in licensed exempt bands

• 802.16i, (with drawned). Mobile WiMAX management information base, merging into 802.16-2008

• 802.16j, multi-hop operation

• 802.16k, bridging amendment

• 802.16m, advanced air interface for next generations, higher data rate and higher speeds


802.16s Comparison, (1)

Standards 802.16 802.16a 802.16d 802.16e

Completed Dec‟01 Jan‟03 June‟04 Dec‟05

Alignment Mode LOS only LOS & NLOS LOS & NLOS NLOS

Spectrum 10-66 GHz 2-11 GHz,

licensed &

license exempt

2-11GHz 2-11GHz for

fixed, 2-6GHz

for mobility

Bit Rate Up to 134 Mbps Up to 75 Mbps Up to 75 Mbps Up to 15 Mbps

Bandwidth 28 MHz 20 MHz 20 MHz 5* MHz

Modulation Single carrier

only

Single carrier,

256 OFDM or

2048 OFDM

sub-carriers

Single carrier,

256 OFDM or

2048 OFDM

sub-carriers

Single carrier,

256 OFDM or

scalable

OFDMA (128,

512, 1024, 2048

sub-carriers)


802.16s Comparison, (2)

Standards 802.16 802.16a 802.16d 802.16e

Modulation 4/16/64 QAM 4/16/64 QAM,

256Q optional

4/16/64 QAM,

256Q optional

4/16/64 QAM,

256Q optional

Mobility Fixed Fixed/Portable Fixed/Nomadic Mobile/Portable

Bandwidth 20, 25, 28 MHz 1.25-20 MHz 1.25-25 MHz 1.25-20 MHz,

uplink to

conserve Po

MAC

Architecture

PMP mesh,

TDD and FDD

PMP mesh,

TDD and FDD

PMP mesh,

TDD and FDD

PMP mesh,

TDD and FDD

Applications E1/T1 services,

backhauling

hot spots

E1/T1 services,

backhauling

hot spots.

Wireless DSL

Indoor

broadband

access for

residential

users (HSpeed

internet, VoIP)

Portable

broadband

access for

consumers.

Mobile internet.

Always best

connected


WiMAX & Others

Parameter Fixed WiMAX Mobile WiMAX HSPA

1xEV-DO

Rev A WiBRO Wi-Fi

Standards IEEE 802.16d IEEE 802.16e 3GPP, Release 6 3GPP2 IEEE 802.16e 802.11a,b,g

FTT size 256, 2048

2048, 1024, 512,

128 N/A N/A 1024 64

User carriers 1680/1728 various N/A N/A 864/840 52

Pilot carriers 166/192 various N/A N/A 96 4

MIMO Yes Yes No No Yes No

Guard period 1/4, 1/8, 1/16, 1/32 1/4, 1/8, 1/16, 1/32 N/A N/A 1/4,1/8,1/16,1/32 1/4Multiple users

over frequency

(@1 symb time) Yes Yes No No Yes No

Multiple users

over time (@1

channel) Yes Yes No No Yes No

Peak DL data

rate

~4.5bps/Hz,

9.4Mbps in 3.5MHz

with 3:1

DL to UL ratio TDD,

6.1 Mbps with 1:1

~4.5 bps/Hz,

46Mbps with 3:1

DL to UL ratio TDD,

32 Mbps with 1:1

14.4 Mbps using

all

15 codes, 7.2

Mbps with 10

codes

3.1Mbps, Rev B

will support

4.9Mbps

~4.5 bps/Hz, 38

Mbps with 3:1

DL to UL ratio

TDD, 27 Mbps

with 1:1

~2.7 bps/Hz peak data

rate, 54Mbps shared in

20MHz using

802.11a/g more than

100Mbps peak layer 2

throughput using

802.11n

Peak UL data

rate

3.3Mbps in 3.5MHz

with 3:1

DL to UL ratio TDD,

6.5 Mbps with 1:1

7Mbps in 10MHz

with 3:1

DL to UL ratio, 4

Mbps with 1:1

1.4Mbps initially,

5.8Mbps later 1.8Mbps

5.9 Mbps with 3:1

DL to UL ratio, 3.4

Mbps with 1:1 same as above


WiMAX & Others (conti’d)

Parameter Fixed WiMAX Mobile WiMAX HSPA

1xEV-DO

Rev A WiBRO Wi-Fi

Bandwidth

3.5MHz and 7MHz

in

3.5GHz band,

10MHz in 5.8GHz

band, 28M max

3.5M, 7M, 5M,

10M & 8.75M

initially, 28M max 5M 1.25 x 2 M 8.75M

20M for 802.11 a/g,

20/40M for 802.11a,

11M for 11b

Modulation

QPSK, 16QAM,

64QAM

QPSK, 16QAM,

64QAM QPSK, 16QAM

QPSK,

8PSK,16QAM

QPSK, 16QAM,

64QAM

BPSK, QPSK,

16QAM,

64QAM

Multiplexing TDM TDM/OFDMA TDM/CDMA TDM/CDMA TDM/OFDMA CSMA

Duplexing TDD, FDD TDD initially FDD FDD TDD TDD

Frequency

3.5G and 5.8G

initially

2.3G, 2.5G

&3.5G initially

800/900/1800

/1900/2100M

800/900/1800

/1900M 2.3G 2.4G, 5G

Coverage

(typical) 3-5 miles <2 miles 1-3 mile 1-3 miles 1-3 miles

<100ft indoor

<1000ft outdoor

Mobility Not applicable Mid High High High Low

QoS

QoS designed in

for voice/video,

differentiated

services.

Grant request MAC

QoS designed in

for

voice/video,

differentiated

services. Grant

request MAC DL only DL only

QoS designed in

for

voice/video,

differentiated

services. Grant

request MAC

No QoS

support.802.11e

working to

standardize.

Contention based

MAC


OFDM, Orthogonal Freq Division Multiplexing

• A combination of modulation and multiplexing technique

• Mapping of information on changing in the carrier phase, freq., amplitude or a combination

• Method of sharing bandwidth with other data channels

• Channel bandwidth divided by a number of sub-channels

– Aggregate data rate throughput is about the same but data rate on each sub-carrier is much lower

– Longer symbol time practically eliminates the effects of variable time delays

• Integer number of cycles to complete for each sub-channel

• OFDM bit rate is based on number of active data sub-carriers, not the bandwidth

• Orthogonality allows simultaneous transmission on many sub-carriers in tight frequency space without interfering each others


OFDM, Interference Response, Example 1

• Single carrier like water flowing from a faucet

• Multi-carrier like water flowing from a shower head


OFDM, Interference Response, Example 2

• Reliable delivery mechanism


OFDM Basic-1

• Converts single bit stream (wider bit rate) into multiple (smaller bit rate) parallel bit

streams

• Efficient BW (No BPF between sub-channels) usage

• Orthogonal approach using FFT technique (50 yrs old, used to be expensive to

implement). Signal orthogonality happens in frequency domain (peak of one signal

at zeros of all others)

• Time & freq domain representation

Ch1

Po

we

r

Ch5Ch4Ch3Ch2

Po

we

r

Bandwidth saving

Ch1 Ch5

Ch4

Ch3Ch2

FDM-Frequency

OFDM-Frequency


Signal in Time & Frequency Domain

Tg

TFFT

Ts

Subcarriers

Guard

Intervals

Symbols

Time

Frequency

FFT


OFDM Basic-2

• An area under a complete sine/cosine

wave is always zero

– When multiplied by another integer or

non-harmonics, the result is zero

– It is non-zero only when multiplied by the

same harmonics

– Integral 0 to T of sin2π(ft)*sin2π(2f)t dt=0

where T is multiple of 1/f

• FDM must apply a bank of RRC filters

– OFDM uses optimized channel spacing

and RRC does not buy you much

• A proven technology, already being

used in cable modem, WiFi, DSL, DVB

and DAB


OFDM Basic-3

Rs Rs/N f2

fN

f1

Mod N

Mod 2

Mod 1

S/P

∑

Time Wavefom Carrier Freq using N point FFT

Time

Fre

qu

en

cy

One OFDM Symbol

Data

T=1/f0

Time-frequency Grid

f0

Subcarrier


OFDM Basic-4

• Treats source Symbol as frequency domain and converts it into time domain with IFFT

• N carriers results in N orthogonal-sinewaves

• N tones are generated digitally to avoid bank of phase locked oscillators

• Each N determines a complex Amplitude & phase for that sub-carrier

• Sin(x)/x spectra for unfiltered sub-channels but their orthogonality prevents interfering to one another

• Output of IFFT is sum of all sub-carriers and makes up a single OFDM symbol (whose length = NT, T is IFFT input sampling period) in time domain

DMD

M-QAMFFT BB OFDM

A/DData Out

IFFTData In

DACMod

M-QAM

BB OFDM


Example of an OFDM Parameters

• FFT/IFFT process must use 2^N tones (use zero filling for non-used tones)

• Each sub-carrier produces its sin(x)/x spectra

• Relatively simple DSP algorithms

• 256 subcarriers (192 data + 8 pilot + 28 null at start + 27 null at end + 1 DC),

center subcarrier is not used due to being easily susceptible to RF carrier feed-

through

• Constant number of subcarriers regardless of BW. Advantage at Narrow BW. The

subcarriers are variable for S-OFDMA.

• IFFT Symbol (active subcarriers) vs. OFDM Symbol (IFFT plus the gap)

• Configurable DL and UL frame length from 2.5 to 20 ms

• Each UL subframe is preceded by preamble to allow BS to sync on each individual

SS

B1 RTGTTG P B3B2 PP B4PP H B1 B2 B4B3

1 Frame (2.5 to 20 ms)

Downlink subframe (basestation) Uplink subframe (subscriber)


SC (Single-carrier) vs. MC (Multi-carrier)

Modulated System

• SC: each user transmits & receives data with only one carrier at the same time

– Serially modulated scheme

– To increase throughput would require higher symbol rate which is more susceptible to channel effects

• Highly susceptible to crosstalk, ISI, multipath fading

– BW = 1+ % RRC alpha roll-off, 3 dB bandwidth of SQRT raised cosine filter

• Decreasing roll-off increases the PAPR and interference

• MC: each user can employ number of subcarriers to transmit data simultaneously

– Parallel modulation scheme

– Slower subcarrier rate that makes it easier to process and more rugged against interference

– Simpler frequency domain equalizer

– Manageable DSP

– Robust to interference


MC, Multi-carrier

• Divide the shared wideband channel into N sub-channels

– Data divided into Na active substreams

• Substream modulated onto separate subcarriers

– Substream bandwidth is B/N where B a total bandwidth

– B/N < Bc implies flat fading on each subcarriers (no ISI), Bc a

coherence time bandwidth

• Spacing between two carriers is proportional to 1/T, where T

is the IFFT symbol duration

– BWOccupied = NaxTIFFT, sharper roll off due to lower sub-carrier frequency (higher

IFFT rate) & DSP process. Na active subcarriers, TIFFT IFFT sampling duration


Impairment Affects on SC vs. MC

• Impairments affect differently on SC vs. MC systems

Impairment OFDM Single Carrier

IQ gain balance State spreading (uniform/carrier) Distortion of constellation

IQ Quadrature skew State spreading (uniform/carrier) Distortion of constellation

IQ channel mismatch State spreading (nonuniform/carrier) State spreading

Uncompensated freq. error State spreading Spinning constellation

Phase noise State spreading (uniform/carrier) Constellation phase arcing

Nonlinear distortion State spreading

State spreading (may be more

pronounced on outer state)

Linear distortion Usually no effect (equalize) State spreading if not equalized

Carrier leakage

Offset constellation for center

carrier only (if used) Offset constellation

Frequency error State spreading Constellation phase arcing

Amplifier droop Radial constellation distortion Radial constellation distortion

Spurious

State spreading or shifting of

affected subcarrier

State spreading,

generally circular


Multiple Carrier Modulated System

• MC Advantages– Data are shared among several

subcarriers and simultaneously transmitted. Pulse length ~N/B

– Flat fading per subcarrier

– N long pulses

– ISI is comparatively short

– N short EQs needed

– Facilitate NLOS operation with added guard interval

– Manage spectral efficiency with null subcarriers

– Easier to exploit frequency diversity

– Allows to deploy 2D coding techniques

– Dynamic signaling

– Exploit MIMO operation

• MC Disadvantages

– Higher linearity requirements due to

PAPR

• Reduced system gain due to

additional back off

• Higher power devices require

more power dissipation, real

estate space and cost

– Sensitive to phase-noise and clock

inaccuracy

– Additional circuit, processing

resources and cost for IFFT/FFT

– Reduced spectral efficiency due to

added guard interval


SC, Single Carrier

• SC Advantages

– Efficient and lower power

consumption

– Complexity of transmission is

much simpler than that of

reception, making it suitable for

asymmetrical operations

– High level of narrow-band noise

immunity due to inherent

capability by use of adaptive

equalization

– Less susceptible to phase noise

– Lower peak to average ratio

– Frequency domain equalization

for performance improvement

• SC Disadvantages

– Data are transmitted over only

one carrier. Pulse length ~1/B

– Selective fading

– Very short pulse

– ISI is comparatively long

– EQs are then very long

– Poor spectral efficiency because

of band guards

– Sensitive to group delays


MC, Multi-carrier

• For each subcarrier, Rx receives a composite of sinusoids

– Same frequency but different phase and amplitude

– Fairly robust in frequency selective fading channel

• For single carrier transmission system, if the channel encounters interference at this frequency, the entire transmission can fail

• In OFDM, the problem is reduced since only a few of the N subcarriers will be affected. This means loss of a few bits instead of the entire OFDM symbol

• Powerful error correcting codes can be used to help restoring the erroneous bits in the corrupted subcarriers

Delay

SCHi-Freq Signal

Combined Signal

Multipath Signal

Delay

OFDMLow Freq Signal

Combined Signal

Multipath Signal

Time

Fre

qu

en

cy

One OFDM Symbol

Data bits

T=1/f0

Time-frequency Grid

f0

Bad sub-

carriers

Use

d B

an

dw

idth


SC vs. MC, (1)

• The ability to overcome

delay spread, multi-path,

and ISI in an efficient

manner that allows for

higher data rate throughput

• As an example, it is easier

to equalize the individual

OFDM subcarriers than it is

to equalize the broader

single carrier signal

Frequency

Single Carrier OFDM Mode

Frequency

Symbol have

narrow freq. long

symbol time

Each of the symbols is used to

modulate a separate carrier

Symbol have

wide freq. short

symbol time

Serial symbol stream used to

modulate a single wide band carrier

Level

TimeS0 S5S4S3

S2

S1

S0

S5

S4

S3

S2

S1

OFDM Mode

FrequencyFrequency

Single CarrierLevel

The dotted area represents the transmitted

spectrum. The solid area is the receiver input

Deep Fade


SC vs. MC, (2)

• Single Carrier systems are fairly robust to frequency offset

errors and are more appropriate for mobile environment

that experience large frequency offset errors

• The complexity of the equalizer for Single Carrier system

is much greater than the multi-carrier system

• Multicarrier systems are fairly robust to timing errors

compared to a Single Carrier system. Their performance is

similarly affected by the loss in SNR caused by frequency

offset errors. Intuitively, this is easily understood from the

fact that the Multicarrier symbol duration is N-times longer

than its single carrier counterpart operating at the same

data rate


LOS vs. NLOS

• LOS, direct non-obstructed path

– LFSL = 10Log(4πDm /λ)2, Dm distance in m, λ=c/f wavelength in m

– Optical LOS, Dkm = 3.57 SQRT (H), H antenna height

– Radio LOS, Dkm = 3.57 [SQRT (kHB)+SQRT(kHM)], k effective earth k-

factor

• NLOS, Rx signal reaches through reflections, scatterings

and diffractions

– Signal have different delay spreads, attenuations, polarization & stability

relative to direct path

– OFDM technology takes advantage of this phenomena

– LNLOS = Free space loss + terrain induced loss


Fresnel Zone

• Fresnel zone clearance depends on frequency & path length

– 1st Fz = 0.5 wavelength = 17.31 SQRT{(d1* d2) / (Dkm fGHz)}, d1 & d2 distance from obstruction to antenna, Dkm total distance

– Destructive affects at even orders of Fz

• Signal summation of same and or opposite phase


Multipath Fading

• More than one transmission path between Tx and Rx

• Receive signal is the sum of many versions of the Tx signal with varied delay and

attenuation

• Reflection occurs when a propagating electromagnetic wave impinges upon a smooth

surface with very large dimensions relative to the RF signal wavelength (λ=c/f, c speed of

light, f operating frequency)

– Buildings, ground, billboards, media

• Diffraction occurs when the propagation path between Tx and Rx is obstructed by a dense

body with dimension that are large relative to λ. Wave bends around sharp objects

– Terrain, top of buildings

• Scattering occurs when a radio wave impinges on either a large, rough surface or any

surface whose dimensions are on the order of λ or less, causing the energy to be spread

out or reflect in all directions. In an urban areas it is caused by lamppost, street signs and

foliage.

• Multipath propagated signal affected by

– Velocity, path, attenuation, time delay, Doppler shift, number of paths, etc.

• PRX = PTX GTX GRX (λ / 4πD)2 for LOS path

Reflection

Diffraction

Scattering


Two-Ray Ground Propagation Model

• If there are obstructions between the transmitter and receiver, wave will traverse multiple paths

– Radio waves arrive at receiver from different directions and with different time delays

– Resultant signal at receiving antenna is vector addition of incoming signals

– Individual signals can add constructively (resultant signal has large power) or destructively (resultant signal has small power) depending on relative phases

• EM wave impinging on surface will be reflected with some attenuation (determined by reflection coefficient)

• Two-ray model assumes one direct LOS path and one reflection path each reaching receiver with significant power


Typical Signal Attenuation

EnvironmentSig. Attn

dB @ 2.5GHzDrywall, per cm 2.1

Sheetrock wall, 2x4 6

Office whiteboard, per cm 0.3

Clear glass, per cm 20

Mesh glass, per cm 24.1

Office wall 10

Wooden wall 15Brick wall 30

Metal wall 45

Foliage, 3 m deep 8.3

FSL, 1 km 100.4

Rural open space, 1 km 104

Suburban, 1 km 117

Urban, Newark, 1 km 119

Urban, Philadelphia, 1 km 125Urban, Tokyo, 1 km 139


Sector Antenna Pattern, example


Frequency Selective Scheduling

• OFDMA is fairly resistive to frequency selective

fading since its parallel nature allows errors in

sub-carriers to be corrected

• Mobile WiMAX signal occupies a portion of the

bandwidth. In broadband wireless channels,

propagation conditions can vary over different

portions of the spectrum in different ways for

different users. Mobile WiMAX supports

frequency selective scheduling to take full

advantage of multi-user frequency diversity and

improve QoS. WiMAX makes it possible to

allocate a subset of sub-carriers to mobile users

based on relative signal strength. By allocating a

subset of sub-carriers to each MS for which the

MS enjoys the strongest path gains, this multi-

user diversity technique can achieve significant

capacity gains.

F1F1

F1F1

F1F1

F1

F1F1

F1

F1

F1F1

F1

F1

F1

F1

F1

F1

F1

F1

F1

F1

F1


Frequency Reuse Approach (1)

• Sectorized vs. Omni antenna

– Higher directivity (higher gain). Must do up front

performance tradeoffs

– Interference, range and cost tradeoffs

– Interference from overlapping section of the sector’s

edge

• Single & dual-pole antenna

– Lower order modulated signal may use the same

frequency adjacent sectors on single polarized antenna

– Higher XPD requirements on a dual poled antenna

especially for higher order modulated signal

• Applies dynamic frequency reuse across sectors based

on loading and interference conditions

– Allocating non-overlapping subchannels for poor SINR

area at the expense of spectral efficiency

F1

F2

F3

F1

F2

F3

F1

F2

F3

F1

F2

F3

F1

F2

F3

F1

F2

F3

F1

F2

F3


Frequency Reuse Approach (2)

• (NcxNsxNf definition), Cluster of cell x Sectors per cell x Frequencies per cell

– 1x3x1, Higher interference (CCI) higher spectral efficiency

– 1x3x3, Lowest interference (CCI & ACI) lower spectral efficiency

– 1x3x11-3 Subchannels, Lower interference & lower spectral efficiency at edge

• Transmission from BS (all sectors) and SSs must be synchronized while using different

permutation subchannels to minimize interference

• PUSC typically uses 1/3 subcarriers per sector

– Randomly assigns subcarriers to subchannels using PUSC scheme in an unloaded network

(not very affective when load increases)

• FUSC & AMC would result in large coverage holes

F1

F2

F3

F1

F2

F3

F1

F2

F3

F1

F2

F3

F1

F2

F3

F1

F2

F3

F1

F2

F3

F1S3a

F1S2a

F1S1a

F1S3a

F1S2a

F1S1a

F1S3a F1S1a

F1S2a

F1S3

F1S2

F1S1

F1S3

F1S2

F1S1

F1S3

F1S2

F1S1


Frequency Reuse (3)

• Mobile WiMAX also support frequency reuse one, i.e. all

cells/sectors operate on one frequency channel to

maximize spectrum utilization. However, due to heavy

interference in (common frequency) reuse „1‟ deployment,

users at the cell edge may suffer low connection quality

• In WiMAX the sub-channel reuse pattern can be

configured so that users close to the base station operate

on the zone with all sub-channels available. While for the

edge users, each cell/sector operates on the zone with a

fraction of all sub-channels available.


Fractional Frequency Reuse (4)

• Subchannel reuse planning can improve the coverage across the cells/sectors

based on network load and prevents interference

• F1a, F1b and F1c represent different sets of sub-channels of the same frequency

• Full frequency use (maximum sub-channels) at the center while fractional

frequency use at the edges

• The sub-channel reuse planning can be dynamically optimized

• Other implementation forms include time-coordinated and power-coordinated

transmission

• Transmission across BSs and sectors are coordinated in order to achieve maximal

interference avoidance

Fractional Freq Reuse

1x3x1 Reuse

1x3x3 Reuse

F1

F3

F2

F

F

F

F1

F3

F2

F

F

F

F1

F3

F2

F

F

F

F=F1+F2+F3

F1: F,S1 F2: F,S2 F3: F,S3

F1=F1a + F1b + F1c

F1a

F1b F1c

F1

F1 F1



• .

fre

qu

en

cy

Reuse 1 Area Reuse 3 Area

All

Resources

Reuse

Partition 3

time

Reuse

Partition 2

Reuse

Partition 1

time

Pre-

amble

Center cell

FFR = 1

Whole cell

FFR = 3

DL subframe UL subframe

G

A

P

Pre-

amble

Center cell

FFR = 1

Whole cell

FFR = 3



• .

Frequency

Ce

ll 2

Ce

ll 1

Ce

ll 3

Power

Frequency

Ce

ll 2

Ce

ll 1

Ce

ll 3

Power

Frequency

Ce

ll 2

Ce

ll 1

Ce

ll 3

Power

Frequency

Ce

ll 2

Ce

ll 1

Ce

ll 3

Power

Uniform Hard reuse 3 Fractional reuse 3 Soft reuse 3

1

2

3

1

2

3

1

2

3

1

2

3

FFR-A Scheme

Frequency

Cell 2

Cell 1

Cell 3

1

2

3

FFR-B Scheme

Frequency

Cell 2

Cell 1

Cell 3

1

2

3

Reuse-3 Scheme

Frequency

Cell 2

Cell 1

Cell 3

1

2

3


DFS, Dynamic Frequency Selection

• Feature used in license-exempt frequency band only

• Automatically detects and avoids interference by moving

to a different frequency location within the band

• Prevents harmful interference into other users

• Provides improved system performance

• A mandatory feature


RF/ Mixed Signal Impairments

• RF-LO phase noise, not correctable by adaptive equalizers

• Inter-modulation distortions, at inputs, at outputs or both

• Amplitude & group delay distortions

• Improper cable termination

• PA compression

• PA switching and settling time

• Antenna mismatch and low isolation

• Low signal combiner isolation

• Burst shaping error

• Recovered clock jitter

• IQ gain imbalance, IQ phase imbalance, (in-band spurious)

• Filtering distortions (normally compensated by signal processing)

• Distortion due to DC offset compensation

• DC offset (canceled in analog and signal processing)

• A/D and D/A converter non-linearities

• Thermal noise


Digital Baseband Impairments

• Improper channel estimation

• IFFT, FFT

• Equalizer

• Incorrect coefficients

• Viterbi decoder

• BB-LO phase noise

• Timing / Frequency Sync

• Digital insertion loss

• Latency

• Processing circuit noise


Commonly Used Implementation

Architectures & its Characteristics

• Super-heterodyne (dual conversion)

– Needs of channel IF filtering (external component) and two synthesizers. Less stringent IF filtering

– Good immunity from interfering signals and good selectivity performance

– Image is not a serious problem

• Heterodyne with not fixed wide IF (2nd LO by division of 1st LO)

– Removes part of DC offset issues, LO emission, pulling and flicker noise

– Higher complexity, more spurious, IQ imbalance & power consumption

– Spurious associated with the 2nd LO and IF frequencies: careful frequency plan required

• Low IF

– Removes part of DC offset issues and flicker noise

– Higher complexity, sensitive to IQ paths asymmetry

• Homodyne (Zero-IF)

– Reduced parts count, saves die/board size and power consumption

– Simple Frequency Plan – Spurious and higher order mixing products associated with the 2nd LO and the IF frequencies are also eliminated from the frequency plan

– Isolation and dynamic rage trade-off

– Sensitive to DC offset , LO emission, LO pulling, flicker noise


Common Path Related Impairments

• Distance dependent decay of the signal power

• Blockage due to obstructions

• Large variation in received signal envelope

– Due to constructive/destructive additions of multi-path signals

• ISI due to time dispersion

• ICI due to local clock inaccuracy & phase noise

– More critical for TDD than the FDD system

– If occurred, it is not correctable

• Synchronization vs. clock drift

• Frequency dispersion due to motion

• Noise

• Interference from own & or intra-network equipment


Other Impairments

• Atmospheric absorption – water vapor and oxygen

contribute to attenuation (not relevant for low freq

WiMAX)

• Multipath effects by terrain and environmental

conditions

–Obstacles reflect signals so that multiple copies with

varying delays are received

• Refraction – bending of radio waves as they

propagate through the atmosphere


TX PO

• Max PO, regulated by local regulatory agency

• Asymmetric power level at SS, MS & BS

• Different device sizes may yield asymmetric performance

at each end

• PO, determined by sum of power from all active sub-

carriers measured over certain number of symbols in time

– Sub-carrier power varies depending on type of sub-carrier,

modulation and content

– Total data PO =DataPWR of 1subcar +10Log(# of act data subcar)

– Total pilot PO =pilotPWR of 1subcar +10Log(# of act pilot subcar)

– Total symbol PO =10 Log(10^dataPWR/10 + 10^pilotPWR/10)


TX PO Constraints and Impairments

• Requires vector power meter to measure specific symbol power (not feasible with traditional power meter due to TDD, DL/UL ratio, adaptive modulation, burst rate, training sequence etc.)

• MC system demands increased PA linearity for reliable high performance

• PAPR = 10 Log(# of subcarriers), additional requirements than SC

– Creates extreme peaks and valleys

– Normally not a serious occurrence issue due to data scrambling

• PA linearity may improve at the expense of

– Increased power back off

– Larger device (resulting in increased cost, power, real estate, thermal rise and lower reliability). Perform upfront trade offs

– Adaptive distortion control

• Increases cost

• Increases control algorithm complications

• Reduces processing resources

• Inappropriate power level affects system performance, spectral mask, spectral flatness, spurious, interference to other equipment, etc.


TX PO Impairments

• PA non-linearity causes IMD that results in spectral regrowth

– Select higher OIP3, OIP5, IIP3, IIP5, P1dB, SFDR to improve performance

– Demands more linearity at higher order modulation

– Non-linear distortion can not be corrected by equalizer

• Spurious may also originate at other areas of the circuit such as

in non-linear mixer, LO phase noise, DAC, IQs, filters, etc.

• Performance degradation affects at its own near-end / far-end Rx

and other operators in the vicinity

• Regulatory agency controls the Tx signal quality (Po, Freq., BW,

spectrum, spurious, noise floor, ACLR, etc.) in order to protect

other operators


TX Impairments

• Tx impairments affect the performance of its own and other neighbors in the

vicinity

• Mitigation techniques includes power back-off, distortion control, larger device,

signal clipping, selective mapping, partial IFFT, etc

• Destructive effects resulting from IMD3 & IMD5

1dB

3rd

Intermod

Fundamental

SFDR

BDR

P1dB IIP3

Input

Power

Ou

tpu

t P

ow

er

Noise

Floor

3rd

5th

Fundamental

OIP3

Rx Thres

NF

SNR

kTB

Freq

Output

Power

f1 f22f1-f2 2f2-f1

IMD3IMD3

3f2-2f1

Fundamental

SFDR

IMD5

Δf

Δf


Spectral Spreading Control

• Commonly used mitigation techniques includes:

– Operate at increased power back-off

– Forcing counter distortion

– Using larger PA devices

– Signal clipping

• Digital domain clipping also introduces spreading and minimizes the

effective SNR

• Passing the clipped signal through BPF prior to PA eliminates

spreading

– Selective mapping

– Partial IFFT

From OFDM

modulator Clip to

specified Pre-

filter OBO

BPF user

FIR

Simulated

Transmitted

signalClip to specified

Output Power

Amplifier OBO


TX spectral mask

• Reference ETSI EN302 326-2 & EN320 544-1

• RBW is generally set to about 1% of the BW if not specified

• ACLR: 44.2 dB at x1 CS, 49.2 dB at x2 CS (Channel Spacing)

• The spectral mask basically specifies the accuracy of the out of

band signal

4QAM

16QAM

64QAM

-55

-45

-35

-25

-15

-5

0 1 2Frequency/CS

Att

en

uati

on

(d

Br)


TX Spectrum, (1)

• Frequency domain representation of one OFDM symbol

• Modulation scheme & power adjustable per sub-channel


TX Spectrum, (2)

• Higher spectrum efficiency

– Place unused sub-carriers at the beginning & end of OFDM symbol

– Rectangular spectrum shape (almost like brick wall)

– For larger number of subcarriers the spectrum goes down rapidly in the

beginning, which is caused by the fact that the side lobes are closer

together

– Roll off relative to sub-carrier rate

– Small frequency guard band

• BW= ½ BOU + ½ BOL + (FOH – FOL)

– BOU, BW of upper subcarrier

– BOL, BW of lower subcarrier

– FOH, Upper frequency edge

– FOL, Lower frequency edge

dB

FreqX MHz

-80

OFDM Single Carrier


Tx Spectrum, (3)

• Use vector spectrum analyzer to capture a non-traditional signal: TDD, DL/UL

ratio, adaptive modulation, burst rate, training sequence, etc.

• Sharp almost brick wall like spectrum, allows more data in the allowed BW

• Tx spectral flatness to be within 2 dB over all active tones for spectral lines

starting from -50 to -1 and +1 to +50. +2/-4 dB over all active tones for spectral

lines from -100 to -1 and +1 to +100. To be within 0.1 dB for adjacent sub-

carriers


TX Spectrum, (4)

Preamble


TX BW

• Tx BW is determined by total active data and pilot

sub-carriers

– For example, BWAllowed =20 M, NFFT = 2048, active data = 1440,

pilots = 240, subcarrier spacing = 11.160714 kHz

– BWOccupied = (1440+240) * 11.160714 = 18.75 MHz

– If an input data rate R bps, Nused of FFT, M Modulation order, 3/4

FEC then each active data subcarrier carries {(R/Nused) * (4/3) * M}

load


TX Frequency

• MC system demands higher frequency stability &

accuracy to deliver a consistently reliable

performance

–≤ 1 ppm is required for FDD & TTD operation over the life

of product

• This equates to 2.5 kHz for Tx only at 2.5 GHz

• The subcarriers frequency is typically about 10 KHz

–SS to BS synchronization tolerance to be ≤ 2 Hz

–Timing accuracy of 5-25 us required for TDD system

–Frequency inaccuracy increases the ICI


Thermal Noise, (1)

• Thermal noise due to agitation of electrons

• Present in all electronic devices and transmission

media

• It cannot be eliminated

• Function of temperature (increases at higher

temperature, 1.68 dB from -33 C to +80 C)


Thermal Noise, (2)

• Amount of thermal noise to be found in a bandwidth of 1 Hz in any device or conductor is:

• N0 = noise power density in watts per 1 Hz of bandwidth

• N0 = -173.93 dBm/Hz into a 50 Ohms load (antenna) at room temperature

• k = Boltzmann's constant = 1.3803 * 10-23 J/K

• T = temperature, in Kelvin's (absolute temperature)

• Noise is assumed to be independent of frequency

W/Hz k0 TN


Noise and Threshold

• RX Threshold = kT +10 Log (BW) + NF + SNR

– kT =10 Log(1.38e-23 * 293 ºK) = -204 dBW/Hz = -174 dBm/Hz into a 50 ohm

antenna

– k Boltsmann’s constant, 1.38e-23 W/Hz/ºK

– BW is the RX signal’s 3 dB bandwidth

• BW is computed differently for MC system

– Post processing SNR

• NF varies with Freq band, RF filter & cable losses (adds dB for dB)

– T room temp in Absolute term, (273+20) ºK

– NF increases due to rigid filter requirements for narrow FDD T-R spacing

– NF increases due to higher insertion loss in narrow band filters

– NF increases with temperature rise

• SNR requirements vary with Mod level, FEC power & modem design approach


Threshold & Interference

• N = ktB * NF, also known as noise floor in non-log terms

• I, Interference to cause 1 dB threshold degradation at 1e-6 BER

• I, dBm = N+I =N+(-6)

• SNR, dB = Signal / Noise

• SINR, dB = SNR + 1

• SIR = S/(N-6)

• SIR = SNR when S ≥ +6 dB

Unfaded RF RX Level

SNR

6 dB, objective for 1 dB

threshold degradation

T/I

N

Interference Level

1e-6

BER Threshold with Noise

Noise Floor

SINR

S

I

Thermal Static Fade Margin

SIR

T1e

-6 BER with I+N

Noise Floor + Interference

64QAM¾

o

o

4QAM-¾

4QAM-½

BPSK-½

1dB


RX Constellation Error

• Rx relative constellation error includes transmit constellation error, Rx constellation error plus the channel impairments

• Provides Tx-Rx and channel condition prior to error correction

• Tested under any of the normal operating conditions

OFDM

Burst Type

Relative Cons

Error for SS (dB)

Relative Cons

Error for BS (dB)

BPSK-1/2 ≤-13.0 ≤-13.0

4QAM-1/2 ≤-16.0 ≤-16.0

4QAM-3/4 ≤-18.5 ≤-18.5

16QAM-1/2 ≤-21.5 ≤-21.5

16QAM-3/4 ≤-25.0 ≤-25.0

64QAM-2/3 ≤-29.0 ≤-29.0

64QAM-3/4 ≤-30.0 ≤-31.0

OFDMA

Burst Type

Relative Cons

Error for SS (dB)

Relative Cons

Error for BS (dB)

4QAM-1/2 ≤-15.0 ≤-15.0

4QAM-3/4 ≤-18.5 ≤-18.5

16QAM-1/2 ≤-20.5 ≤-20.5

16QAM-3/4 ≤-24.0 ≤-24.0

64QAM-1/2 ≤-26.0 ≤-26.0

64QAM-2/3 ≤-28.0 ≤-28.0

64QAM-3/4 ≤-30.0 ≤-30.0


RX Threshold

• Rx to remain operational at signal up to -30 dBm for all modulations

• No damage to equipment at signal up to -0 dBm

– Requires higher IIP3 and IIP5 devices at RF front end

• Hi-RSL is more sensitive to higher modulation

• Rx must detect Rx signal up to -90 dBm min

• Rx dynamic range of 50 dB min

• PER to be better than 0.49%

• Image rejection to be 60 dB min

• Receive threshold is determined by the number of data subcarriers & frame length while excluding the pilots & preambles


RSSI, Receive Signal Strength Indicator

• Measurement referenced at RF Rx input

• Signal detection over a wide signal range (-10 to -90 dBm for 16d, -

40 to -90 dBm for 16e). Measurements to continue down to -123

dBm

• Tolerance accuracy over environmental conditions (within 2 dB

relative, 4 dB absolute)

• Controlling parameter for other critical functions

• Covers the full RF filter bandwidth

• Performs fast estimation to determine available BW, mod & FEC

allocations

• RSSI determined by excluding pilots & preambles

• Fixed power DL RSSI, adaptive power UL RSSI


Typical SNR vs. Modulation

• Ratio of the RX signal power to noise

power

• Required minimum SNR for 1e-6 BER

• Table shows typical SNR using CC &

RS FEC types vs. Mod level

– Further reduction with more powerful codes

• A key quality metrics for RX signal

• SNR requirements vary with Mod type,

FEC power & modem design techniques

– Minimum post processing requirement

– Lower requirements with higher powered FEC

– Lower requirements at lower modulation

Modulation SNRReq

BPSK-1/2 3.0

4QAM-3/4 6.0

4QAM-3/4 8.5

16QAM-1/2 11.5

16QAM-3/4 15.5

64QAM-2/3 19.0

64QAM-3/4 21.0


SNR

• S/W performs fast background

computation to determine

channel conditions and fade

margin

• Useful to check the presence of

steady interference at normal

RSL

• SNR value decreases (worse)

with increased path impairments

• Provides current channel status

conditions to optimize

transmission

16QAMError

Distance 'd'

For SC System


Phase Noise Effect

• Accumulated phase noise from all sources (Tx-BB, Rx-BB, TX-LO, Rx-LO)

• Phase noise effect appears different on OFDM system compares to a single

carrier system

-2 0 2

-3

-2

-1

0

1

2

3

In-phase Amplitude

Quadra

ture

Am

plit

ude

RX Const

-1 -0.5 0 0.5 1

-1

-0.5

0

0.5

1

In-phase Amplitude

Quadra

ture

Am

plit

ude

Scatter Plot

MulticarrierSingle carrier


Interference, (1)

• Spurious

– Caused by different combinations of signals in the Tx and Rx

– Harmonics are integer multiple of the primary transmitter/receiver frequency

• Predictable location and traceable back to primary frequency source

• Typically harmonics are measured up to 5x the frequency or up to 17.5 GHz

– Spurious signals are typically image frequencies caused by internal mixing of

an oscillator or clock freq with the primary transmitter/receiver frequency

• Difficult to trace due to change in level and mixing location

• Others types such as CIR, CINR, PCINR, ECINR, SINR, CCI, ACI, CW, ICI, ISI

• Easy to avoid/reject (with null carriers) narrowband interference with subchannels

– Less interfered part of the carrier can still be used

Freq

f1 f22f1-f22f2-f1

Rx Filter

IMD3

3f2-2f1

Interferer

IMD5

Δf

Δf


Interference, (2)

• Tx, Rx spurious interference to be ≤-47 dBm in 1 MHz BW

Power Rx Filter

FrequencyCCI

Out of channel

interference

ACI

Thermal Noise

Desired

Signal


CCI & ACI (Co & Adjacent Channel

Interference)

• Co-channel interference occurs when another transmission on the same carrier frequency affects the receiver

• Adjacent-channel interference occurs when energy from a carrier spills over into adjacent channels

• Standards specify a reference 1 & 3 dB degradation from interference within co-channel and from adjacent channel bandwidth

– Using same BW and type of signal

• Tests are performed with a same order modulated signal and bandwidth

• Degradation referred to 1e-6 BER

• Interference tolerance to determine from inversed T/I curve -25

25

15

20

10

5

3016QAM

64QAM

Adj-Ch

Sensitivity

Le

ve

l a

bo

ve

se

nsitiv

ity (

dB

)

2x-Adjacent

Channel


Interference

• CCI and ACI requirements at 1/3 dB degradation

• Reference 1e-6 BERModulation

CCI (dB)

1 dB deg

x1 ACI (dB)

3 dB deg

x2 ACI (dB)

3 dB deg

BPSK-1/2 4.0 -11.0 -30.0

4QAM-1/2 8.0 -11.0 -30.0

4QAM-3/4 9.5 -11.0 -30.0

16QAM-1/2 12.5 -11.0 -30.0

16QAM-3/4 16.0 -11.0 -30.0

64QAM-2/3 19.0 -4.0 -23.0

64QAM-3/4 22.0 -4.0 -23.0


CW (Continuous Wave) Interference

• Narrow out of band signal

may affect the normal Rx

AGC operation

• It may exceed the device

maximum overload

tolerance

• Affects RSSI detector

accuracy

• Mixing products may fall in-

band

• Tolerable limit is specified

by ETSI standardFreq

f1 f22f1-f22f2-f1

Rx Filter

IMD3

3f2-2f1

Interferer

IMD5

Δf

Δf


ISI, ICI, LO Phase Noise and Clock Offset

• ICI & ISI are normally caused by phase

noise, poor synchronization, unstable

subcarriers clock, insufficient delay

spread and Doppler shift

• ICI & ISI become more sensitive at higher

modulation

• To avoid inter-carrier interference, the

inter-carrier spacing is set to be equal to

the inverse of the symbol duration.

• Non-linear distortion and phase noise are

the two largest contributing factors to a

loss of orthogonality, creating an ICI.

Poor frequency estimation in the receiver

is another contributing factor

• ISI introduces an irreducible error floor

which can not be removed by increasing

transmit power


ISI & ICI

• Signal arriving late from secondary path

– Affects both ISI & ICI

(i-2)

OFDM symbol

(i-1) (i+2)(i+1)(i+0)

OFDM symbolOFDM symbolOFDM symbolOFDM symbol

OFDM symbol

OFDM symbol

t

|Ca(t)|

Magnitude of channel

impulse response

Fade in (ICI)

Fade out (ISI)


ISI, Inter Symbol Interference

• Delay spread: defined as the RMS time difference between the

arrival of the first and the last multipath signal seen by the receiver

– The delay is affected by distance, frequency and environment

– For mobile, it is also dependent on the speed (Doppler shift)

• Typical delay spread: 40 to 200 ns for indoors (50 ns in homes, 100

ns in offices, 300 ns in industrial environment), 1 to 20 us for

outdoors

• ISI becomes more serious as the bit rate increases (σ/Ts gets worse

i.e., bigger). σ is delay spread, Ts sample time

• For OFDM ratio of σ /NTs to become smaller (better ISI)

• Sensitive to LO phase noise (from all sources)

• For MC system, ISI is less sensitive to narrowband interfering signal

and frequency selective fading


ICI, Inter Carrier Interference

• OFDM systems becomes more susceptible to time-variations as symbol length

increases

– Increase the CP length and number of pilot tones to mitigate the ICI

– Lower FFT size increases the subcarrier spacing that improves the ICI and

more tolerant to Doppler shift

• Time variations introduce ICI in frequency domain

• Signal arriving from multipath causes ICI. If occurred, it is not correctable

• LO phase noise and clock recovery error produces wider overlapping skirt at the

lower part of the subcarriers in frequency domain. This phenomena is independent

of clock stability

F0F-2 F2F1F-1

ICI


Dynamic Range

• Receive signal ratio between the maximum possible signal and the

minimum signal that gives the desired signal level over noise at

demodulator input

• The range includes input power (signal, noise and interference) over

which receiver performs adequately

• Performance determined at a reference 1e-6 BER

• Determined by aggregate AGC in the receiver chain

-RSSI

BER

1e-2

1e-4

1e-6

1e-10

ThresOverload


TDD, Time-Division Duplex

• DL & UL timeshare the same RF channel

– With a gap period at transition to accommodate Tx/Rx mode switching and

PA settling time

• BS or SS, neither transmit/receive simultaneously

• On DL, SS is associated with a specific burst

• On UL, SS is allotted a variable length time slot for their usage

• Single RF filter and single RF-LO

• Less stringent filter requirements

• Less data throughput

• Increased MAC control complexity (less hardware complexity)

• Readily available lower cost parts due to higher usage in unlicensed band

• Dynamic asymmetry ratio of DL/UL

• Unlicensed operation is limited to using TDD format


FDD and TDD Frame Structure

RNG BW

Contention

UL

Burst #kPUL

Burst #1P

UL SS #1

UL Subframe (PHY PDU)

UL SS #k

P FCH DL Burst #m...

DL Burst #1

DL Subframe (PHY PDU)

...

P FCH DL Burst #m...DL Burst #1

DL Subframe (PHY PDU)

Frame n-1 Frame n Frame n+2Frame n+1

RNG BW

Contention

UL

Burst #kP

UL

Burst #1P

UL SS #1

UL Subframe (PHY PDU)

UL SS #k

...TDD

FDD

FCH: Frame Control Header

P: PreambleRNG: Contention Slot for Ranging Request

BW: Contention Slot for BW Request

TTG: Tx/Rx Transmission Gap

RTG: Rx/Tx Transmission Gap

TTG RTG

Time

Fre

qu

en

cy


TDD & FDD, HFDD

TDD

FDD

HFDD


FDD, Frequency-Division Duplex

• DL & UL on separate RF frequency channels

• BS & SS transmit/receive simultaneously

• Static asymmetry

• Half-duplex SSs supported

–SS does not transmit/receive (lower cost)

• Continuous operation, no switch settling time required

• Requires two frequency channels

• Higher performance front end RF

• Simpler MAC control operation


TDD & FDD

• TDD

– Advantages

• Asymmetric DL/UL ratio

• Lower cost of RF elements

• More options with channels

size

• Simple AAS in MIMO

implementation

– Disadvantages

• Vulnerable to interference

• Synchronization of receiver

• Synchronization of network

• FDD

– Advantages

• Better protection against

interferences (separate DL/UL

ratio)

• Stronger synchronization of

receiver

• Network planning is easier

– Disadvantages

• Fixed DL/UL ratio

• More expensive

• Less options with channel sizes


WiMAX

• WiMAX architecture consists of two key

items

– PHY

• BB & RF processor (Frequency

source, Mod, IFFT/FFT, timing

recovery, Sync, multiple interface

access, error detection &

correction etc.)

– MAC

• Standard compliant LAN and end

to end interface

• Protocol control / process /

manage and QoS toward LAN

interface

• Protocol control / process /

manage and QoS toward end-to-

end system


Base Station & Subscriber Station

• Base Station (BS):

– Controls the entire system, frame size, scheduling, admission control, QoS, Ranging, clock synchronization, power control, handoff, privacy key and PHY management

– All traffic goes through BS

• Subscriber Station (SS):

– Finds BS, acquire PHY synchronization, obtain MAC parameters, generate bandwidth requests, make local scheduling decisions, follows transmission/reception schedule from BS, performs initial ranging, maintenance ranging and power control

• Mobile Station (MS):

– In addition to the SS functions, mobility management, handoff, power conversion and power management


System Design Requirements

• BW, Bandwidth

• Bit rate

• Subcarrier spacing

• Tolerable delay spread

• Doppler shift value

• Bits per OFDM sym = Bits rate * (active data subcarriers) * FEC

* Log2(Mod)

• QoS

Source

Data

Inter

leaving

Channel

Coding S/P

Modulation (M)

Modulation (M)

Modulation (M)

Modulation (M)

IFFT

Mapping

from

size

N/2 to N

S/PTo

ChannelCP


Key System Design Parameters

• Channel bandwidth

• Number of subcarriers

• CP, cyclic prefix

• Subcarrier spacing

• Modulation

• FEC

• PO

• Dynamic range

• Threshold


OFDM Subcarriers

• An OFDMA symbol consists of

three types of subcarriers:

– Pilot subcarriers

– Data subcarriers

– Null subcarriers

• Subcarriers can be turned

on/off dynamically based on

channel conditions and to meet

the required BW

Im

Received symbolReceived symbol

Transmitted symbol

Real


Pilot Sub-carrier

• Pilot subcarriers contain signal values that are known to the receiver

– Facilitate signal recovery and synchronization

• Pilot subcarriers are used in the receiver for correcting the magnitude (important in QAM) and phase shift offsets of the received symbols (see signal constellation example on previous page)

– Magnitude and phase of these subcarriers are known to receiver that helps to speed up channel estimation

• Always BPSK-1/2 modulated & its transmission repeated

• Higher power level (2.5 dB higher than the average power of the non-boosted data tones

• Transmitted with embedded Pseudo random code

• Inserted after the FEC stage so as not to destroy the fixed time and amplitude relationships that these signals must possess to be effective

• 8 pilots for OFDM (Configurable number for each transmitter in OFDMA)

• More pilots increases noise resiliance & processor loading while reducing the overall throughput

• For OFDM, pilots are common to all UL-subchannels

• For OFDMA, certain numbers are dedicated to specific subchannels


Data Sub-carriers

• Used to transport over head control and user data

• Part of the DL/UL data subcarriers contain preamble symbols for training purposes

– DL has two long preamble symbols of QPSK: Two training cycles at the start of

each 8 us (1st containing 50 subcarriers and called short training sequence,

every 4th subcarrier with a phase relationship that minimize the PAPR. This

period is used for RX gain setting and course frequency correction. All have the

same levels. 2nd containing 100 subcarriers and called long, 8 us, all

subcarriers turned on. Allows RX to calculate frequency response of the

channel and to fine tune the frequency errors). Preambles are 3 dB stronger

than all other symbols in the DL frame.

– UL always starts with preamble (called short preamble, 100 subcarriers of

QPSK. Preamble has no pilot carriers. Helps Rx to sync and perform additional

channel estimation). Modulation remains the same within burst but changes

from burst to burst.

• Following DL Preamble is the FCH (single symbol of 88 bits, BPSK-1/2 for OFDM,

QPSK-1/2 for OFDMA), occupies 1st two subcarriers in the 1st data symbol

• Remainder of the data subcarriers carry the user data


Null Sub-carriers

• To avoid difficulties in DAC and ADC converter offsets,

and to avoid DC offset and PA saturation, the sub-carrier

falling at DC is not used

– Relaxes anti-aliasing and filtering requirements

– DC subcarrier power must be at least 15 dB lower than the

average of all other subcarriers

• Provides a frequency guard band before the Nyquist

frequency and allows for a realistic roll off in the analog

anti-aliasing reconstruction filters

• Used for spectrum shaping and to fit the regulatory mask

• Null subcarriers contain no power


Typical OFDMA Blockdiagram

• Variable FFT & subcarriers size based on the BW

• Constant subcarrier frequency spacing

• Configurable over-sampling factor

S/P IFFTQAM

ModP/S

SCRM

&

FEC

Interl

eaverPilot RFTX

CP

&

Win

DAC

&

FLTR

S/PFFTQAM

DMDFD

EQL

CP

RemoveP/S RFRX

ADC

&

FLTR

TMG

&

FreqSyn

FEC

&

DScrm

D-Int


Scrambler

• Randomization prevents long sequences of 1‟s or 0‟s in the incoming data stream

– Helps speed up and maintain clock recovery

• DL & UL data is randomized by modulo-2 addition of every data bit with output of a

pseudo random binary sequence generator

• Randomization is performed on data bits only

• A pseudo random binary sequence of 1+X14+X15

• Each frame starts with initialization sequence of 100101010000000

• Randomization is performed on each allocation (DL or UL) independently


FEC (1)

• Probability of symbol location after

passing through AWGN channel

• When does error occur?

– Expected symbol ends up in

neighbor’s territory

• Symbol vs. bit error

– Error multiplication at higher order

modulation

Receive symbol position

Pro

ba

bili

ty D

en

sity

Probability Density Function


FEC (2)

• Hard decision declares error the moment it crosses the decision boundary

• Soft decision further adds statistical values in error computation

• Error distance decreases on higher modulation making it more susceptible to error

– BPSK = 2.0

– QPSK = 1.414

– 16QAM = 0.471

– 64QAM = 0.283

– 256QAM = 0.202

• Signal compression at outer most

• SER, BER. PER, FER

1/4th constellation View

Modulation Error

BPSK 90°

QPSK 45°

16QAM 16.9°

64QAM 7.7°

4QAM

97531 1311 15

9

7

5

3

1

13

11

15

825026102 170122 226

250218194178170 338290 394

202170146130122 290242 346

1621301069082 250202 306

13098745850 218170 274

10674503426 194146 250

9058341810 178130 234

306274250234226 394346 450

128QAM

64QAM

32QAM

16QAM

Error

Distance 'd'

256QAM


FEC (3)

• Process computes and adds additional parity bits at the transmit

which helps identify error location & possible correction by receiver

• FEC implementation varies by Cap/BW/Modulation

– Reed Solomon (RS) only

– RS + Convolution

– RS + Convolution + Interleaver

• Detected error quality is used to control adaptive modulation, coding

rate, data integrity, error performance, bandwidth allocation,

subchannelization, AAS (adaptive antenna system) & MIMO

calculations

• If the last FEC block is not filled, that block may be left shortened

• Shortening in both UL and DL is controlled by the BS and is implicitly

communicated in the UL-MAP and DL-MAP


FEC (4)

• Adds redundancy to data bits

• Programmable concatenated Reed Solomon (good for low BER,

≤1e-8) and Convolution coding (good for 1e-4 to 1e-7 BER)

• Total of 7 different rate dependent combinations

• Support of Block Turbo Coding (BTC), Convolutional Turbo Coding

(CTC) and low density parity coding (LDPC) is optional

• The Reed Solomon encoding shall be derived from a systematic

varied length RS code (k, n, t) where n is the number of overall bytes

after encoding, k is the number of data bytes before encoding, t is

the number of data bytes which can be corrected

• Encoder supports shortened and punctured codes to accommodate

variable block size

• Reduces overall throughput according to the selected coding rate


FEC (5)

• Incoming data bytes are processed serially (byte by byte) over a

fixed RS block length then adds the parity bytes at the end

• Low rate coding may be punctured by deleting zeros to lower

overhead (i.e., deleting 2 out of 6 bits of ½ to create a ¾ rate)

Data In

X Out

Y Out

1 bit

delay

+

1 bit

delay

1 bit

delay

1 bit

delay

1 bit

delay

1 bit

delay

+

Raw Data (lower speed)

Block 2 Block 1 Block 0

After RS (higher speed)

X X X X X X

Parity Bytes

Data Bytes Direction of Data Flow


FEC (6)

• Four FEC schemes defined in 802.16

• 802.16 defines concatenated coding schemes: inner code

(random errors) and outer code (burst errors)

• Code type 1 (used for large data block or high coding

requirements):

– No inner code

– Outer codes: systematic Reed-Solomon (corrects errors: 16 to 0

bytes)

– Two modes of operation:

• Fixed codeword: number of information bytes same for every RS

codeword

• Shortened codeword: number of information bytes in the final RS

block is reduced


FEC (7)

• Code type 2 (useful for low to moderate coding rates that

provide good performance):

– Outer code: almost same as RS code as in code type 1

– Inner code is a (24, 16) block convolutional code (BCC)

• 16 bits input block code, bi

• 24 bits output codeword ci (each symbol: combination of others

symbol: c23 = b15+b0+b1)

• Code type 3 (optional):

– Outer code: almost same as RS code as in code type 1 and 2

– Inner code: (9, 8) parity check code (code adds one parity bit to

every eight bits)


FEC (8)

• Code type 4 (used to extend the range of a BS or

increase the data rate at the same range):

– No inner code

– Outer code: block turbo code (BTC): The idea is to encode the

data twice

– Option: bit interleaving

k1 n1

n2

k2 Information bits Checks

Checks on checksChecks


BER Curve

• BER vs. Eb/No before and after the RS only FEC

• Performance tradeoffs

FEC Gain


Modulation & Coding Combination Summary

• WiMAX supports 7 possible Mod / FEC rates to provide

optimal data throughput

• Other modulations and FEC types are optional

Modulation

Uncoded

Blocks (bytes) RS Code CC Code

Coded Blocks

(bytes)

Overall

Coding

BPSK 12 (12, 12, 0) 1/2 24 1/2

4-QAM 24 (32, 24, 4) 2/3 48 1/2

4-QAM 36 (40, 36, 2) 5/6 48 3/4

16-QAM 48 (64, 48, 8) 2/3 96 1/2

16-QAM 72 (80, 72, 4) 5/6 96 3/4

64-QAM 96 (108, 96, 6) 3/4 144 2/3

64-QAM 108 (120, 108, 6) 5/6 144 3/4


Preamble (added after the FEC)

• It is important that the frame control section of the DL frame be

encoded with fixed set of parameters known to the SS at

initialization in order to ensure that all subscribers stations can

read the information

• The control portion of the frame is encoded with a Type 2 FEC

where the outer code is a (46, 26) RS code and the inner code

is a (24, 16) BCC

S/P IFFTQAM

ModP/S

SCRM

&

FEC

Interl

eaverPilot RFTX

CP

&

Win

DAC

&

FLTR

S/PFFTQAM

DMDFD

EQL

CP

RemoveP/S RFRX

ADC

&

FLTR

TMG

&

FreqSyn

FEC

&

DScrm

D-Int


Interleaving

• Data interleaving is very affective against burst (clustered) typed errors

• Process increases latency through the system

• All encoded data bits shall be interleaved by a block interleaver

• Interleaver block size corresponds to the number of coded bits per specified allocations

• The number of coded bits per carrier is 2, 4 or 6 for QPSK, 16QAM or 64QAM, respectively

Before Interleaving

X X X X X

No error block Burst errored block No error block

After Interleaving

X X X X X

Correctable block Correctable block Correctable block

Data Input b1 b4 b7 b10 Data Out

Fill in b2 b5 b8 b11 b1 b4 b7 b10 b2 b5 b8 b11 b3 b6 b9 b12

b3 b6 b9 b12


Constellation Mapping

• Serial data bits are mapped into selected modulated

symbol

• Supports Gray-mapped BPSK, 4/ 16/ 64-QAM modulation

• Support of 256-QAM is optional

• Normalized to achieve unity average power regardless of

modulation scheme

• Constellations must be normalized to achieve equal

average power

• Supports adaptive modulation and coding


Constellation Map and Spectral Efficiency

Q = b3 b4 b5

. . b5 b4 b3 b2 b1 b0

6 bits I = b0 b1 b2

1 Symbol

7

5

-3

-1

110 000010 000011 000001 000000 000 101 000111 000 100 000

110 101010 101011 101001 101000 101 101 101111 101 100 101

110 111010 111011 111001 111000 111 101 111111 111 100 111

110 110010 110011 110001 110000 110 101 110111 110 100 110

110 010010 010011 010001 010000010 101 010111 010 100 010

110 011010 011011 011001 011000 011 101 011111 011 100 011

110 001010 001011 001001 001000 001 101 001111 001 100 001

110 100010 100011 100001 100000 100 101 100111 100 100 100

b0b1b2 b3b4b5Imaginay

Real

3

1

-5

-7

-7 -5 -3 -1 7531

Input Bits

(b0b1b2)I-Out

Input bits

(b3b4b5)Q-Out

000 -7 000 -7

001 -5 001 -5

011 -3 011 -3

010 -1 010 -1

110 1 110 1

111 3 111 3

101 5 101 5100 7 100 7

Modulation

Spectral

Efficiency

BPSK 1

QPSK 2

16QAM 4

64QAM 6


Constellation Mapping (2)

• Digital modulation: how bits are mapped to symbols.

Constellation can be selected per subscriber (quality

of the RF channel)

• In the DL: QPSK, 16-QAM and 64-QAM

• Distance to origin: power that sends the signal,

follows 2 adjustment rules

–Constant constellation peak power and constant

constellation mean power

–Before Mod I & Q signals filtered by square root raised

cosine pulse shaping filter:

• S(t) = I(t)*cos (2πfct) * Q(t)sin(2 π fct)


Constellation Mapping (3)

• After bit interleaving, the data bits are entered serially to the

constellation mapper

• Mapped to form symbol for selected modulation

• Modulation with Gray coding

• Normalized to achieve unity average power regardless of modulation

scheme in order to facilitate timing recovery

S/P IFFTQAM

ModP/S

SCRM

&

FEC

Interl

eaverPilot RFTX

CP

&

Win

DAC

&

FLTR

S/PFFTQAM

DMDFD

EQL

CP

RemoveP/S RFRX

ADC

&

FLTR

TMG

&

FreqSyn

FEC

&

DScrm

D-Int


Constellation Display

• Multiple modulations captured in a single frame

• Constant average power to stabilize Rx AGC loop gain


Constellation Map

• The Frequency Domain description includes the basic

structure of an OFDM symbol

• An OFDM symbol is made up from subcarriers, the number of

which determines the FFT size used. There are several

carrier types

– Data subcarriers: for data transmission & down stream synchronization

– Pilot subcarriers: for various estimation purposes

– Null subcarriers: no transmission at all, for guard bands and DC carrier


IFFT/FFT

• A DSP process that uses N points IFFT of a signal X(k)

• Parallel data streams are used as inputs to an IFFT

• IFFT output contains N times data buckets

– Each bucket contains sum of many samples of many sinusoids

– Same frequency, different amplitude and phase

– At center of the subcarrier there is no cross talks from other subcarriers and

hence makes receiver to correctly recover data

• IFFT does modulation and multiplexing in one step

• Normal DFT would require (N-1)^2 operation whereas the FFT would require only

N/2*Log2(N) operations (i.e., 65025 vs. 1024 multiplications for 256 point FFT)


IFFT/FFT

• The IFFT operation in OFDM partitions a wide band channel into multiple

narrowband subchannels

• The IFFT & FFT operations are almost identical. The IFFT can be made using an

FFT by conjugating input and output of the FFT and dividing the output by the FFT

size. May use the same hardware for Tx & Rx in TDD mode

• IFFT modulates and multiplexes the signal in one step

• DSP algorithms replace a required bank of IQ Mod-DMD that would otherwise be

required

Input

Data

Output

Base

Band

OFDM

Signal

I

Q

Frequency Domain Time Domain

Symbol StartGuard Period

Zeros

xxx

IFFT

Pa

ralle

l to S

eria

l

xxx

Subcarrier

Modulation

Data

IQ Vector

01

199


Guard Time

• Signal from multiple reflected paths arrive at various delays

• Delayed signal may corrupts the front part of the next symbol

• The guard time acts as a buffer to allow time for multipath signals from previous

symbol to die away before the information from the current symbol to get collected

by receiver

• It is like water splash when driving too close to a car in front

• A simple gap is not acceptable for optimal signal recovery at Rx

• Adding guard time lowers the symbol rate but does not affects the subcarrier

spacing Environment Sig. Delay, ns

Office/home NLoS 50

Open space office NLoS 100

Large open space office NLoS 150

Manufacturing area 200-300

Microcell 500

Large open space LoS 140

Large open space NLoS 250

Mobile city 2500

Mobile rural area 25000


CP, Cyclic Prefix Plot

DC Carrier

Guard Band Low (28) Guard Band High (27)

Pilot Carrier (8)Data Carrier (192)

TgTb = FFT symbol duration

Ts = OFDM symbol duration

CP x(0), x(1), ..., x(N-2-v), x(N-1-v) ,...,x(N-1)

Sampling startξmax

Frequency Domain

CP vs. Guard Band

Time Domain

Sample

Tg

Ts=Tb+Tg

NFFT*1/FS=Tb(=1/Δf) Tg=G-Tb

NFFT

NFFT-1

Time

1/FsLevel

~

~

SC Sym periodEqualiazer Length

MC


Delayed Signal

• Signal travels through various paths and ultimately arrives at

different time

• FFT symbol portion must contain integer number of cycles

• Append tail part of the FFT symbol to its front part in order to make it

a continuous signal

– ICI & timing recovery issues if not appended

• Guard interval length may or may not contain integer cycles


Guard Time Consideration

• Guard interval reduces the signal energy available at receiver

• Guard interval reduces the data rate throughput while increasing the noise bandwidth (spectrally inefficient)

• Adding a guard interval lowers the symbol rate, however it does not affects the subcarrier spacing see by the receiver

– Subcarrier spacing Δf = Fs / FFFT

• Guard interval simplifies equalization at the Rx if guard interval time is greater than the maximum delay spread

• Guard interval should be short (performance trade offs)

• Guard interval should be chosen longer than the actual RMS delay spread, 3x to 4x longer (≈ 0.1 of symbol length, SNR ≤ 1 dB = -10log(1- Tg/TOFDM Sym))

• Guard interval is discarded by the receiver

– SNR Loss, in dB = -10Log (1 - Tguard interval length / TOFDM symbol duration)


Guard Time Effects

• Insufficient guard time (CP) causes ISI & ICI

• Higher order modulations & timing recovery circuits are more

sensitive to ICI & ISI

• 16QAM 256 points FFT receive constellation plots (a, delay ≤

guard time. b delay exceeds guard time by 3% of FFT

internal. c, delay exceeds guard time by 10% of the FFT

interval)

-2 0 2

-3

-2

-1

0

1

2

3

In-phase Amplitude

Quadra

ture

Am

plit

ude

RX Const

-2 0 2

-3

-2

-1

0

1

2

3

In-phase Amplitude

Quadra

ture

Am

plit

ude

RX Const

0 5 10 15

-40

-35

-30

-25

-20

-15

-10

Frequency (MHz)

Magnitude-s

quare

d,

dB

-2 0 2

-3

-2

-1

0

1

2

3

In-phase Amplitude

Quadra

ture

Am

plit

ude

RX Consta b c


CP, Cyclic Prefix

• Usage of CP is necessary to combat MP distortions

• CP reduces the BW efficiency (a tradeoff between throughput performance vs. BW)

• CP should be longer than the maximum expected RMS delay spread

• Programmable (1/4, 1/8, 1/16, 1/32), ¼ is the most robust in the multipath

• Delayed replicas of the OFDM symbol always have an integer number of cycles within FFT interval

• A copy of the last OFDM symbol is appended to the front of transmitted OFDM symbol

• Actually the Tg can be realized by adding zeros, but using the CP as guard interval can transform the linear convolution with the channel into circular convolution

• CP is added after the IFFT on a combined signal rather than for each sub-carrier

• Accommodates the decaying transient of the previous symbol

• Smooth initial transient to reach the current symbol

• Impact of CP is similar to the roll-off factor in raised cosine filtered SC systems


Doppler Shift

• A measure of spectral broadening caused by the channel time variation

• Motion of the mobile causes periodic phase shifts which change with time. The rate of change of phase gives rise to Doppler frequency, which varies with mobile speed and arrival angle of rays

• fd, Hz = v/λ, where v velocity, λ wavelength

– Inter-carrier spacing must be at least 10 times higher than the maximum fd

– Value increases if moving toward source and lower when moving away from each other

• Symbol rate must be much higher than the Doppler shift. Inter-carrier spacing of the system must be chosen large, compared to the maximal Doppler frequency of the fading channel.

• Coherence time, Tc = 0.423/fd , fd Doppler shift

– TC >> T, slow fading

– TC ≤ T, fast fading

– Coherence time is a time period to correlate the channel’s value

• Coherence distance, DC = 0.179 λ, to determine antenna spacing


Group Delay

• It is defined as derivative of the phase response versus

frequency, that is the slope of the phase response

• Prime contribution of the group delay comes from tighter

band pass filter response in the baseband, IF & RF sections.

It is also contributed from improper cable termination, non-

compensated sin(x)/x, antenna mismatch, signal combiners

and multipath effects

• MC systems are more tolerant to group delays

• Equalizer is an effective tool to remove linear distortion


OFDM Symbol

• Chosen NFFT = 2^n

• Carrier filler for unused carriers

• Total OFDM Symbol time, Ts = (1/subcarrier frequency

spacing) + Tg

• Tb: Useful symbol time, N/Fs

• Tg or CP (to improve the impulse response) guard time

–Generally kept under 1 dB (pick about 4x the delay spread)

– Increased roll off time reduces the spread tolerance

• G = Tg/Ts, four programmable intervals (1/4, 1/8, 1/16,

1/32)


Windowing

• Makes amplitude go down smoothly to zero at symbol boundaries (minimizes

interference to others)

• Tx signal without windowing will have wide bandwidth due to the side lobes of the

IFFT being a Sinc function

• Tx signal is band limited in time domain by using windowing technique (raised

cosine function). There is no band limits in frequency domain

• Applied window must not influence the signal in its effective period. In other words

pulse-shaping affects on the CPTguard TFFT

Twin

T

Twin

Prefix PostfixEffective TX-time Time

TFFT

Sym+1Current SymbolSym-1


TX PO

• BS to provide 10 dB min settable attenuation in 1 dB min step size with better than 1.5 dB (within 0.5 dB for relative steps) measurement accuracy

• SS to provide 30 dB (40 dB for 16e) min settable transmit attenuation in 1 dB min step size

– 50 dB min settable attenuation for devices with sub-channelization

– The measurement accuracy for 1 dB step size must be within 1.5 dB (within 0.5 dB for relative step) for the first 30 dB range, 3 dB for larger step size

• Preamble level between adjacent sub-carriers must be within 0.1 dB

• Preamble bursts are 3 dB (4.6 dB for 16e) higher than the FCH & DL data

• Training symbol (for sync., estimation and tracking time-varying channels) are transmitted via preamble or pilot carriers

• UL pilots are not boosted


BS, TX Performance Requirements

• PO +38.5 dBm/MHz max or per local regulatory requirements

• Tx output noise spectral density of -80 dBm/MHz max when Tx is not transmitting

• Ramp up & down time of ≤ 8 sym

• Mod accuracy

– 12% for QPSK, 6% for 16QAM without equalizer

– 10% for QPSK, 3% for 16QAM, 1.5% for 64QAM with equalizer, linear

distortion removed

• Symbol timing accuracy

– ≤ 0.02 pk-pk of nominal sym relative to previous sym over 2s duration

– Tx Sym clock accuracy to be within 1e-6

• Tx burst timing step size 0.25 of a sym

• Tx burst timing step accuracy 0.125 of a sym

• SNDR ≤ -31 dBc

• Tx spectral mask regulated per local regulatory agency


TX Waveform Accuracy

• The accuracy of the modulated waveform is affected internally by

– Root raised cosine filter length and coefficients accuracy

– D/A converter accuracy

– Modulator imbalances

– Synthesizer phase noise

– PA nonlinearities

• Externally affected by

– Cable mismatch

– Antenna mismatch

– Interference

– Terrain

– Environmental

Magnitude

error

Error Vector

Magnitude

Actual

Ideal

Phase Error

Carrier Leakage

Error

I

Q


BS, min TX Performance

• Provides TX Mod accuracy and PA

linearity conditions

• Tx relative constellation error is to

measure with ideal receiver with carrier

recovery loop BW of 1% of the symbol

rate

• Measurement method determines the

magnitude error of each constellation

point at the sampling instances and

RMS averages them together across

multiple symbols, frame, and packets

• Provides signal quality prior to channel

impairments at the sampling instances

OFDM

Burst Type

Relative Cons

Error for SS (dB)

Relative Cons

Error for BS (dB)

BPSK-1/2 ≤-13.0 ≤-13.0

4QAM-1/2 ≤-16.0 ≤-16.0

4QAM-3/4 ≤-18.5 ≤-18.5

16QAM-1/2 ≤-21.5 ≤-21.5

16QAM-3/4 ≤-25.0 ≤-25.0

64QAM-2/3 ≤-29.0 ≤-29.0

64QAM-3/4 ≤-30.0 ≤-31.0

OFDMA

Burst Type

Relative Cons

Error for SS (dB)

Relative Cons

Error for BS (dB)

4QAM-1/2 ≤-15.0 ≤-15.0

4QAM-3/4 ≤-18.5 ≤-18.5

16QAM-1/2 ≤-20.5 ≤-20.5

16QAM-3/4 ≤-24.0 ≤-24.0

64QAM-1/2 ≤-26.0 ≤-26.0

64QAM-2/3 ≤-28.0 ≤-28.0

64QAM-3/4 ≤-30.0 ≤-30.0


SS/MS, min TX Perf

• ≤ +39.5 dBm/MHz or per regulatory requirements

• BER at 1e-6 & carrier symbol rate of R in Mbps

– -90dBm+10Log(R) for QPSK

– -83dBm+10Log(R) for 16QAM

– -74dBm+10Log(R) for 64QPSK

• Transmission time from Tx to Rx, 2 us for TDD, 20 us for FDD &

HD-FD

• ACI at 1e-6 BER & 1 dB degradation

– -1 dB for QPSK, +6 dB for 16QAM, +13 dB for 64QAM

• 2nd ACI at 1e-6 BER & 1 dB degradation

– -30 dB for QPSK, -30 dB for 16QAM, -23 dB for 64QAM


SS/MS, min TX Perf-2

• 40 dB min dynamic range

• Tx PO of +15 dBm min for QPSK

• Tx PO adjustment in 0.5 dB step

• Tx pk-pk jitter, ≤ 0.02 of the symbol duration in over 2 s period

• Symbol clock to be locked on BS

• TX burst timing accuracy, self correction for burst step up to 0.5 of a symbol with step accuracy of 0.25 of symbol

• TX RF frequency accuracy, 1 ppm

• Spectral mask per local regulatory requirements

• Ramp up and down time ≤ 8 symbols

• Noise density, -80 dBm/MHz when not transmitting

• Modulation accuracy: 10% (QPSK), 3% (16QAM) 1.5% (64QAM) with equalizer distortion removed


SS/MS, min TX Perf-3 (Flatness)

• DC subcarrier to be suppressed by 15 dB min

relative to the total average power from all data and

pilot subcarriers

• The outer subcarriers need to be within +2/-4 dB

from average power transmitted from all active

subcarriers

–The inner subcarriers must be within 2 dB

–The adjacent subcarriers must be within 0.4 dB


Po, Attenuation & Accuracy

• Preamble power level may change from burst to burst

• Preamble aids in synchronizing the RX, perform channel estimation & Equalization processes

• Preamble spectral flatness is specified across all sub-carriers

– Po within 0.1 dB of adjacent subcarriers for both DL/UL

– Preamble applies to every second or 4th channel (computational adjacent)

– 2 dB ave over all active tones from -50 to -1 & +1 to +50, +2/-4 dB ave over all active tones from -100 to -50 and +50 to 100 for both DL/UL

– Preamble symbol contains no pilot

– 3 dB higher power than all other data subcarriers in the DL subframe

– Requires extremely sharp notch filters for reliable measurement.

• Requires 10 dB min range for BS. 30 dB min range for OFDM SS. 50 dB min range for OFDMA SS.

• 1 dB step with 1.5 dB min relative accuracy for 30 dB, Larger steps with 3 dB min relative accuracy for over 30-50 dB for MS/SS

• Power control to support 30 dB/s signal fluctuations


BS performance

• TX center freq tolerance

–BS 1 ppm, SS must be locked to BS, SS to be within 1

ppm of BS

• Tx symbol clock frequency tolerance

• Rx freq & timing requirement

• Time accuracy

–5 to 25 us for TDD

–N/A for FDD

–GPS option (more expensive and difficult to access open

sky if in the basement)


MS Performance Parameters

• RSL power determined from per subcarrier level

• RSL per unboosted subcarrier = RSSI-10Log(8)-

10Log(number of preamble subcarriers)

• Fast feedback channel from MS to up date the time critical

information such as CINR, MIMO, AAS, spatial multiplexing,

etc.


RX Constellation Error

• Relative constellation error

• Intended to ensure that RX SNR does not degrade more than 0.5 dB due to TX

SNR. Measured by an ideal receiver with carrier recovery loop bandwidth of 1% of

the symbol rate

• EVM value includes PA nonlinearities, untracked phase noise, inband amplitude

ripple and DAC inaccuracies

• Results are independent of the FEC

OFDM

Burst Type

Relative Cons

Error for SS (dB)

Relative Cons

Error for BS (dB)

BPSK-1/2 ≤-13.0 ≤-13.0

4QAM-1/2 ≤-16.0 ≤-16.0

4QAM-3/4 ≤-18.5 ≤-18.5

16QAM-1/2 ≤-21.5 ≤-21.5

16QAM-3/4 ≤-25.0 ≤-25.0

64QAM-2/3 ≤-29.0 ≤-29.0

64QAM-3/4 ≤-30.0 ≤-31.0

OFDMA

Burst Type

Relative Cons

Error for SS (dB)

Relative Cons

Error for BS (dB)

4QAM-1/2 ≤-15.0 ≤-15.0

4QAM-3/4 ≤-18.5 ≤-18.5

16QAM-1/2 ≤-20.5 ≤-20.5

16QAM-3/4 ≤-24.0 ≤-24.0

64QAM-1/2 ≤-26.0 ≤-26.0

64QAM-2/3 ≤-28.0 ≤-28.0

64QAM-3/4 ≤-30.0 ≤-30.0

Magnitude

error

Error Vector

Magnitude

Actual

Ideal

Phase Error

Carrier Leakage

Error

I

Q


Adaptive Frequency Domain Equalizer

• A perfectly normal signal becomes distorted after going through multipath channel

• Same frequency but different amplitude & phased signals are received

• Equalizer estimates the error amplitude

• Multiply the subcarrier by inverse phase and magnitude of the estimated channel

• Time domain equalization in a time dispersive channel becomes prohibitively

expensive for a SC system as data rate increases. For MC, each subchannel can

be modeled as flat fading channel requiring simple (N-short) Frequency Domain

EQ. FDE is an attractive alternative to mitigate complexity.

• OFDM moves the IFFT operation to Tx to load balance complexity between Tx

and Rx

• An adaptive process for varied conditions

• Removes only the linear distortions

S/PFFTQAM

DMDFD

EQL

CP

RemoveP/S FLTR

ADC

Single

Transmitted

Subcarrier

Multipath

Channel

Single

Received

Subcarrier

Distorted

Phase & Magnitude

|a|ejφ

1/|a| e-jφ

-1 -0.5 0 0.5 1

-1

-0.5

0

0.5

1

In-phase Amplitude

Quadra

ture

Am

plit

ude

TX Const

-1 0 1

-1

-0.5

0

0.5

1

1.5

In-phase Amplitude

Quadra

ture

Am

plit

ude

Scatter Plot

-1 -0.5 0 0.5 1

-1

-0.5

0

0.5

1

In-phase Amplitude

Quadra

ture

Am

plit

ude

Scatter Plot


OFDM, Rx Threshold

• OFDM receive signal is made up of a sum of

attenuated, phase shifted and time delayed versions

of the transmitted signal

• Rx Thresh=-114 -10Log(R) +10Log(FS * NUsed/NFFT)

+NF + SNR + LImp

–Add 10Log(Nsubchannel used/32) for OFDMA, when using less

subchannels in the BS Rx

–NF = SNRIn / SNRO, for front end cascaded Rx chain

–R, number of repetitions for the modulation/FEC rate

–FS, sampling frequency in MHz


Packet vs. Frame

• Packet address stays with user data up till the final

destination

• Link address of the frame changes at each physical device

Packet

Network

(destination)

Address

Control

InfoData Payload Pad CRC

Link Address

(Destination-source addresses change

along the path)

User Data

Frame


FER, Frame Error Rate

• Error rate, a quality metrics for data applications

• Performance dependent on the frame length & BER

• Receiving equipment discards the entire packet upon receiving error(s) and requests retransmission

– Lower data throughput vs. BER

– Further reduction in data throughput due to retransmission

• FER1 = BER * (1-BER)^FR*8-1 * (FR*8), for 1 bit error per frame

• FER2 = BER2 *(1-BER)^FR*8-1 *(FR*8)*(FR*8 -1)/2, for 2 bits error per frame

– For example, FER1&2 for frame rate of 64 bytes

• FERActual = 1 - # of non-errored frames Frame received / # of frames transmitted

BER FER1 FER2

1e-6 5.12e-4 1.31e-7

1e-7 5.12e-5 1.31e-9

1e-8 5.12e-6 1.31e-11

1e-10 5.12e-8 1.31e-15

1e-12 5.12e-10 1.31e-19


RX Requirements

• Minimum threshold requirements based on 7 dB NF & 1e-6 BER

• Residual bit error rate to be ≤1e-10

Rx Threshold vs. Modulation & Coding rate, dBm

BPSK QPSK QPSK 16QAM 16QAM 64QAM 64QAM

1/2 1/2 3/4 1/2 3/4 2/3 3/4

1.5 -94 -91 -89 -84 -82 -78 -76

1.75 -93 -90 -87 -83 -81 -77 -75

3 -91 -88 -86 -81 -79 -75 -73

3.5 -90 -87 -85 -80 -78 -74 -72

5 -89 -86 -84 -79 -77 -72 -71

6 -88 -85 -83 -78 -76 -72 -70

7 -87 -84 -82 -77 -75 -71 -69

10 -86 -83 -81 -76 -74 -69 -68

12 -85 -82 -80 -75 -73 -69 -67

14 -84 -81 -79 -74 -72 -68 -66

20 -83 -80 -78 -73 -71 -66 -65

Rx SNR, dB 3.0 5.0 8.0 10.5 14.0 18.0 20.0

Bandwidth

(MHz)


SS/MS Power Restraint

• BS configures each SS/MS PO such that the Rx

power arriving at the BS to remain constant and

consistent with all others regardless of the distance

• Receiver front end must be able to tolerate high

incoming signal level demanding linearity with higher

IIP3 devices

–For direct conversion system, IIP3 demand further

increases due to lack of sharp IF filtering and limited AGC

range


Fade Mitigation

• Narrow band system

–Time diversity

–Freq diversity

–Diversity type interactions

• Wide band system

–Equalization


OFDM, Benefit Summary, (1)

• High spectral efficiency

• Simple implementation by FFT,

modulate by switching between time

and frequency domain

• Lower Rx complexity as Tx combat

the channel effect to some extends

• Resilient to ICI, ISI (by increasing

symbol time)

• Immunity to delay spread and

resilient to MPF

• Equalization is simplified or

eliminated altogether

• Suitable for high data rate transmission

• Highly flexible in term of link adaptation

• Low complexity multiple access (OFDMA)

• Mod/code change on frame to frame and SS to SS depending on robustness (trade-off cap vs. robustness in real time)

• QoS based on latency, jitter & reliable throughput

• Channel impairments and timing problems are both solved with simple phase and channel estimators


OFDM, Benefit Summary, (2)

• Interoperability

• Higher subchannel count

smaller guard band

• Simple equalization. EQ

complexity B*Log(BTg) vs. SC

is B^2 * Tg

• Frequency diversity capable

• Graceful degradation due to

delay spread (ideal for AMC)

• Multi-access using OFDMA

• Robust against narrow band

interference

• Suitable for coherent

demodulation

• TDD, FDD or half FDD

• NLOS

• Easier time-frequency

synchronization

• No inter-carrier guard band

• Resistance to frequency-

selective fading


OFDM, Disadvantages

• Synchronization

– Requires more complex algorithms for time / frequency sync

• Additional circuit for FFT and IFFT is needed

• Greater complexity

• More expensive Tx & Rx

• Reduced efficiency due to guard interval

• Sensitive to phase noise, timing & frequency offsets

– Tight specifications for local oscillators

– Doppler limitation

• High peak to average ratio (PAPR)

– Approximately 10 Log (N), in dB

– Large signal peaks require higher power amplifiers

– Amplifier cost grows nonlinearly with required power

– Need very linear amplifiers with large dynamic range


OFDM Snapshot

• 8 BPSK pilots at fixed location

• 192 data subcarriers, 55 null subcarriers and 1 DC subcarrier

• OFDM Symbol = (1+Cyclic Prefix)/Δf

– Δf (Sub-carrier spacing) is proportional to Channel BW/FFT size

• Sub-channel spacing varies according to the BW

– For narrow BW, sub-channel spacing becomes closer that makes the symbol time longer yielding better performance in NLOS channel

8 BPSK pilots

Fixed location BPSK, QPSK, 16QAM, 64QAMActive Subcarriers: 200

Subcarrier Spacing: 90 kHz

Active Subcarriers: 200

Subcarrier Spacing: 5.6 kHz

20 MHz BW

1.25 MHz BW

14 MHz

o

o

o

2.5 MHz

8 BPSK pilots

Fixed location

GuardbandDC


OFDMA

• Combination of FDMA and OFDM. No guard band between sub-

carriers.

• FFT size is scalable from 128 to 2048

• Fill unused channels with null subcarriers to bring up to next 2N

• Increase the FFT as the BW increases such that subcarriers spacing

remains 10.94 kHz (depends on configurable over-sampling rate)

• Keeps constant symbol duration and have minimal impact on higher

layers

• Sub-carrier spacing can support delay spread up to 20 us, 125 kmph

at frequency 3.5 GHz

• 4QAM,16QAM & 64QAM are used for data. BPSK is used during

preamble, pilot & when modulating subcarriers in the ranging

channel


OFDMA, Example

• Sub-carrier separation remains constant regardless of the BW

• FFT and Subcarriers (Data, Pilot & Null) vary with the BW

• Number of OFDM symbols remain constant regardless of BW in a specific frame rate

1.25 MHz

20 MHz

10 MHz

5 MHz

166-240 QPSK Pilots

Fixed & variable location

82-120 pilots

42-60 pilots

10-16 pilots

FFT: 2048

Active Carriers: 1680

Subcarrier Spacing: ≈11.161 kHz

FFT: 1024


Subcar. Spacing: ≈11.161 kHz

FFT: 512



FFT: 128

Active Carriers: 84


QPSK, 16QAM, 64QAM

DC


OFDMA

• In OFDM, only one MS is transmitted in one time slot

• In OFDMA, several MSs can be transmitted in the same time slot

over several sub-channels

• Time-frequency allocations are done dynamically to improve

performance at the expense of complexity

OFDM OFDMA

Time Time

Su

bca

rrie

rs, fr

eq

ue

ncy

Su

bch

an

ne

ls, fr

eq

ue

ncy

User 1

FFT symbol

User 4

User 3

User 2


OFDMA Symbol Parameters

• Adds additional multiple access features in the frequency domain

• BW is divided into slots for the user in the time and the frequency domain

• OFDMA carriers for different users are very close together (10kHz) & that the order of the

physical carriers may change from Symbol to Symbol

• Difficult to design variable subcarriers spacing. This is mitigated by using FFT size vs. BWParameter Fixed Mode

System bandwidth, MHz BW 3.5-28 1.25 2.5 5 10 20

Sampling frequency, MHz Fs=8/7*BW 4.000 1.429 2.857 5.714 11.429 22.857

Sample time, us Ts=1/Fs 0.250 0.700 0.350 0.175 0.088 0.044

FFT size N 256 128 256 512 1024 2048

# of used data subcarriers NData 192 72 180 360 720 1440

# of pilot subcarriers NPilot 8 12 30 60 120 240

# of null/guard subcarriers NNull/Guard 56 44 46 92 184 368

Subcarrier spacing, kHz Δf=Fs/N 15.625

Useful symbol time, us Tb = 1/Δf 64

Useful symbol BW, MHz Δf*(Ndata+Npilot) 3.125 0.9 2.3 4.687 9.4 18.7

11.160714

89.6 (exact)

Mobile Mode

Available guard time settings Tg = 12.50% Tb/4 Tb/8 Tb/16 Tb/32

Guard time, us Tg 8 22.4 11.2 5.6 2.8

OFDMA symbol time, us Ts=Tb+Tg 72 112.0 100.8 95.2 92.4

TTG+TRG, us PS=4/Fs 1.000 2.800 1.400 0.700 0.175

Symbol per 10 ms frame 10/TS 137 87 98 104 108


IEEE 802.16-2005 OFDMA Physical Layer

Parameters

Modulation QPSK, 16-QAM, 64-QAM

Error correction code CC, BTC, CTC

Overall coding rate ½, ¾, 2/3

Cyclic Prefix 1/32, 1/16, 1/8, ¼

Subchannels 1, 2, 4, 8,16, 32

Bandwidth 1.25, 2.5, 5, 10, 20

FFT 128, 256, 512, 1024, 2048


OFDMA Frame Structure for TDD, Example

• No UL preamble at start of UL subframe but an increased number of pilots. Pilots

in the UL are never transmitted without data subcarriers

• SSs use PN CDMA technique to access the BS in the contention region

• Pilot and null subcarriers are not shown


OFDMA Frame

• A frame is one complete set of DL & UL transmission,

meaning the time between two preambles of the DL signal

• Frame consists of DL & UL subframe with flexible boundaries

• PS (physical slot) is a unit of time defined as 4 modulation-

Symbol length

• FS, sampling freq. = FFT-size * channel-spacing

• UL has no preamble except for system using AAS, but there

are increased number of pilots. Data is transmitted in bursts

that are as long as the UL sub-frame zone allows and

wrapped to further sub-channel as required.


OFDM Frame Structure, Example

• Adaptive & variable

length duration for DL-

MAP, UL-MAP, fast

feedback, ranging and

data burst

• Pilot and null

subcarriers are not

shown

OFDM symbol number (time) Timek,k+1 k+2 k+3 k+4 k+7 k+9 k+11 k+13 ... k+17 k+20 k+23 ... ... k+31 k+33

S FCH FCH

S+1

S+2 DL burst #2 UL burst #1

UL burst #2

DL burst #1 DL burst #3 UL burst #3

UL burst #4

DL burst #4

UL burst #5

DL burst #5

S+L Fast Feedback

RangingDL subframe UL subframe

TTG RTG

S

ub

ca

rrie

r (f

req

ue

nc

y)

P

rea

mb

le

U

L-M

AP

DL

-MA

P

U

L-M

AP

(c

on

ti'd

)

D

L-M

AP

UL

-MA

P

Pre

am

ble


OFDMA Subchannels, (1)

• A subchannel describes the smallest logical allocation unit in the frequency domain. It contains one or more physical carriers, which are adjacent or non-adjacent and whose order may change within a burst from symbol to symbol. Subchannelization is a sophisticated form of frequency division multiple access where multiple subcarriers are grouped into subchannels to enhance system performance. (The number of subchannels varies from 32 to 96, depending on the zone type)

• Subchannel is the basis of OFDMA Multiple Access Method

– Frequency space is divided into subchannels, i.e. Group of subcarriers forms subchannel

– Dynamically allocating time-frequency resources to DL/UL subframe

– At certain moment subchannel is utilized by one transmitter only

– Basic OFDMA time-frequency unit utilized for communication is determined through subchannel-OFDMA symbol combination

– WiMAX uses the term slot for minimum data allocation unit and a slot contains 48 data subcarriers


Zone and Burst

• Zone (contains bursts)

– A zone is one complete logical part of a frame. There are DL and UL zones,

and there are different zone types that may use all subchannels of the OFDMA

frequency range (full usage of subchannels = FUSC) or only parts of them

(partial usage of subchannels = PUSC).

– Grouping of contiguous symbols that use a specific type of subchannel

assignment

– All zones except for AMC use the distributed allocation of subcarriers for

subchannelization

– OFDMA PHY specifies 7 different zones: FUSC, OFUSC, PUSC, OPUSC,

AMC, TUSC1 and TUSC2

• Burst (contains slots)

– A burst is an area within a zone which is assigned to one dedicated user. It

uses a certain number of subchannels (frequency) and a certain number of

symbols (time).


Slot

• Slot

– A slot is the minimum possible data allocation unit within OFDMA,

defined in time and frequency (number of contiguous symbols

times number of subcarriers). It always contains one subchannel

and can contain one to three symbols (depending on the zone

type). A DL-PUSC slot is two symbols wide, a UL-PUSC slot three

symbols wide.

• Minislot

– A unit of UL BW allocation equivalent to n physical slots, where

n=2^m, m is an integer ranging from 0 through 7


OFDMA Subchannels Terms

OFDMA Symbol Number Timek+0 k+1 k+2 k+3 k+4 k+5 k+6 k+7 k+8 k+9 k+10 k+11 k+12 k+13 k+14

0

1

2 Slot Subchannel offset3

4

5

6

7 Sym offset

8 Subcarriers

9

10

11

12 # of OFDMA symbols13

14

15

Slot Data Region Segment Permutation zone

Su

bch

an

nel L

og

ica

l Nu

mb

er


Mandatory and Optional Zones

• In OFDMA PHY, the mapping from data bit to physical subcarriers is achieved in

two steps:

– The 1st step is to map the data to one or more time slots and one or more

logical subchannels

– The 2nd step is called permutation, in which the logical subchannels are

mapped to physical subcarriers

• Multiple permutation zones marked by Zone Switch IEs (AAS_DL_IE, AAS_UL_IE,

STC_DL_Zone_IE)

• Switching from Non-STC to STC, and Non-AAS zones is defined by the IEs

Must appear in every frame

DL Subframe

Zone switch IEs in DL-MAPMay appear in frame

Pre

am

ble

PU

SC

(DL

_P

erm

Ba

se

X)

UL Subframe

PU

SC

(1st z

on

e c

on

tain

s

FC

H &

DL

-MA

P)

FU

SC

(DL

_P

erm

Ba

se

Z)

FU

SC

(DL

_P

erm

Ba

se

Y)

AM

C

TU

SC

1

TU

SC

2

AM

C

Op

tio

na

l P

US

C

PU

SC

Op

tio

na

l F

US

C



• Like an OFDM, OFDMA symbol contains subcarriers

• Subchannel means splitting a normal channel BW into more than one

• Subchannel contains a group of subcarriers

• User is assigned one or more subchannel

– 1, 2, 4, 8, 16 or 32 subchannels for UL. Two of the UL subchannels are used for ranging and BW request

– Link budget improvement in UL for fixed WiMAX, 12 dB when used 1/16th of BW

• Power concentration on few subcarriers

• Reduces over head control algorithm complexity

• Subcarriers of a subchannel can be contiguous or distributed over the BW

– Randomly distributed type provides frequency diversity in frequency-selective fading channels and inter-cell interference averaging.

• Well suited for mobile applications

– Adjacent type is useful for frequency non-selective and slowly fading channels, and for implementing ACM. Typically used for fixed or low mobility applications.

• Channel estimation is easier as the subcarriers are adjacent

• Well suited for fixed and low mobility applications



• Basic principle is to trade off mobility for throughput

• Subchannels are dynamically allocated to users for UL &

DL data based on CINR

• Subcarriers are randomly assigned to subchannel and

changed every symbol time

• Different subchannel allocation methods to

– Optimize the frequency band for stationary or mobile usage

– High or low interference from neighboring sectors, cells

– Optimize diversity and beamforming techniques performance

• 16e system uses subchannels in both directions whereas

16d applies in the UL only


Subchannels, Power Concentration

• Increases the Tx power & improve asymmetric link budget on cost prohibitive CPE

• OFDMA example: 5 MHz BW, 480 active subcarriers, 200 mW CPE PO, subchannelization assigns 120 subcarriers to CPE

• PSD = PWR/BW = 200mW/5M =40 uW/Hz for all subcarriers

• PSD = PWR/BW = 200mW/1.25M =160 uW/Hz for 120 subcarriers, yielding 6 dB gain, 10Log(480/120)– 1/2 of tones yield 3 dB power gain

– 1/4 of tones yield 6 dB power gain




DL Bandwidth

UL All Sub-channels

UL 1/4 Active Subchannels

4x sub-ch

Po

All-ch Po


OFDMA Subcarriers Mapping into

Subchannel

Subchannel1

DC subcarrier

Subchannel NSubchannel N-1Subchannel2

guard band guard band

Subchannel1

DC subcarrier

Subchannel 3Subchannel2

guard band guard band

Distributed allocation

Adjacent Allocation

Subchannel N


Subchannelization

• Very effective under stressful channel conditions


Subchannelization

• Narrowband interference rejection

• Easy to avoid/reject narrowband dominant interference

• Less interfered part of the carrier can still be used

User Subcarriers

AllocationInterference

Subcarriers

User Subcarriers

AllocationInterference

Subcarriers

Null

Subcarriers

Total Frequency Band

Before

After


FUSC, Full Usage of Subcarriers

• DL FUSC

– Fixed & Variable pilot tones are added for each OFDMA symbol independently

• One set of common pilot subcarriers and always at the same location

– Remaining subcarriers are divided into subchannels that are used exclusively

for data

– User are allocated slots for DL data transfer

• One slot is a single subchannel with 48 subcarriers by one OFDMA symbol

(48 tone-symbols)

– Provides full frequency diversity and inter-cell interference averaging by

spreading the subcarriers over the entire band

BW

(MHz)FFT Size

AMC Subchannels

(downlin/Uplink)

(downlink/Uplink)

PUSC Subchannels

(downlin/Uplink)

(downlink/Uplink)

FUSC Subchannels

(downlink only)

1.25 128 2/2 3/4 2

5 512 8/8 15/17 8

10 1024 16/16 30/35 16

20 2048 32/32 60/92 32


PUSC, Partial Usage of Subcarriers

• DL & UL PUSC

– The set of used subcarriers is partitioned into subchannels

• Groups the sub-carriers into tiles to enable fractional frequency reuse scheme (FFRS)

– Pilot subcarrier are allocated from within each subchannel

• Each subchannel contains its own set of pilot subcarriers

– User are allocated slots for DL/UL data transfer

• For DL, one slot is a single subchannel by two OFDMA symbols

– DL-PUSC slot uses a cluster structure. One subchannel contains two cluster. One cluster contains 12 data subcarriers and two plot subcarriers

• For UL, one slot is a single subchannel by three OFDMA symbols

– One subchannel contains 6 tiles. One tile contains 4 subcarriers. One tile with three symbols contains 4 pilots subcarriers and 8 data subcarriers

• One subchannel symbol consists of 48 tones minimum

– Provides frequency diversity function (minimizes interference between Inter-cell as well as between adjacent sectors)


OFDMA Permutation

OFDMA Permutation

1.25 MHz

BW

5 MHz

BW

10 MHz

BW

20 MHz

BW

DL FUSC No. of subcarrier 105 426 851 1702

No. of data subcarrier 96 384 768 1536

DL PUSC No. of subcarrier 85 421 841 1681


DL O-FUSC No. of subcarrier 108 432 864 1728


DL O-AMC No. of subcarrier 108 432 864 1728


UL PUSC No. of subcarrier 97 408 840 1681

No. of data subcarrier 272 840

UL O-PUSC No. of subcarrier


Ul O-AMC No. of subcarrier 108 432 864 1728



PUSC, Example

• PUSC with STC

OFDMA Symbol Number

Unallocated

User 3, Matrix BUser 3

Matrix A

User 2

Matrix BUser 1

Matrix B

User 2

Matrix A

Su

bch

an

ne

l L

og

ica

l N

um

be

r

User 1, Matrix A

DL-PUSC Subframe with STC Zone (2 Antennas

OFDMA Symbol Number

Unallocated

User 1 None, 1 Tx Antenna

User 2 Matrix B, 2 Antennas

Su

bch

an

ne

l L

og

ica

l N

um

be

r User 1, Matrix A, 2 Tx Antennas

UL-PUSC Subframe with STC Zone


OFDMA Highlights

• 16e intended for mobile broadband connection for pedestrians and

automobiles in 1-3 mile radius range

Maximum subscriber throughput 3 Mbps per DL, 1 Mbps per UL

Maximum sector throughput

(10MHz band)

18 Mbps per DL, 6 Mbps per UL

Frequency reuse 1

Mobility Up to 120 km/h

Handoff Under 150 ms

Service coverage Macro (1km), Micro (400m), Pico (100m)

Roaming Seamless roaming with cellular and

WLAN

QoS offering Unsolicited grant service, extended real-

time, real time, non real-time, best-effort

Uplink/Downlink ratio Software adjustable


OFDMA Data Rates

Mod Code rate 5MHz 5MHz 10MHz 10MHz

DL rate Mbps DL rate

Mbps

UL rate

Mbps

DL rate

Mbps

UL rate

Mbps

4QAM ½ CTC, 6x 0.53 0.38 1.06 0.78

½ CTC, 4x 0.79 0.57 1.58 1.18

½ CTC, 2x 1.58 1.14 3.17 2.35

½ CTC, 1x 3.17 2.28 6.34 4.70

3/4 CTC 4.75 3.43 9.50 7.06

16QAM ½ CTC 6.34 4.57 12.07 9.41

¾ CTC 9.50 6.85 19.01 14.11

64QAM ½ CTC 9.50 6.85 19.01 14.11

2/3 CTC 12.67 9.14 26.34 18.82

3/4 CTC 14.26 10.28 28.51 21.17

5/6 CTC 15.84 11.42 31.68 23.52


OFDMA Advantages-Summary, (1)

• Enables adaptive modulation for every user using QPSK,

16QAM & 64QAM

• Performs adaptive FEC based on CINR, RSSI & error rate

• Enables dynamic subcarrier allocation

– Efficient use of air resources for mobile applications

• Enables spatial diversity by using antenna diversity at the

Base station and possible at the Subscriber Unit

• Gives frequency diversity by spreading the carriers all over

the used spectrum

• Gives time diversity by optional interleaving of carrier

groups in time



• Using the cell capacity to the outmost by adaptively using

the highest modulation a user can use, this is allowed by

the gain added when less subcarriers are allocated,

therefore gaining in overall cell capacity

• The power gain can be translated to distance

– Doubles the distance for each 6 dB gain in LOS conditions

• Enabling the usage of indoor Omni Directional antennas

for the users

• MAC complexity is the same as for TDMA systems



• Allocating carrier by OFDMA/TDMA strategy

• Minimal delay per OFDMA symbol of 300us

• Using small burst per user of about 100 symbols for better

statistical multiplexing and smaller jitter

• User symbol is several times longer than for TDMA

systems

• Using the FEC to overcome error in disturbed frequencies

• Adaptive modulation and coding

• OFDM is a proven technology for transporting high data

rates for NLOS, long ranges with multipath conditions like

for DVB-T, DAB etc.


Drawbacks of OFDMA

• OFDMA signal due to its extreme amplitude variation does not behave well in non-linear channel.

• PAPR, dB = 10Log(N), OFDM composite signal exhibits significant peaks and valleys (when all carriers add in phase) with depths of more than 50 dB but the probability of its occurring is low due to scrambling and higher number of N. Some uses advanced coding techniques to minimize its effect. Higher N means higher peak, requiring linear & expensive PA. Increases ADC/DAC complexity.

• Highly sensitive to timing jitter and frequency offsets

• Doppler limitation

• Susceptible to phase noise as each subcarrier is Mod by phase noise of the LO

• Loss due to guard band, typically set ≤ 1 dB

• Reduced channel spacing at higher N increases chance of ICI

• Sync: requires more complex algorithms for time/frequency sync


AMC, Adaptive Modulation & Coding, (1)

• Uses adjacent sub-carriers for each subchannel for use with beam forming

• Modulation, power and or Coding change on a burst by burst basis per link

• Channel response remains flat over narrow subcarrier

• A closed loop control process

– TX controls the Capacity, QoS, ECC, symbol mapping and power

– Rx feeds back the current SINR, BER & RSL information

– Basic idea is to transmit as mush data as possible and throttle down when channel is not good

• AMC for OFDMA: each user is allocated a group of subcarriers, each having different SINR. Care needs to be paid in selecting Mod-coding based on varying SINR across subcarriers

• DL burst goes to one or more SS using the same mod & coding. The UL burst comes from individual users with individual devices. The SS are told when to transmit.

• Demands fast settling response from PA

Subcarriers

256QAM

Not used due to

low SNR

64QAM

16QAM

4QAM

BPSK

Ave

rag

e S

NR

BPSK

SNR = 6 dB

4QAM

SNR = 9 dB

64QAM

SNR = 22 dB

16QAM

SNR = 16 dB



• Down Link

– Modulation

• BPSK, QPSK, 16QAM, 64QAM

• BPSK optional for OFDMA-PHY

• 256QAM optional

• Adaptive power back-off per

Modulation

– Coding

• Mandatory: concatenated

convolutional codes at rate ½, 2/3,

¾, 5/6

• Optional: convolutional turbo codes

at rate ½, 2/3, ¾, 5/6 & LDPC

– Capacity, QoS and power back-off

management

• Up Link

– Modulation

• BPSK, QPSK, 16QAM

• 64QAM optional

• Adaptive power back-off per

Modulation

– Coding

• Mandatory: concatenated

convolutional codes at rate ½,

2/3, ¾, 5/6

• Optional: convolutional turbo

codes at rate ½, 2/3, ¾, 5/6.

Repetition codes at rate ½, 1/3,

1/6 & LDPC

– Capacity, QoS and power back-off

management



• A closed loop controlled process

• Affects on incoming data queuing, FEC, Mod and PO in the transmit

direction

• Dynamically adapts the transmitting signal based on channel status

(RSL, CINR & PER) from the far end receiver

– Assigns adjacent subcarriers to specific SS/MS

• Improved performance in OFDMA subchannelization

4

Transmitter Receiver

Bits

Out

Feedback Channel

RSL, SINR, PER

Queue

Bits

In

Select

Code

Select

Constellation

Power

Control

Symbol

Mapper POECC

Encoder

Adaptive Modulation and Coding

Controller

Channel

SINR Demod Decoder

Channel

Estimation


PHY Parameters

• Table assumes 5 ms frame rate and a Tg 12.5% (1/8) of Tb

Parameter

Fixed

WiMAX

OFDM-PHY

FFT size 256 128 512 1024 2048

Number of used data subcarriers 192 72 360 720 1440

Number of pilot subcarriers 8 12 60 120 240

Number of null/guardband subcarriers 56 44 92 184 368

Cyclic prefix or guard time (Tg/Tb) 12.5%

Oversample rate Fs/BW

8/7 for OFDM, 28/25 for OFDMA 8/7 for multiples of 1.75 MHz, 2 MHz or 2.75 MHz

Channel bandwidth (MHz) 3.5-28 1.25 5 10 20

Subcarrier frequency spacing (kHz),ΔF 15.625

Useful FFT symbol time (us), Tfft=1/ΔF 64

Guard time assuing 12.5% (us), Tg 8

OFDM symbol duration (us), T=Tfft+Tg 72

Number of OFDM symbol in 5 ms frame 69

1/32, 1/16, 1/8, 1/4

Depends on bandwidth: 7/6 for 256 OFDM, 28/25

Mobile WiMAX Scalable

OFDMA-PHY

48

10.94

91.4

11.4

102.9


PHY Estimated Data Rate vs. BW

Parameter

Channel Bandwidth

(MHz)

FFT size

Oversampling Rate

Mod & Code Rate

DL UL DL UL DL UL DL UL DL UL

BPSK, 1/2 946 326

QPSK, 1/2 1882 653 504 154 2520 653 5040 1344 4464 1120

QPSK, 3/4 2822 979 756 230 3780 979 7560 2016 6696 1680

16QAM, 1/2 3763 1306 1008 307 5040 1306 10080 2688 8928 2240

16QAM, 3/4 5645 1958 1512 461 7560 1958 15120 4032 13392 3360

64QAM, 1/4 5645 1958 1512 461 7560 1958 15120 4032 13392 3360

64QAM, 2/3 7526 2611 2016 614 10080 2611 20160 5376 17856 4480

64QAM, 3/4 8467 2938 2268 691 11340 2938 22680 6048 20088 5040

64QAM, 5/6 9408 3264 2520 768 12600 3264 25200 6720 22320 5600

PHY-Layer Data Rate (kbps)

Not applicable

Mobile WiMAX Scalable

OFDMA

Fixed WiMAX

OFDM

10

1024

28/25

8.75

1024

28/25

5

512

28/25

3.5

256

8/7

1.25

128

28/25


Interference, (1)

• Presence of one or more undesirable signals that degrades a normal performance

• Potential source could be from its own internal, external equipment or both

• Non-optimized path, network and frequency planning

• Sharing dual pole antenna & crossed pole interference

– Low antenna XPD

• Co-located antennas, low discrimination antenna & sector spill over

• Most interference detection tests are traffic affecting

• Maintenance activity may introduce interference

• EMI and interference from co-located equipment

• ATPC helps minimizing the interference affect

• OFDMA spreads the energy of an impulse noise over an OFDMA burst that results in smaller increased noise rather than losing symbols

• Triple transient signal interference due to poor termination, return loss & simple Frequency Domain Equalizer


Interference, (2)

• Ways to check interference

– Easier to test prior to commissioning

– Scan RX frequency & power facing toward remote with spectrum analyzer

– Difficult to detect intermittent types of interference

– TX fade test with built-in attenuator

– Capturing and plotting RSSI & CINR over time

– Mute remote adjacent/opposite polarity TX and check RSSI & PER

– Mute remote opposite polarity TX and check RSSI, CINR & PER

– Mute remote TX and check RSSI, CINR & PER

– Change polarization and check RSSI & BER

– Change operating frequency within the band and check RSSI, CINR & PER


Interference, (3)

• Acceptable interference level = 1e-6 Threshold - T/I - MEA

– T/I, ratio of the RX threshold and interference vs. Freq offset

– Interference level not to cause more than 1 dB degradation at 1e-6 BER

– Using same bandwidth and type of signal

– MEA maximum number of exposure allowed

• The terms 1e-3 BER threshold, 3 dB degradation, C/I &

MEA are more relevant to analog radios

• Determining threshold degradation at specific interference

level

– 10Log[1+10(I-N)/10]

• I interference level in dB

• N noise level in dB


Noise & Interference Relationship

Unfaded RF RX Level

SNR

6 dB, objective for 1 dB

threshold degradation

T/I

N

Interference Level

1e-6

BER Threshold with Noise

Noise Floor

SINRReq

S

I

Thermal Static Fade Margin

SIR

T1e

-6 BER with I+N

Noise Floor + Interference

1dB


PCINR, Physical CINR

• A quick and accurate CINR estimation information is required from the SS in order for BS to select an appropriate modulation coding scheme for that SS

• I+N conditions over a symbol or channel vary rapidly therefore it is important to estimate both average and instantaneous CINR

• SS/MS measures the CINR from DL preambles and reports back to BS in REP-RSP message

• C/N, Carrier to thermal noise ratio

• C/(N+I), Carrier to thermal noise plus interference ratio

– Interference level that causes 1 dB degradation

– During non-boosted data subcarriers

• PCINR:

– PCINR = (3/8 *Cpmbl) / (3/8 * Ipmbl + N) = Cpmbl / (Ipmbl + 8/3 * N)

• Cpmbl & Ipmbl power measured during preamble, 3/8 to scale down due to preamble

• ECINR, CINR estimation based on the pilot sub-carriers


MAC, Media Access Control Overview (1)

• Designed for Point-to-Multipoint based on collision sense multiple access with collision avoidance (CSMA/CA – listen before transmit)

• Targeted for Metropolitan Area Network applications

• Connection-oriented MAC– Connection ID (CID), Service Flows(SF)

• Supports difficult user environments

– High bandwidth, hundreds of users per channel

– Continuous and burst traffic

– Very efficient use of spectrum

• Protocol-Independent core (ATM, IP, Ethernet,...)

• Balances between stability of contentionless and efficiency of contention-based operation

• Data control plane (traffic scheduling to provide QoS) with speed up to 268 Mbps each way

• Supports multiple 802.16 PHYs


MAC, Media Access Control Overview, (2)

• Interfaces between higher transport layers and PHY. MAC layer takes

packets from upper layer. These packets are called MAC service data

units (MSDUs) and organizes into MAC protocol data units (MPDUs) for

transmission over the air. For Rx it performs reverse process. MAC

includes convergence sub-layer that can interface with a variety of higher

layer protocols such as ATM, TDM, voice, Ethernet, IP and other unknown

future protocols. Beside providing a mapping to and from the higher layer,

the convergence sub-layer supports MSDU header suppression to reduce

the higher layer overhead on all packets.

• Supports very high peak bit rates while delivering QoS similar to ATM and

DOCSIS (data over cable service interface specifications).

• Uses variable length MPDU and offers a lot of flexibility to a lower layer for

their efficient transmission (multiple MPDUs of the same or different length

may be aggregated into a single burst to save PHY overhead. Conversely,

large MSDUs may be fragmented into smaller MPDUs and sent across

multiple frame)


MAC, (1)

• Management messages: FCH, DL-MAP, UL-MAP, DCD, UCD, Data.

– FCH frame control header, DCD downlink channel description (PHY

characteristics. DL-frame prefix (24 bits)

• OFDM MAC is designed for efficient use of spectrum

• Very high bit rate, DL & UL broadband services

• Supports multi-services simultaneously with full QoS

– Efficient transport IPv4, IPv6, ATM, Ethernet, VLAN, etc.

• Flexible QoS offerings

– CBR (unsolicited grant service, highest), rt-VBR, ert-VBR, nrt-VBR, BE

(lowest), with granularity within classes

– QoS per user and per connection basis

• Protocol-independent engine

– Convergence layers to ATM, IP, Ethernet, ...

• Extensive & strong security types encryption/decryption (DES, 3DES, AES, RC4,

data encryption standards)


MAC, (2)

• ARQ, Automatic Repeat Query

– Done at MAC layer rather than at TCP layer 4 (yields less outage)

– Adds error detection ability in data stream

– Bad packets are retransmitted

– Detecting errors using CRC-32 codes

– Not efficient in broadcast systems

– Not used in voice services

• OFDM/OFDMA support

• Dynamic Frequency Selection

– For license-exempt applications

• Adaptive antenna system support

• Mesh mode

– Optional topology for license-exempt operation

– Subscriber to subscriber communications

– Complex topology and messaging, but:

• Addresses license-exempt interference

• Scales well

• Alternative approach to non-line-of-sight


MAC, (3)

• Packet classification – IP and Ethernet

– DSCP / TOS (any bit)

– Source / destination MAC and / or IP

– Source / destination Port ranges

• Packet convergence sublayer support for:

– IPv4 and IPv6

– Packet IPv4 & IPv6 over 802.3 (Ethernet)

• Dynamic service flow creation – BS/MS initiated

• PHS (packet header suppression) & ROHC (robust header

compression Rel 4.x)

• PKMv2 (privacy key management) support for

– EAP based authorization

– Cryptographic suites (for data encryption, CCM-mode 128 bit AES key)


MAC Operation


MAC Requirements

• Provide Network Access

• Address the Wireless environment

– e.g., very efficient use of spectrum

• Broadband services

– Very high bit rates, downlink and uplink

– A range of QoS requirements

– Ethernet, IPv4, IPv6, ATM, ...

• Likelihood of terminal being shared

– Base station may be heavily loaded

• Security

• Protocol-Independent Engine

– Convergence layers to ATM, IP, Ethernet,...

• Supports for both TDD and FDD in PHY


Physical Layer Convergence Procedure

(PLCP Packet Structure 802.11)

• PLCP Preamble

– Short training symbol t1 to t10

– Long training symbol (T1 & T2)

• PLCP Signal field

• PLCP Data

• The training length period varies for 802.16 due to varied IFFT size

T2T1t10t1 t7t6t5t3 t4 t8 t9t2 GI2 Data 1 Data 2GI GIGISignal

8+8 = 16 us, Preamble10x0.8 = 8 us 2x0.8 + 2x3.2 = 8 us 0.8+3.2 = 4 us 0.8+3.2 = 4 us

0.8+3.2 = 4 us

OFDM Service+DataRate

Length

Channel and Fine

Frequency Offset

Estimation

Coarse Freq

Offset Estimation

Timing Sync

Signal Detect, AGC,

Diversity Selection


MAC Management Messages, (1)

• Connection orienteded

– For each direction, connection identified with a 16 bit CID

– Each CID is associated with a Service Flow ID (SFID) that determines the QoS parameter for that CID

• Admission control plane (ensures that new flows do not degrade the quality of established flows)

• Channel access:

– UL-MAP

• Defines uplink channel access

• Defines uplink data burst profiles

– DL-MAP

• Defines downlink data burst profiles

• QPSK-1/2, Defines allocated data regions for UL-MAP

• Defines UL BW-request, Ranging, CQICH ... regions

• Defines allocated data regions for SS/MS DL/UL reception/transmission

• Defines multiple permutation zones (if present)

– UL-MAP and DL-MAP are both transmitted in the beginning of each downlink subframe (FDD and TDD)


MAC Management Messages, (2)

• From BS: Preamble>FCH>DL-MAP>UL-MAP>DCD>UCD>Data>

Data...

– Preamble: Provides fixed known pattern to aid in Rx timing recover

– FCH: The FCH specifies the burst profile and length of one or more bursts that

follows the FCH. DL_Frame Prefix (24 bits)

– FCH: In fixed WiMAX, FCH is transmitted at the lowest mod and highest

coding rate. BPSK ½ rate and occupies 1st two subcarriers

– FCH: In mobile WiMAX, FCH transmission is repeated for robustness.

QPSK ½ rate, repetitions coding of 4 and occupies 1st two subchannels

– DL-MAP: describes the DL allocations, PHY sync info, BS identifier,

DCD identifier that is used in the allocation.

– UL-MAP: describes the UL allocations. Consists of UL Ch ID, UCD (UL

Ch descriptor) identifier, start time, PHY specific UL-MAP elements that

define allocations.

– Configurable fixed duration frames


QoS

• Advanced QoS features:

– Weighted fair queuing

– Traffic shaping

– Congestion management

– Random early detection

– Hierarchical QoS


DL BW Calculation, Example

input chan size

(MHz)

Calculates

frame O/H

Calculates

frame BW

Calculates

preamble O/H

Calculates

useful Channel BW

input mod and

coding distribution

input cyclic prefix

(¼ to 1/32)

Calculates

DL map O/H

Calculates

FCH O/H

Calculates

useful frame BW

Calculates

UL map O/H

Calculates

useful MAC BW

Calculates

MAC hdr O/H

Calculates

MAC subhdr O/H

Calculates

MAC CRC O/H

input avg user pk

(B)

input frame length

(2 to 20 ms)


UL BW Calculation, Example

input chan size

(MHz)

Calculates

frame O/H

Calculates

frame BW

Calculates

ranging O/H

Calculates

useful Channel BW

input mod and

coding distribution

input cyclic prefix

(¼ to 1/32)

input

subchannel size

input burst sizee

Calculates

contention O/H

Calculates

MAU

Calculates

useful frame BW

Calculates

Preamble O/H

Calculates

Subchannel O/H

Calculates

useful MAC BW

Calculates

MAC hdr O/H

Calculates

MAC subhdr O/H

Calculates

MAC CRC O/H

input avg user pk

(B)

input frame length

(2 to 20 ms)


Allocating Network BW, Example

Calculate user

Channel BW

Calculate VBR

MS BW

Calculate BE

MS BW

Calc VBR MS %

remaining BW

Allocate VBR

MR BW

Calc remaining

Channel BW

Allocate CBR BW

Calc BE MS%

remaining BW

Allocate BE

MS BW

No remaining BW

done

Calc remaining

Channel BW

Allocate VBR

MR BW


Power Saving Modes

• Three types of subscriber power management support

–Normal operation

–Sleep mode

– Idle with paging support


PHY Operating Modes, Optional

• Sleep

– MS with active connection temporarily disrupt connection for a predetermined time followed by a listen window

– Sleep and listen windows are negotiated between the BS and the MS

• Duration depends on the saving class

– Class I, sleep window increases exponentially from minimum to maximum. Used for BE or non-real time traffic

– Class II, fixed length sleep window and used for UGS service

– Class III, One time sleep window typically used for multicast or management

• Idle

– Increased power saving than sleep mode

– MS to receive broadcast DL transmission from BS without registering itself with the network

– No handover action

– BS conserves PHY and MAC resources


Security

• Private Key Management (PKM) for MAC layer

security

–56 bit DES-CBC, 128 bit AES-CCM encryption

–X.509 certification

–RSA authorization

–HMAC message integrity protection

• PKM version 2 supports Extensible Authentication

Protocol (EAP)


MBS, Multicast and Broadcast Services

• Multicast polling is done when there is insufficient bandwidth to poll each MS/SS individually

• Signaling mechanisms for MS to request and establish MBS

• SS access to MBS over a single or multi BS, depending on its capability & desire

• MBS associated QoS & encryption using a globally defined traffic encryption key

• A separate zone within the MAC frame with its own MAP information for MBS traffic

• Method for delivering MBS traffic to idle mode SS

• Support for macro diversity to enhance the delivery performance of MBS traffic

• Certain CIDs are reserved for multicast groups and for broadcast messages


Advanced Features

• H-ARQ, Packing, PHS, PKMv1

• Hitless Handoff

• Antennas: SIMO, MISO, MIMO

• MIMO Matrix-C (4x4)

– Four data streams are transmitted in parallel from four antennas per symbol yielding four times the baseline data rate

– Multiple separately encoding (horizontal) streams are transmitted over multiple antennas

• Adaptive MIMO mode switching


MIMO

• Multiple Inputs to the TX antenna(s) and Multiple Outputs of RX antenna(s)

PHY

PHY

RF

RF

RF

RF

SSPHY

PHY

RF

RF

RF

RF

MAC

MAC

BS

RF

RxTX

TX

Rx


MIMO

• MIMO techniques improves system performance, robustness, throughput and coverage

• It takes advantage of multi-path and reflected signals that occur in NLOS environments

– BW of each subcarrier is small that enables the low cost DSP (PHY layer technology) to practically calculate the MIMO coefficient

• MIMO needs a better SNR than SISO

• Reduces interference and improves fade margin by using multiple adaptive antennas in TX/RX diversity. Approximate gain increase of 10Log(# of antenna array elements)

• Space time coding (transmit diversity technique by taking pair of symbols, time reverse each pair for transmission on a second antenna)

– Space-time diversity coding (up to NxCap but no increase in peak data rate)

• MIMO (spatial division multiplexing), beam forming with multiple antennas.

– Spatial multiplexing increases peak data rate by up to Nx with Nx antennas


Switched SISO

• Receiver performs a quick AGC and level check on arriving packet

and switches to a stronger signal based on the signal level

• Transmitter keeps knowledge of the channel performance and

switches to the better TX.

Transmitter Signal Strength

Transmitter

Select AntennaReceiver

Knowledge of

the channel

Introduction to WiMAX TechnologyPage 245

MIMO Matrix A (STC)

• MIMO Matrix-A uses two or more antennas at transmitter (2x1) and one or more at far end receiver (1x2)

• Space time coding (transmit diversity) is a method which yields diversity gain without channel knowledge in the transmitter by coding across antennas (space) and across time– Applies well known Alamouti code in the downlink direction– Provide reliability improvement via diversity with transmitting two redundantly

encoded data streams (time reversing each pair of symbols for transmission on second antenna) during the same symbol and enables to utilize spatial (or polarization) diversity gain

– Overall data transfer rate remains the same as the baseline data but holds the throughput under difficult conditions

– Increases signal strength (3 dB higher SNR at stable conditions and about 10 dB at faded conditions compared to non-STC) by coherently combining multiple signals

– The Tx streams must originate at the same frequency and phase

• High order modulations are more sensitive to multi-path and other impairments– One remedy is an aggressive use spatial / frequency / time diversity – Space time coding (STC) is a well proven way to improve system performance – Performance equivalent to maximum ratio combining with two RX antennas


STC, (1)

• There are two transmit antennas at the BS side and one reception antenna at the SS side (MISO system)

• Each TX antenna has its own OFDM chain

– Distinct pilot subcarrier location for each antenna

– Common location for data subcarrier but its content in a different order

– This technique requires Multiple Inputs Single Output channel estimation

• Decoding is very similar to maximum ratio combining

• STC achieves near optimal diversity gain in slow fading (coherence time is ≥ 10 times the channel estimate update period) environment

Time

Frequency

DC Subcarrier

Data Subcarrier

Pilot Subcarrier

Antenna 1 Antenna 2


STC, (2)

• Cheaper to implement in BS than the SS

• Applies cyclic shift into one Tx path (typical delay of 50 to

200 ns)

• Two forms: with coding and without coding

TX A

TX B

Modified

SignalRx

Data B AC

B ACBAC‟


STC, (3)

• First channel uses: antenna 1 transmits S0 and antenna 2

transmits S1

• Second channel uses: antenna 1 transmits -S1* and antenna 2

transmits S0*

IFFT Input

Packing

[S0, -S1*]

TX Diversity

Coding

[S1, S0*]

IFFT

Diversity

CombiningFFT

RF/ADC

IFFT DAC/RF

Tx-1

Tx-2

Rx

Subch.

Mod

Filter

Subchannel

detection

Log

Likelyhood

Ratio Decoder

Filter

Filter

DAC/RF


STC, Space Time Coding

• DL Tx diversity that can add up 3 to 10 dB link margin in

faded NLOS environments

• Up to N times the capacity/frequency but no increase in peak

data rate

• Provides large coverage regardless of channel condition

• Adds robustness to time fluctuations and decreases

frequency selectivity

• Applies new coefficients to the computation equations upon

receiving signal offset by half wavelength

• BS continuously optimizes algorithm by obtaining

performance data results from the MS


MIMO Matrix B (SDM)

• Two independent data streams are transmitted over two antennas in the same

time-frequency slots

• An independent data stream is mapped to each transmit antenna and sends only

once (unlike STC which sends the data twice)

• Requires two receive antennas at the MS with proper signal equalization and

decoding

• Enables spatial multiplexing in down link, doubling the capacity and providing

unmatched spectral efficiency when channel conditions are poor

• Up to N-times the peak data rate increase with N-times antennas

• Good signal quality is required and the correlation has to be low enough

• Huge capacity increase is expected for pico and nano cell

• Improves robustness to multipath fading using space diversity

• BS continuously optimizes algorithm by obtaining performance data from MS


MIMO Matrix B (SDM)

• SDM: Space Division Multiplexing (2x2) increases the capacity by

transmitting multiple data streams in parallel on different antennas while

reducing the signal quality

– No increase in cell range because users near the cell edge typically have low SNR


MRC, Maximal Ratio Combining

• Both BS and MS receivers are equipped with two receive antennas

performing Maximal Ratio Combining (MRC) technique for both DL

& UL

• Increases SNR and robustness, especially in dynamic and

frequency selective channel by averaging symbol error probability in

an additive white noise channel

• Adds spatial diversity gain at the receiver to further increase the link

budget

Transmitter+


Adaptive MIMO Mode Switching

• Matrix A adds robustness and coverage, while Matrix B increases

capacity, but decreases the robustness

• Smart adaptation algorithms are required for making decisions when

to use Matrix A and when to use Matrix B

• Switching and algorithms decisions are made based on CINR and

antenna correlation, but also on the capacity and QoS requirements

• Certain MIMO techniques apply pre-coded transmission and use fast

feedback slot (6-bit payload) for coefficient update

– 1 subchannel x 6 Tiles x 8 Data-subcarriers = 48 QPSK modulated subcarriers

– Mapping of the MIMO coefficients to the 6-bits payload done by a codebook

and a sequence of signal phase


Adaptive MIMO Switching


Collaborative Spatial Multiplexing

• Two MS transmitters, each with one Tx antenna, may

transmit at the same time and on the subchannels

• In the UL, BS can receive signals simultaneously from

two MSs in the same time-frequency slot

• Achieves multiplexing gain (capacity increase) through

collaboration of MSs

• Does not increase the peak UL data rate of the modem, but can

double the cumulative uplink data rate in a sector

• This technique is also called space division multiple

access (SDMA) and requires multiple antennas at the

base station


Beam Forming (SDM)

Antenna array focuses energy in selective area or null steering

in interferer

– The narrowness of the beam is directly proportional to number of

antennas and their gain

– Achieve additional robustness and capacity

– Higher peak rate at cell’s edge

– Robustness against inter-cell interference

• Multiple antennas are required at the BS side

– Emission patterns are controlled with phase and amplitude

• Especially beneficial in larger cell with higher antennas

• Increases link budget and decreases interferences


Two Types of Beamforming

• Beamforming with phase array antenna

– applies to LOS & SC

– applies to TX or RX

– Requires regular scanning mechanism like omnidirection beacon

– Interference rejection at RX is equally important as increased wanted signal

– RX signal strength depends on phase alignment of the incoming signal

• Beamforming with MIMO SDM

Co-Channel

Algorithm

+

Transmitter

Multipath

Desired

Delay

Gain=A+jB

Gain=C+jD


Interference Mitigation with Beam Forming


AAS, Adaptive Antenna System

• AAS optional feature improves system capacity by spatially

overlay coverage area by adding additional independent

antennas systems

• Increases SNR gains toward SS while placing nulls on

interfering transmitter

• Increases or decreases antenna gains toward affected

direction

• Enables transmission of DL and UL burst using directed beams

to intended one or more SSs

• Increases expense

• Implemented by using multi-element phase array BS antennas


ARQ (16d), Automatic Repeat Query

• Provides a rapid retransmission

• Implemented at below the MAC layer

– Process hides the errors from TCP stack and simulates TCP error correction at lower layer

• Allows selective repeat (stop, wait, go back to n)

• ARQ block size negotiated at connection setup (depends upon the type of service, expected delay, etc.)

• ARQ block cannot be fragmented

• Monitors Rx packets and requests retransmission if found in error(s)

• Protocol overhead and processing resource burden

• Not used in VoIP applications

• Ineffective in broadcast system

• Configurable enable/disable function

• Latency impact


HARQ, Hybrid ARQ (16e)

• ARQ discards the previously transmitted data while HARQ combines the previous and retransmitted data to gain time diversity

• Uses dedicated ACK channel and PHY functions to implement a stop & wait protocol

• Makes use of the faster responding physical layer

• Complement to FEC

• HARQ combines ARQ with FEC such as convolutional or turbo codes.

– Transmitter sends a coded block. If transmission cannot be recovered by the decoder then: 1. coded data blocks are stored at the receiver. 2. retransmission is initiated. When additional coded data block is received: 1. both coded data blocks are combined and fed to decoder, 2. adds incremental redundancy and hence improves probability of recovering the data

• Greatly increases the data rate when SNR is very low, hence increases the coverage

• Typically increases the ideal BLER (block error rate) operating point by about a factor of 10

• Latency impact


Adaptive Burst Profiles

• Burst profiles are transmitted in decreasing robustness

– Modulation and or FEC

• Dynamically throttle up or down according to the link

conditions

– Burst by burst, per subscriber station

– Trade-off capacity vs. robustness in real time

• Roughly doubles the capacity for the same cell area

• Burst profile for DL broadcast channel is well known and

robust

– Other burst profiles can be configured on the fly

– SS looks for its MAC header to receive rest of the data

– SS capabilities are recognized at registration


Frame Diagram in Time Domain

Frame n-1

Preamble FCH DL Burst

#1

DL Subframe

DL Burst

#n

UL Subframe

DL-PHY PDUUL-PHY PDU

from SS#1

UL-PHY

PDU from SS#n

UL BurstPreamble

CR CBR

Frame n

MAC

HeaderMSDU CRC

DL-MAP, UL-MAP

DCD, UCDDLFP PAD

MAC

PDU

MAC

PDU

MAC

PDUs

Frame n+1

G

A

P

G

A

P

Broadcast Message

1 ODFM Sym

with BPSK-1/2

one UL burst per

UL PHY PDU

MAC

HeaderMSDU CRC


Frame Partitioning

• Normal region: frequency-diverse sub-channels

– Time scheduling possible but no frequency-specific scheduling

• i.e., used for voice services without scheduling or for flat channels

• Band AMC region: adjacent sub-carriers

– Time and frequency scheduling possible

• Broadcast region: frequency-diverse sub-channels in simulcast mode

– Borrows concept of single frequency network (SFN) from DVB/DAB etc.

Pre

am

ble

Frame

Normal RegionBand AMC

Region

FCH &

DL-MAP

(signaling)

PUSC

(Cell ID Y)

Broadcast

Region

CQ

I/A

CK

Pre

am

ble

PUSC

(Cell ID Z)

FUSC

(Cell ID Z)

Normal

Region

PUSC

(Cell ID Y) AMC

Band

AMC

Region

AMC

DL Subframe UL Subframe

DL Subframe UL Subframe

Frame

NNG NAMG Guard NAMGNNGNBR


Frame structure (1)

• Frame consists of DL & UL sub-frames

– Asymmetric traffic distribution between DL and UL is always expected

• DL subframe consists of only one DL PHY PDU and followed by one or more UL

sub-frames

• UL subframe consists of:

– Contention slot for initial ranging

– Contention slot for BW requesting

– UL PHY PDUs from different SS

– Each UL PHY PDU consists of UL preamble and UL burstFrame n-1

DL PHY

PDU

Frame n Frame n+1 Frame n+1

Contention

slot A

Contention

slot B

UL PHY

burst 1

UL PHY

burst n

TDM signal in

DL

For initial

ranging

For BW

requests

TDMA burst from different SSs

(each with its own preamble)

Adaptive


Frame Structure (2)

• DL: One transmitter and multiple receivers (multiplexed TDM)

• UL: Several transmitters and one receiver (TDMA)

– In UL, all transmitters have unique time and frequency offset, thus, UL system

design is more difficult than the DL

– The SSs are accurately synchronized such that their transmission do not

overlap each other as they arrive at the BS

• 7 different frame durations (2.5 to 20 ms, 5 ms typical)

• TTG transmit/receive transition gap between DL & UL

• RTG receive/transmit transition gap after UL before DL

• Transition gap duration is a function of channel BW and OFDM symbol time

– This is also used for Tx/Rx mode selection and PA to settle gracefully at both

ends

• Header suppression, packing and fragmentation techniques are applied in the

frame structure for efficient use of spectrum


DL/UL Sub-frame Sample Trace


TDD frame structure, Sample

OFDM symbol number (time) Timek k+1 k+3 k+5 k+7 k+9 k+11 k+13 k+15 k+17 k+20 k+23 k+26 k+30 k+31 k+33

S FCH FCH

S+1

S+2 DL burst #2 UL burst #1

UL burst #2

DL burst #1 DL burst #3 UL burst #3

UL burst #4

DL burst #4

UL burst #5

DL burst #5

S+L Fast Feedback

RangingDL subframe UL subframe

TTG RTG

D

L-M

AP

UL

-MA

P

S

ub

ca

rrie

r (f

req

ue

nc

y)

P

rea

mb

le

Pre

am

ble

U

L-M

AP

DL

-MA

P

U

L-M

AP

(co

nti

'd)


OFDMA Frame Structure

• DL-MAP and UL-MAP indicate the current frame structure

• BS periodically broadcasts Downlink Channel Descriptor (DCD) and Uplink

• Channel Descriptor (UCD) messages to indicate burst profiles (modulation and

• FEC schemes)


MAC Data Frame Format, Basic 802.3

• Flexible frame structure allows terminals to be dynamically assigned

UL & DL burst profiles according to their link conditions

Transmission order: left to right, bit serial

FCS error detection coverage

FCS generation span

PRE SFD DA SA Length/Type Data Pad FCS

7 1 6 6 4 46 to 1500 4

Field length in bytes

PRE = Preamble

SFD = Start of frame delimiter

DA = Destination address

SA = Source address

FCS = Frame check sequence


MAC

• WiMAX system can be deployed as TDD, FDD or half-duplex FDD

• A short gap between each DL & UL

• SS to remain synchronized to BS

• Each UL preceded by preamble (called short) that allows BS to sync with each

individual SS

• DL starts with preamble followed by FC header then one or more DL bursts of data

– All symbols in the FCH and DL data bursts are transmitted with equal power to

simplify the Tx & Rx design

• Mod-Coding remains the same within a burst but may change from burst to burst

• Preamble bursts are 3 dB higher than the FCH & DL data

• Burst generally starts with BPSK or QPSK then moves up depending on the

performance

B1 RTGTTG P B3B2 PP B4PP H B1 B2 B4B3

1 Frame (2.5 to 20 ms)

Downlink subframe (basestation) Uplink subframe (subscriber)


TDD Frame (1)

• DL subframe starts with preamble that helps SS to do time and

frequency synchronization and initial channel estimation

• FCH provides frame configuration information such as MAP

message length, modulation, coding and usable carriers. Multiple

users are allocated data regions within frame and it is relayed by DL-

MAP & UL-MAP.

• MAP contains burst profile for each user such as modulation,

coding. It is usually sent in BPSK-½ coding and repeated. Potential

of increasing overhead when too many users with small packet like

VoIP. Possible mitigation by use of multiple sub-MAP messages at

higher rate (if there is good SNR), compress or use broadcast MAP.

• BPSK is used for preamble, pilot & when modulating sub-carriers in

the ranging channels


TDD frame (2)

• Support multiple users on a same frame

• Varied size, type of data for several users

• Variable frame size (2-20 ms but typically 5 ms), variable

packets or fragmented packets from higher layers

• UL sub-frame has a channel quality information that is

used by scheduler (change the modulation & coding).

• Repeat pilots in lower modulation to improve recovery

• Supports Convolution, RS and optionally turbo LDPC

coding


Frame Format (3)S

ch

ed

ule Broadcast

control

DIUC = 0

4QAM

TDM

DIUC a

4QAM

TDM

DIUC c

64QAM

TDM

DIUC b

16QAM

TDM portion

TDMA portion

Pre

am

ble

Pre

am

ble

Pre

am

ble

Pre

am

ble

TDMA

DIUC d

TDMA

DIUC g

TDMA

DIUC fTDMA

DIUC e

Pre

am

ble

UL-MAP

MAC-Cntl

DL-MAP

PHY-Cntl

Burst start points

G

A

P

Tx/Rx Transition Gap

(TDD only)

For FDD

G

A

P

TDD

G

A

P


Various MAC PDU, Example

GMHOther

SH . . . CRC

MAC PDU frame carrying several fixed length MSDUs packed together

GMHOther

SHFSH MSDU Fragment CRC

MAC PDU frame carrying a single fragment MSDU

GMHOther

SHFSH PSH . . . CRC

MAC PDU frame carrying several variable length MSDUs packed together payload

GMHOther

SHFSH CRC

MAC PDU frame carrying ARQ payload

GMHOther

SHPSH PSH . . . CRC

MAC PDU frame carrying ARQ and MSDU payload

GMH CRC

MAC management Frame

CRC: Cyclic redundancy check GMH: Generic MAC Header

FSH: Fragmentation Subheader PSH: Packing Subheader

PDU: Packet Data Unit SH: Subheader

Packet Fixed

Sized MSDU

Packet Fixed

Sized MSDU

Packet Fixed

Sized MSDU

Variable Sized

MSDU

Variable Sized MSDU

or Fragments

MAC Management Message

ARQ Feedback

ARQ FeedbackVariable Sized

MSDU or Fragment


DL Subframe (1)

• The first DL burst contains

– DL map (DL MAP)

• DL MAP always refers to current frame

– UL map (UL MAP)

• UL MAP may be broadcasted one frame ahead

– DL channel descriptor (DCD)

– UL channel descriptor (UCD)

• DL bursts are broadcasted in order of decreasing robustness

BPSK> QPSK> 16QAM> 64QAM

• A SS listens to all bursts it is capable of decoding

• A SS does not know which DL burst (s) contain(s)

information sent to it


TDD Downlink Sub-frame (2)

• DL subframe starts with

– Preamble

– FCH, Frame control header

• DIUC: downlink interval usage code

• TTG/RTG, this gap is an integer number of PS (physical slot = 4 modulation

symbols) durations and starts on a PS boundary

• A portion of the DL subframe can be designated as zone for STC and AAS

applications

Preamble

DL PHY

PDU

FCH DL burst 1 DL burst n

Contention

slot A

Contention

slot B

UL PHY

burst 1

UL PHY

burst n

DL-MAP, UL-MAP

DCD, UCD

MAC

PDUs


DL Map Message (3)

• DL-MAP message defines usage of DL and contains

carrier-specific data

–DL allocation can be of broadcast, multicast and unicast

• DL-MAP is the first message in each frame

• Decoding is very time-critical

–Typically done in hardware

• Entries denote instant when the burst profile change


Typical Uplink Sub-frame (1)

• Initial maintenance opportunities

– Ranging (a procedure for MS to gain access to the BS)

– To determine network delay and to request power or profile changes

– Collisions may occur in this interval

• Request opportunities

– SSs request bandwith in response to polling from BS

– Collisions may occur in this interval as well

• Data grants period

– SSs transmit data bursts in the intervals granted by the BS

– Transition gaps between data intervals for synchronization purposes

– Any of these burst classes may be present in any given frame

• in any order and any quantity (limited by the number of available PSs) within

the frame

• at the discretion of the BS UL scheduler as indicated by UL-MAP


UL Subframe Structure (2)

• The SSs transmit in their assigned allocation using the burst profile specified by

the UIUC (UL interval usage code) in the UL-MAP entry granting them bandwidth

• UL subframe starts with

– Contention slot for initial ranging requests

– Contention slot for bandwidth request messages

SS transition

gap

Initial

maintenance

opprtunities

(UIUC = 2)

Bandwidth

request

collisionAccess

burst

Rx/Tx transition

gap (RTG)

Request

contention

opportunities

(UIUC = 1)

SS 1

scheduled

data

(UIUC = i)

SS N

scheduled

data

(UIUC = j)

Bandwidth

requestcollisionAccess

burstGap

Tx/Rx transition

gap (TTG)

Pre

am

ble

Pre

am

ble

Pre

am

ble


UL Transmission

• UL is considered to be invited transmission and is more complicated than

the DL

• Transmissions in initial ranging slots

– Ranging Requests (RNG-REQ)

– Contention resolved using truncated binary exponential back-off algorithm

• Transmissions in contention slots

– Bandwidth requests

– Contention resolved using truncated binary exponential back-off algorithm

• Each of these contention slots is further divided into minislots

• Bursts defined by UIUCs (UL interval usage code) by BS and the SS

adapts and adjusts accordingly

• Transmissions allocated by the UL-MAP message

• All transmissions have synchronization preamble

• Ideally, all data from a single SS is concatenated into a single PHY burst


UL Physical Layer

• The UL transmission convergence sub-layer is identical to

the DL one. The UL PMD (physical media device) layer

coding and modulation are as follows:

– Three classes of bursts transmitted during the UL sub-frame:

• Burst transmitted in contention opportunities reserved for initial maintenance

• Burst transmitted in contention opportunities provided by multicast and

broadcast polls

• Bursts transmitted in intervals specifically allocated to individual SS

• All UL transmissions are made according to the UL burst

profiles, specified by the BS

• Each UL burst begins with an uplink preamble


UL Channel Descriptor

• Defines uplink burst profiles

• Sends regularly

• All UL burst profiles are acquired

• Burst profiles can be changed on the fly

• Establishes association between UIUC (UL interval

usage code) and actual PHY parameters


UL-MAP Message

• UL-MAP message defines usage of the UL

• Contains the “grants”

• Grants addressed to the SS

• Time given in mini-slots (A unit of UL BW allocation

equivalent to n physical slots, where n = 2^m, m is an

integer ranging from 0 through 7)

–unit of UL bandwidth allocation

–2m physical slots

• in 10-66GHz PHY physical slot is 4 modulation symbols long

• Time expressed as arrival time at BS


UL Contention Resolution

• Based on a truncated binary exponential backoff

– The initial/maximal backoff window is controlled by the BS

• The SS shall randomly select a number within its backoff window

– This random value indicates the number of contention transmission

opportunities that the SS shall defer before transmitting

• For bandwidth requests, if the SS receives a Unicast Request IE or Data Grant

Burst Type IE at any time while deferring for this CID, it shall stop the contention

resolution process

Transmission

Opportunity #1

One Request IE

Transmission

Opportunity #2

Transmission

Opportunity #3

Preamble

(2 minislots)

BW Req Message

(3 minislots)

SSTG

(3 minislots)


Downlink Preamble

• DL subframe starts with two OFDM symbols containing preamble (called long)

– Symbol 1 contains 50 subcarriers (every fourth subcarrier with no data or pilot subcarriers resulting in wider adjacent channel spacing), QPSK

– Symbol 2 contains 100 subcarriers (every even subcarriers with no data or pilot subcarriers resulting in wider adjacent channel spacing), QPSK

– Transmitted at 3 dB higher level than all others to make Rx to easily recover information

– Preamble followed by FCH then one or more data Symbols

• All symbols in FCH and DL data burst are transmitted with equal power

– Same modulation is kept within the burst but it may change from burst to burst

– Data initially starts out with low level modulation then gradually increases depending on RSL and CINR

– Generally inserts a few short mid-preamble in extremely long DL burst


OFDM Frame Structure Diagram

• Variable number of subcarriers for OFDMA

50 CXR BPSK

Preamble

1 symbol

200 carriers, BPSK/ QPSK/16QAM/64QAM

Preamble

1 symbol

FCH

1 symbol

DL

Burst #1

Preamble

1 symbol

UL

Burst #1

UL

Burst #2

DL Subframe

variable number of OFDM symbols

Frame #2 Frame #nFrame #1

100 CXR BPSK BPSK

DL

Burst #2

DL

Burst #n

UL

Burst #n

200 carriers, BPSK/ QPSK/16QAM/64QAM100 CXR BPSK

UL Subframe

variable number of OFDM symbols

Can contain DL MAP

(if FCH is too small, DL BURST is used)Long preamble

2 symbols

G

A

P

Ranging BW

CDMA


Downlink/Uplink Preamble

• UL subframe starts with short single OFDM Symbol that

synchronizes the BS to the SS

– Preamble (called short) consists of 100 even number subcarriers

– Uses QPSK-1/2 modulation

– Same power as data sub-carriers

– Symbol contains no data or pilot subcarriers

• Following DL preamble is a FCH (single OFDM Symbol of

BPSK, 88 bits of overhead data that describes critical

system decoding information such as BS ID and DL burst

profile). DL burst contains one or more Symbols. Each

symbol in the DL burst contains 12 to 108 bytes of payload

data, depending on the modulation & coding types


Preamble Plot

T2T1t10t1 t7t6t5t3 t4 t8 t9t2 GI2 Data 1 Data 2GI GIGISignal

8+8 = 16 us, Preamble10x0.8 = 8 us 2x0.8 + 2x3.2 = 8 us 0.8+3.2 = 4 us 0.8+3.2 = 4 us

0.8+3.2 = 4 us

OFDM Service+DataRate

Length

Channel and Fine

Frequency Offset

Estimation

Coarse Freq

Offset Estimation

Timing Sync

Signal Detect, AGC,

Diversity Selection

Preamble

Data


Training Symbol Structure

• Flexible usage in OFDMA and MIMO

User DataPreamble

1 OFDM symbol 3 OFDM symbol

OFDM Packet (time domain)

Preamble-based

Fre

qu

en

cy

Time

Data SymbolTraining Symbol

2D Time-

Frequency

Interpolation

1D Frequency

Interpolation

1D Time

Interpolation

Pilot-Based

Time

Fre

qu

en

cy

{ {


DL Physical Layer, (1)

• Available bandwidth in DL direction: physical slots

• Available bandwidth in UL direction: mini-slots (mini slot

length = 2^m physical slots where m is an integer ranging

from 0 through 7)

• Number of physical slots with each frame is a function of

symbol rate (20 Mbps: 5000 PHY. Slots within 1 ms frame)

• DL frames can be TDD (the subframe contains preamble

for synchronization and equalization, frame control section

to see where bursts begin, and data) and FDD (preamble,

frame control section and TDM portion organized into

bursts transmitted in decreasing order of burst profile

robustness)


DL & UL Physical Layers, (2)

• Physical layer allows for flexible spectrum usage and

support, both TDD and FDD

• Burst transmission format is framed to support

adaptive burst profiling (modulation and coding

schemes can be adjusted individually to each SS)

• The UL physical layer is based on a combination of:

–TDMA (time division multiple access)

–DAMA (demand multiple access)



• UL channel is divided into a number of time slots

– Its various number is controlled by the MAC layer in the BS

• DL channel is a TDM (information for each subscriber is

multiplexed onto a single stream of data)

• The downlink physical layer includes a transmission

convergence sub-layer which helps the receiver to identify

the beginning of a MAC frame.

• The PHY layer performs randomization, FEC encoding and

modulation (QPSK, 16-QAM or 64-QAM)



• The UL PHY layer is based upon TDMA burst transmission

• Each burst is designed to carry variable length MAC

frames

• PMD layer performs randomization, FEC encoding and

modulation

• Frame duration: 2.5 to 20 ms

• Each frame contains a DL sub-frame and an UL sub-frame

• In the TDD case, UL and DL transmissions share the

same frequency but are separated in time

• In FDD case, both transmissions occur at the same time

but the channels are on separated frequencies


Protocol Architecture

• IEEE 802.16 Protocol Architecture has 4 layers: Convergence, MAC,

Transmission and physical, which can be mapped into two lowest

OSI layers: physical and data link

Data Link

Physical

Session

Transport

Network

Application

Presentation

OSI Reference

Model

Medium

MAC Convergence

sublayer

MAC Transmission,

Privacy sublayer

MAC

Physical Layer

Back haul

Virtual

point to point

Frame ralay

ATM

Ethernet, 802.1Q

Internet Protocol

IPATMDigital audio/

video multicast

Digital telephonyBridged LAN

Packing,

Fragmentation,

ARQ,

QoS

Authentication,

Key Exchange,

Privacy (encyption)

OFDM, ranging,

power control,

DFS, Tx, Rx

OSI

physical

layer

OSI

data

layer

1

3

2

4

5

7

6

optical

fiber

Layers

Network

coaxial cable

DQDBToken RingEthernt

IEEE 802.3

twisted

pair

Internet Potocol IP

Transmission Control

Protocol TCP

User Data Protocol

UDP

NFS

XDR

RPC

Name

ServerSMTPHTTP

Application

Internet

Transport

FTP

TCP/IP

protocols

TCP/IP

model


Protocol Structure

• CS: All functions that are specific to a higher layer

protocol

– Receives and adapts higher layer PDUs to MAC CPS

– Classifies SDUs based on MAC address, VLANs,

priorities

– Assigns service flow ID (SFID) and connection

identifier

– Maps data to a CID

• CPS: Provides the core MAC functionality

– Fragmentation and reassembly of large MAC SDUs

– Packing and unpacking of several small MAC SDUs

– QoS control and scheduling

– Bandwidth request and allocation

– Automatic repeat Query (ARQ)

MAC Convergence

sublayer

(CS)

MAC Transmission,

Privacy sublayer

MAC Common Part

Sublayer (CPS)

Physical Layer

(PHY)

PHY

MAC


Security & PHY Sub-layer

• Provides authentication, secured key exchange,

encryption support ARQ scheme

• Supports two protocols:

– Encapsulation protocol for data encryption

• Defines cryptographic suites such as pairings of

data encryption and authentication algorithms

• The rules for applying those algorithms to a MAC

payload

– Privacy key management protocol

• Describes how the BS distributes to SS

• PHY Sub-Layer

– Single carrier, 11-66 GHz

– MC, NLOS, below 11GHz, ARQ, AAS & MIMO

– S-OFDMA, NLOS, H-ARQ, Fast feedback, Handover

MAC Convergence

sublayer

(CS)

MAC Transmission,

Privacy sublayer

MAC Common Part

Sublayer (CPS)

Physical Layer

(PHY)

PHY

MAC


Unicast Polling

1. BS allocates sufficient space for the SS in the uplink subframe

2. SS uses the allocated space to send a BW request

3. BS allocates the requested space for the SS (if available)

4. SS uses allocated space to send data

Poll(UL-MAP)

RequestAlloc(UL-MAP)

Data

BS SS


ATM Convergence Sub-layer

• Supports for:

–VP (Virtual Path) switched connections

–VC (Virtual Channel) switched connections

• Support for end-to-end signaling of dynamically

created connections

• SVCs

• Soft PVCs

• ATM header suppression

• Full QoS support


Packet Convergence Sub-layers

• Initial support for Ethernet, IPv4 and IPv6

• Payload header suppression

–Generic plus IP specific

• Full QoS support

• Possible future support for:

–PPP

–MPLS

–etc.


MAC Addressing

• SS has 48-bit IEEE MAC address

• BS has 48-bit Base Station ID

–Not a MAC address

–24-bit operator indicator

• 16-bit Connection ID (CID)

–Used in MAC PDUs (packet data units)


MAC PDU Transmission

• MAC communicates using MAC protocol data units (MPDUs) that are carried by the PHY

• MAC PDUs (packet data units) are transmitted in PHY bursts

• A single PHY burst can contain multiple Concatenated MAC PDUs

• The PHY burst can contain multiple FEC blocks

• MAC PDUs may span FEC block boundaries

• The TC (transmission conversion) layer between the MAC and the PHY allows for capturing the start of next MAC PDU in case of erroneous FEC blocks

– The TC PDU format allows resynchronization to the next MAC PDU if the previous block had irrecoverable errors

– Without the TC layer, a receiving SS or BS would potentially lose the entire remainder of a burst when an irrecoverable bit error occurred

– Performs conversion of variable length MAC PDUs into fixed length FEC blocks (plus possibly a shortened block at the end) of each burst


MAC PDU Transmission

PDU 1 PDU 5PDU 4PDU 3

SDU 2

PDU 1

MAC Message SDU 1

PDU 2

FEC 1 FEC 2 FEC 3P

MAC

PDUs

PHY

Burst

Preamble FEC blockMAC PDUs

Packing

Concatenation

Shortened

Fragmentation


MAC PDU Format

• There are types of MAC header (generic or BW request)

– Both generic and BW request MAC headers are fixed length and 6 bytes long

• One or more MAC sub-headers may be part of the payload

• The presence of sub-headers is indicated by a type field in the Generic MAC header

• Size varies from 6 byte to 2047 bytes

• Flexibility creates transmission inefficiency

ms

b

Generic MAC Header Payload (optional) CRC (optional)

6 bytes 4 bytes0 to 2041 bytes


GMH, Generic MAC Header

• The GMH is used for transmit data or MAC messages and

may optionally have one or more appended sub-headers

– Fragmented Sub-header (2 bytes, optionally 1 byte)

– Packing (3 bytes, optionally 2 bytes)

– Grant Management (2 bytes)

– Mesh Sub-header (2 bytes)

– Fast-Feedback-Allocation (1 byte)

– Extended Sub-header (variable length)

– The subheader can occur only once per MAC PDU except for the

Packing subheader, which may be inserted before each MAC SDU

packed into the payload


BWH, Bandwidth Header

• The BWH is used by the SS to request more

bandwidth on UL

–ARQ Fast –Feedback and Grant Management sub-

headers are used to communicate ARQ and bandwidth

allocation states between the BS and SS

–Fragmentation and Packing sub-headers are used to

utilize the bandwidth allocation efficiently


PHSF, Payload Header Suppression Format

• If PLHS is enabled at MAC connection, each MAC SDU is

prefixed with a PHSI (payload header suppression index),

which references the PHSF (payload header suppression

field). The classifier (located at the sending entity) uniquely

maps the packets to its associated PHS Rule. The receiving

entity uses the CID and the PHSI to restore the PHSF.

Useful portion Payload

Payload Header

Useful portion

PHSI

PayloadPHSF


Header Suppression for VoIP over WiMAX

• The protocols used in addition to WiMAX are RTP, UDP and

IPv6

Application Layer

UDP

Voice Payload

Voice Payload

Voice Payload

Voice Payload

Voice Payload

Voice PayloadRTP

IPv6

MAC

PHY


Header Suppression for VoIP over WiMAX, 2

• Header sizes of each of these layers:

– between 12 to 72 bytes for RTP

– 8 bytes for UDP

– 40 bytes for IPv6

• the total length of RTP/UDP/IPv6 header is between 60 and 120 bytes

• PHS suppresses repetitive (redundant) parts due to the

higher layers in the payload header of the MAC SDU

• The receiving entity restores the suppressed parts

• Its is the responsibility of the higher-layer service entity to

generate a PHS Rule



• A Payload Header Suppression Index (PHSI), an 8-bit field

which references the Payload Header Suppression Field

(PHSF) that has been used for header suppression

– The PHS rule has also a Payload Header suppression Mask (PHSM)

option to allow the choice of bytes of PHSF that can not be suppressed

Payload

Packet Header

Reconstruction

(using PHSI and CID)

Sender

PHSF

PHSM

Air

Interface

Receiver

PHSM

0 1011

X E‟XC‟A‟

B D

B‟ E‟D‟C‟A‟

B EDCA

0 1011

X E‟XC‟A‟PHSF

Payload

Payload

Packet

Transmission

PHSM

PHSF PHSI 1 byte

MAC header

A-E=curent in

A‟-E‟=cached

X=don‟t care

PHSS=5

=verify

=assign



• Number of suppressed bytes per header:

– IPv6: 37 bytes

–UDP: 4 bytes

–RTP: 4 bytes.

• The RTP/UDP/IPv6 Header drops from 60 bytes to

15 bytes (45 Header bytes or less are suppressed)


Fragmentation

• Partitioning a MAC SDU into fragments then transporting in multiple MAC PDUs

• Longer packet increases probability of losing a packet and hence initiate

retransmission

• Allows better packing of MAC SDUs into the available OFDM freq-time resources

by using all data subcarriers in each OFDM symbol

• Each connection can be in only a single fragmentation state at any time

• Contents of the fragmentation sub-header:

– 2-bit Fragmentation Control (FC)

• Unfragmented, Last fragment, First fragment, Continuing fragment

– 3-bit Fragmentation Sequence Number (FSN)

• Required to detect missing continuing fragments

• Continuous counter across SDUs

• Fragmentation is an optional feature that improves the link efficiency


Packing, (1)

• A process of combining multiple MAC SDUs (or fragments

thereof) into a single MAC SDU

• Allows better packing of MAC SDUs into the available OFDM

frequency-time resources by using all data subcarriers in each

OFDM symbol

– Can, in certain situations, save up to 10% of system bandwidth

• On connections with variable length MAC SDUs

– Packed PDU contains a sub-header for each packed SDU (or

fragment thereof)

• On connections with fixed length MAC SDUs

– no packing sub-header needed

• Packing and fragmentation can be combined


Packing Fixed-Length SDUs, (2)

....

MA

C H

eader

LE

N =

n*k

+j

fixed length

MAC SDU

length = n

fixed length

MAC SDU

length = n

fixed length

MAC SDU

length = n

k MAC SDUs

fixed length

MAC SDU

length = n


Packing Variable Length SDU, (3)

• 2 Byte packing sub-header before each SDU

– Length of the SDU: 11 bits

– Fragmentation control (FC): 2 bits

– Fragmentation sequence number (FS): 3 bits

....

PS

H

Length

= a

+2

variable

length

MAC SDU

length = a PS

H

Length

= b

+2

k MAC SDUs

PS

H

Length

= c

+2

variable

length

MAC SDU

length = b

variable

length

MAC SDU

length = cMA

C H

eader

LE

N =

j

Typ

e =

00001X

b


Packing with Fragmentation

r MAC SDUs

unfragmente

d

MAC SDU

length = 0

PS

H

FC

= 0

0,

FS

N =

n+

2

Length

= c

+1

unfragmente

d

MAC SDU

length = c

first fragment

of MAC SDU

length = d'

PS

H

FC

= 1

0,

FS

H =

x+

y+

1

Length

= d

+2

PS

H

FC

= 0

0,

FS

N =

n+

1Last

fragment

of MAC SDU

length = a

MA

C H

eader

LE

N =

y1

Typ

e =

00001X

bP

SH

FC

= 0

1,

FS

N

=x'

Length

= a

+2

s-I+1 MAC SPUs

. . .

FS

H

FC

=11,

FS

H =

x+

y1

Length

= f

+1

Continuing

fragment of

MAC SDU

length=f

MA

C H

eader

LE

N=

y45

Type =

00010X

b

FS

H

FC

=11,

FS

N=

x+

5

Length

= q

+1

continuing

frragment of

MAC SDU

length=g

MA

C H

eader

LE

N =

y2

Type =

00010xb

FS

H

FC

= 1

1,

FS

N =

x+

y

Length

= e

+1

Continuing

fragment of

MAC SDU

legth =e

MA

C H

eader

LE

N =

y3

Type =

00010xb

r MAC SDUs

PS

H

FC

=0

1,

FS

N=

x+

s+

1

Len

gth

=b

+2

Last

fragment

of MAC SDU

length=h

Unfragmente

d

MAC SDU

length=k

PS

H

FC

=0

0,

FS

N=

x+

s+

2

len

gth

=k+

2

PS

H

FC

=D

0,

FS

N=

x+

s+

3

Len

gth

=1

+2

PS

H

FC

=D

0,

FS

N=

x+

s+

2

Len

gth

=1

+2

unfragmente

d

MAC SDU U

length=f

unfragmente

d MAC SDU

length=f

MA

C H

ea

de

r

LE

N=

y5

Typ

e=

00

00

1xb


OFDMA, Typical TDD Time Frame

• Pilot, null and DC subcarriers are not shownOFDMA Symbol Number Time

k+0 k+1 k+2 k+3 k+4 k+7 k+9 k+11 k+13 ... k+17 k+20 k+23 ... ... k+31 k+33

S+0 FCH Ranging FCHS+1 SubchannelsS+2

DL burst #2 UL burst #1

UL burst #2

DL burst #3

DL burst #4 UL burst #3

DL burst #1 DL burst #5

UL burst #4

S+N UL burst #5DL UL

TTG RTG

P

rea

mb

le

S

ub

ch

an

ne

l L

og

ica

l N

um

be

r

P

rea

mb

le

DL

-MA

P

D

L-M

AP

U

L-M

AP

UL

-MA

P


Chain Transmission

• Randomization and FEC coding in UL are identical to the corresponding in the DL

• The type of modulation and the power adjustment rules are set by the BS

• QPSK, 16QAM and 64QAM are mandatory and the 256QAM is optional

• The mapping of bits to symbols are identical to those in the DL

• Systems shall use Nyquist square-root raised cosine pulse shaping (role off factor 0.25)

• A frame duration of 5 ms is typically used as the compromise between transport efficiency and latency

• Must be able to compensate at most 20 dB/s for 40 dB range

– Actual power control algorithm is left to vendors

– 0.25 dB resolution


Down Link Transmission

• Two kinds of bursts: TDM and TDMA

• All bursts are identified by a DIUC

– Downlink Interval Usage Code

• TDMA bursts have resync preamble

– allows for more flexible scheduling

• Each terminal listens to all bursts at its operational IUC, or at a

more robust one, except when told to transmit

• Each burst may contain data for several terminals

• SS must recognize the PDUs with known CIDs

• DL-Map message signals DL usage


Downlink Channel Descriptor

• Used for advertising DL burst profiles

• Burst profile of DL broadcast channel is well known

• All others are acquired

• Burst profiles can be changed on the fly without

interrupting the service

–Not intended as “super-adaptive” modulation

• Establishes association between DIUC and actual

PHY parameters


Burst Profiles

• Each burst profile has mandatory exit

threshold and minimum entry threshold

• SS allowed to request a less robust

DIUC (DL interval usage code) once

above the minimum entry level

• SS must request fall back to more

robust DIUC once at mandatory exit

threshold

• Requests to change DIUC done with

DBPC-REQ (DL burst profile change

Req.) or RNG-REQ (range Req.)

messages

Overlap

C / (

N+

I)

Burst Profile Z

Burst Profile

Overlap

Burst Profile Y

0


UL-AAS Beam Response Message

• Message contains a total of 48 bits

– Management message type, 8 bits

– Frame number, 8bits

– Feedback request number, 3 bits

– measurement report Types, 2 bits

– Resolution parameter, 3 bits

– Beam bits mask, 4 bits

– reserved, 4 bits

– RSSI mean value, 8 bits

• Quantized in 2 dB increment in range from -48 to -110 dBm

– CINR mean value, 8 bits

• Quantized in 1 dB increment in range from 10 to 53 dB


Admission Control, Scheduling and Link

Adaptation

• Admission Control

– Ensure that new flows do not degrade the quality of established flows

• Scheduling

– BS schedules usage of the air link among the subscribers per specific QoS

– Packet schedulers at the BS and subscribers gives transmission

opportunities to multiple connection queues

• Link Adaptation

– BS determines the contents of the DL and UL portions of each frame

– BS determines the appropriate burst profile (code rate, modulation level and

so on) for each subscriber

– BS determines the BW requirements of the individual subscriber based on

the service classes of the connections and on the status of the traffic

queues at the BS and SS


The QoS Object Model

PDUSFID

[Sevice Class]

[CID]

Payload

N

1

0, 1

1N 0, 1Connection

Connection ID

QoS Parameter Set

Service FlowSFID

Direction

[CID]

[Provisioned QoS ParamSet]

[AdmittedQoSParamSet]

[ActiveQoSParamSet]

Service Class

Sevice Class Name

QoS Parameter Set


QoS Mechanisms

• Classification

–Mapping from MAC SDU fields (e.g., destination IP

address or TOS field) to CID and SFID

• Scheduling

–Downlink scheduling module

• Simple, all queues in BS

–Uplink scheduling module

• Queues are distributed among SSs

• Queue states and QoS requirements are obtained through BW

requests

–Algorithms not defined in standard


QoS Control

.Control Plane

BW request

AC

UL Map

Ctrl/mng

channels

Non

UGS

Conn_req

Conn_rsp

UGS

UL

Scheduling

Applications

Connection Classifiers

BE

CIDs

nrtPS

CIDs

rtPS

CIDs

ertPS

CIDs

UGS

CISs

UL Data packets (data channels)

BSSS

Control

Plane

Priority Scheduler


QOS Mechanism

Packet

Construction

MAC CS MAC CPS MAC CPS MAC CS

Subscriber Station Base Station

VoIP

MPEG

TFTP, FTP

HTTP

New

Connection

TDM/

Voice

Connection

Request

Connection

Response

Implicit

Request

Piggyback

Request

Unicast

Polling

Contention

Based Polling

Data Traffic

Connection

Request

Generator

Admission

Control

UL BW

Grant

Generator

Slot

Allocation

UL BW

Request

Generator

UL BW

Grant

Processor

Packet

Classifier

Connection

Request

CID#6 (UGS)

CID#7 (ert-PS)

CID#8 (rt-PS)

CID#9 (nrt-PS)

CID#10 (BE)

DL Traffic

Processor

VoIP

MPEG

TFTP, FTP

HTTP

TDM/

Voice CID#1

CID#2

CID#3

CID#4

CID#5

Packet

Classifier

VoIP

MPEG

TFTP, FTP

HTTP

TDM/

VoiceCID#1

CID#2

CID#3

CID#4

CID#5

DL Traffic

Processor

DL-MAP

Generator

Data Traffic

DL-MAP

UL-MAP

CID#6

CID#7

CID#8

CID#9

CID#10

Packet

Re-Construction

VoIP

MPEG

TFTP, FTP

E-mail

TDM/

Voice

E-mail

HTTP

E-mailE-mail


5-Types of Scheduling Services

• Unsolicited Grant Service (UGS)

– for constant bit-rate (CBR) or CBR-like service flows (SFs) such as T1/E1

• Extended real-time Polling Service (ertPS)

– for real time variable bit rate in an unsolicited manner and has less

request/grant overhead than the rtPS, VoIP services with silent

suppression

• Real-time Polling Service (rtPS)

– for rt-VBR-like SFs on periodic basis such as MPEG video

• Non-real-time Polling Service (nrtPS)

– for nrt SFs with better than best effort services such as bandwidth-intensive

file transfer (FTP)

• Best Effort (BE)

– for best-effort traffic with no minimum service level required


UGS, Unsolicited Grant Service

• Supports services that generate fixed size data packets on periodic

basis

– T1/E1 services or voice over IP without silence suppression

• No need for explicit BW requests

– Low overhead

• Offers fixed size grants on a real time periodic basis, which

eliminates overhead and latency of SS requests

• No unicast request opportunity provided

• May include a grant Management (GM) sub-header containing

– Slip indicator: indicates that there is a backlog in the buffer due to clock skew or

loss of maps

– Poll-me bit: indicates that the terminal needs to be polled (allows for not polling

terminals with UGS-only services)


ertPS, extended-real-time Polling System (for

16e)

• An extended real-time polling service (ertPS) combines UGS

& rtPS

– Supports VoIP with silence suppression

• Periodic unsolicited grants similar to UGS for data

transmission or for requesting additional BW

• Unlike UGS, allocations are not fixed and may change over

time (on/off UGS)

• Default size is based on maximum sustained traffic rate

• MS may request a change in allocation size, using grant

management sub-header or other means


rtPS, real-time Polling System

• Supports real time flows with variable size data packets on periodic basis such as MPEG video

• Provides periodic request opportunities

– SS specifies the frame size in the BW request in response

– Unicast request opportunities which meets the flow’s real time needs and allows SS to specify the size of desired grant

• Prohibited from using any contention requests

• More overhead, but more flexible and provides optimum data transport efficiency than UGS

• Terminal polled frequently enough to meet the delay requirements of the SFs

• Bandwidth requested with BW request messages (a special MAC PDU header)

• May use Grant Management sub-header

– new request can be piggybacked with each transmitted PDU (protocol data unit)


nrtPS, non-real-time Polling System

• Works like rt-polling except that polls are issued less frequently

• Combines periodic and contention request opportunities

• Base station issues unicast polls on the order of a second or less

• SS may also use contention request opportunities

• Can be used for delay tolerant traffic

– No delay or jitter guarantees

• Intended for non-real-time service flows with better than best effort service

– e.g., bandwidth extensive file transfer


– New request can be piggybacked with each transmitted PDU


BE, Best Effort

• For best-effort traffic in the UL

– Generic Data

– e.g., HTTP, SMTP, etc.

– No QoS guaranteed

• Leftover or unused allocation may be used by SSs

• SS/MS allowed to use contention request opportunities

• BS may allocate unicast opportunities

– Depending on policy

– No guarantees


– New request can be piggybacked with each transmitted PDU


QoS

Downlink

Data

Bandwidth

Requests

Translator

UGS

ertPS

nrtPS

rtPS

BE

Second Phase

Proportionating

Two-Phase

Proportionating

Determine

UL/DL subframe

1st Phase

Proportionating

Qu

eq

ue

s w

ith

ou

t

La

ten

cy

TranslatorAssign slots

to SSs

Uplink

FrameDownlink

Frame

Write in

UL-MAP

Write in

DL-MAP

Assign slots

to SSs

Convergence

Sublayer

Assign slots

for queques

Downlink Uplink

Two-Phase

Proportionating

PHY Layer


Service Flows and QoS, example

• Priority + EDF + WFQ + RR - combined model

BE

CIDs

nrtPS

CIDs

rtPS

CIDs

ertPS

CIDs

UGS

CISs

Prio

rity

Sch

ed

ule

r

UL Map

Fixed Bandwidth

Fixed Bandwidth

with Silent Detect

Earliest

Deadline First

(EDF)

Weighted

Fair Queuing

(WFQ) or WRR

Round Robin


QoS Summary

QoS Category Applications QoS Specifications

UGS

Unsolicited Grant Service

VoIP, T1/E1

Fixed data rate

Maximum sustained rate

Maximum latency tolerance

Jitter tolerance

ertPS

Extended Real Time Packet

Service

Voice with activity detection

(VoIP)

Variable data rate


Minimum reserved rate


Jitter tolerance, Traffic priority

rtPS

Real Time Packet Service

Screaming Audio and MPEG

Video




Committed burst size, Traffic priority

nrtPS

Non-Real Time Packet

Service

File Transfer Protocol

(FTP)



Traffic priority

BE

Best Effort Service

General data transfer, Web

Browsing, etc...


Traffic priority


Scheduling Types

Scheduling

Type

Piggy Back

Request

BW stealing Polling

UGS Not Allowed Not Allowed PM bit is used to request unicast

poll for bandwidth needs of non-

UGS connections

ertPS Allowed Allowed for GPSS

(Grant per SS)

Scheduling only allows unicast

polling

rtPS Allowed Allowed for GPSS Scheduling only allows unicast

polling

nrtPS Allowed Allowed for GPSS Scheduling may restrict a service

flow to unicast polling via the

transmission/request policy;

otherwise all forms of polling are

allowed

BE Allowed Allowed for GPSS All forms of polling allowed


Request / Grant Scheme

• Self Correcting

– No acknowledgement

– All errors are handled in the same way, i.e., periodical aggregate

requests

• Bandwidth Requests are always per Connection

• Grants are either per connection (GPC) or per Subscriber

Station (GPSS)

– Grants (given as durations) are carried in the UL-MAP messages

– SS needs to convert the time to amount of data using information

about the UIUC


BW Requests

• Comes from the Connection

• Several kind of requests:

– Implicit requests (UGS)

– No actual messages, negotiated at connection setup

– BW request messages

• Uses the special BW request header

• Requests up to 32 kb with a single message

• Incremental or aggregate, as indicated by MAC header

– Piggybacked request (for non-UGS services only)

• Presented in GM sub-header and always incremental

• Up to 32 kb per request for CID

– Poll-Me bit (for UGS services only)

• Used by the SS to request a bandwidth poll for non-UGS services


BW Allocation and Burst Placement

• .

Flow 1 Flow 2 Flow 3 Flow 4 Flow 5

Flow 5Flow 4Flow 3Flow 2Flow 1

Flow Queque in DL

Flow 1

Flow 2

UL

Subframe

Flo

w 4

Flow 3

Flow 5


Bandwidth Request and Allocation, (1)

• SSs may request BW in 3 ways:

–Uses ”contention request opportunities” interval upon

being polled by the BS (multicast or broadcast poll)

• Contention is resolved by using back off resolution

–Sends a standalone MAC message called ”BW request” in

an already granted slot

• Due to the predictable signaling delay of the polling scheme,

contention-free mode is suitable for real time applications

–Piggybacks a BW request message on a data packet



• BS grants/allocates bandwidth in one of two modes:

– Grant Per Subscriber Station (GPSS)

• BS scheduler treats all the connections from a single SS as one unit and

grants BW to the SS. An additional scheduler is employed at the SS which

determines the service order for its connections in the granted slot

• More scalable and efficient than the GPC

– Grant Per Connection (GPC)

• BS scheduler treats each connection separately and BW is expilicitly granted

to each connection

• SS transmits according to the order specified by the BS

• Decision based on requested BW and QoS requirements

vs. available resources

• Grants are realized through the UL-MAP



• DL-MAP and UL-MAP indicate the current frame structure

• BS periodically broadcasts DL Channel Descriptor (DCD) and UL Channel

Descriptor (UCD) messages to indicate burst profiles (modulation and FEC

schemes)


GPSS vs. GPC

• Bandwidth Grant per Subscriber Station (GPSS)

– Base station grants bandwidth to the subscriber station

– Subscriber station may re-distribute bandwidth among its connections, maintaining QoS and service level agreements

– Suitable for many connections per terminal; off-loading base station’s work

– Allows more sophisticated real time reaction to QoS needs

– Low overhead but requires intelligent subscriber station

– Mandatory for P802.16 10-66 GHz PHY

• Bandwidth Grant per Connection (GPC)

– Base station grants bandwidth to a connection

– Mostly suitable for a few users per subscriber station

– Higher overhead, but allows simpler subscriber station


Maintaining QoS in GPSS

• Semi-distributed approach

• BS sees the requests for each connection; based on

this, grants bandwidth (BW) to the SSs (maintaining

QoS and fairness)

• SS scheduler maintains QoS among its connections

and is responsible to share the BW among the

connections (maintaining QoS and fairness)

• Algorithm in BS and SS can be very different; SS

may use BW in a way unforeseen by the BS


SS Initialization Steps

• Scans for DL channel and establish synchronization with the BS

• Obtains transmit parameters (from UCD message)

• Performs ranging

• Negotiates basic capabilities

• Authorizes SS and performs key exchange

• Performs registration

• Establishes IP connectivity

• Establishes time of day

• Transfers operational parameters

• Set up connections


Ranging, (1)

• Procedure for MS to gain access to the BS

• Four types of ranging can be defined

– Initial Ranging for Network entry

– Periodic Ranging for synchronization

– Bandwidth requests

– HO ranging

• Single Ranging channel (multiple sub-channels) using 1 to 8 subcarriers defined by the system specified in the UCD

• Ranging process accomplished through PN codes assigned spreading for specific Ranging types

– Also known as CDMA-like (a maximum of 256 sets of 144-bit wide pseudo-noise code) ranging for OFDMA


Ranging, (2)

• For UL transmissions, times are measured at BS

• At start up, SS sends a RNG-REQ message in the

contention slot reserved for this purpose

– SS looks for initial ranging opportunities (UL-MAP) information

present in every frame

• BS measures arrival time and signal power; calculates

required advance and power adjustment

• BS send adjustment in RNG-RSP

• SS adjusts advance and power, sends new RNG-REQ

• Loop between BS & SS is continued until power and timing is

ok


Channel Acquisition

• SS scans for suitable BS DL signal

• SS Sync to this signal and searches the first DL burst

of the DL PHY PDU

–Reads the DL channel descriptor (DCD)

–Reads the UL channel descriptor (UCD)

–Learn the modulation and coding schemes used on the

carrier


Negotiation of Capabilities

• BS sends

– Power adjust information

– Timing adjust information

– CID for the basic management connection

– CID for the primary management

• SS reports its PHY capabilities on the primary management

connection

– Modulation

– Coding scheme

– Half-duplex or full-duplex operation (FDD)

• BS may deny the use of any capability reported by the

subscriber station


SS Authentication

• SS must pass authentication

• SS contains an X.509 digital certificate and the

certificate of the manufacturer

• SS sends these certificates to BS

• BS examines certificates and authenticates (or deny)

the SS

• If authentication is successful, the BS sends the

authorization key

• The AK is used both by SS and BS for securing

further information flow (subsequent key derivation)


Registration

• Registration is a form of capability negotiation

• SS sends a list of capabilities and parts of the

configuration file to the BS in the REG-REQ

message

• BS replies with the REG-RSP message

–Tells which capabilities are supported/allowed

• SS acknowledges the REG-RSP with REG-ACK

message


SS registration

• After successful authentication, the SS registers with

network

• Response from BS contains CID for a secondary

management connection

–Secondary management connection is secured

• SS and BS determines

–Capabilities related to connection set up

–Parameters required for MAC operation

– IP version used


MAC Management Connections

• Upon entering the network, the SS is assigned three management connections in each direction

• Basic management connection

– Exchange of short, time-critical MAC, radio link control management messages with minimal delay

– Used to quickly adapt to wireless environment

• Primary management connection

– Exchange of longer, more delay tolerant MAC management messages

– Authentication and connection setup

• Secondary management connection (higher layer)

– Exchange of delay tolerant IP-based messages (DHCP, SNMP, TFTP, ToD)


IP Connectivity and Configuration File

Download

• IP connectivity established via DHCP or static IP

server

• SS establishes the time of the day via the Internet

Time Protocol

• DHCP server provides the address of the TFTP

server

• Configuration file downloaded via TFTP

• Contains provisioned information

–Operational parameters


Connection(s) set up

• Secondary management connection is also used for

setting up one or more transport connections

• Transport connections carry the actual user traffic

• Service flows defines unidirectional transport of

packets between the subscriber station and BS

–service flows are characterized by a certain set of QoS

parameters

–Service flows are established using a three-way

handshaking establishment procedure


Initial Connection Setup

• BS passes Service Flow Encodings to the SS in

multiple DSA-REQ (dynamic service addition Req)

messages

• SS replies with DSA-RSP messages

• Service Flow Encodings contain either

–Full definition of service attributes (omitting defaultable

items if desired)

–Service class name

• ASCII string which is known at the BS and which indirectly specifies a

set of QoS Parameters


Privacy and Encryption

• Secures over-the-air transmissions

• Authentication (SIM, Universal SIM, removable user identity module RUIM)

– X.509 certification with RSA PKCS

– Strong authentication of SSs (prevents theft of service)

– Prevents cloning

• Data encryption

– Currently 56-bit DES in CBC mode

– IV based on frame number

– Easily exportable

• Message authentication

– Key MAC management messages authenticated with one way hashing (HMAC with SHA-1)

• Designed to allow new/multiple encryption algorithms

• Protocol descends from BPI+ (DOCSIS)


Security Associations

• A set of privacy information

–Shared by a BS and one or more of its client SSs share in

order to support secured communications

– Includes traffic encryption keys and CBC IVs

• Security Association Establishment

–Primary SA established during initial registration

–Other SAs may be provisioned or dynamically created

within the BS


SS Authorization

• Authentication and Authorization

– SS manufacturer’s X.509 certificate binding the SS’s public key to

its other identifying information

– Trust relation assumed between equipment manufacturer and

network operator

– Possibility to accommodate “root authority” if required

• Authorization Key Update Protocol

– The SS is responsible for maintaining valid keys

– Two active AKs with overlapping lifetimes at all times

– Re-athorization process done periodically

– AK lifetime (7 days) & grace timer (1 hr)


Traffic Encryption Key Management

• Two-level key exchange protocol

– Key Encryption Key (symmetric) established with RSA

– Traffic Encryption Keys (TEK) exchanged with symmetric

algorithm negotiated at SA establishment (currently only 3-DES

supported)

– Two sets of overlapping keying material maintained

– No explicit key acknowledgements

– Key synchronization maintained by 2-bit key sequence number in

the MAC PDU header

• Traffic Encryption Key Exchange Protocol

– Defined by the TEK FSM transition Matrix


Data Encryption

• DES in CBC mode with IV derived from the frame number

• Hooks defined for other stronger algorithms, e.g. AES

• Two simultaneous keys with overlapping and offset

lifetimes allow for uninterrupted services

– Rules for key usage

• AP: encryption (older key), decryption (both keys)

• AT: encryption (newer key), decryption (both keys)

• Key sequence number carried in MAC header

• Only MAC PDU payload (including sub-headers) is

encrypted

• Management messages are unencrypted


IP addressing

• Unique address that identifies the network and host

• IPv4 consists of 32 bit wide address

– 4 decimal numbers separated by period

– A valid address ranges 0.0.0.0 to 255.255.255.255

– Class A, B, C, classless, restricted address (0 broadcast, any

address starting with 127 is a loopback, a host with binary all 1’s is

broadcast over the specific network. A host with 0 points to itself.

Network address 0 points to its own network)

– Netmask allows to separate network/host part from address

• Performs bit wide AND function

• IPv6, consists of 128 bit wide address


PHY Sub-block Diagram, Example

• With redundant circuit implementation for STC transmitter

• PHY to support three different modes: SC, OFDM and S-

OFDMA

Interleaver

Frequency

Domain

Time

Domain

Antenna 2

Antenna 1

Channel Encoder

+ Rate Matching

Space Time

Encoder

Symbol

Mapper

Subcarrier

Allocation

+ Pilot

Insertion

Subcarrier

Allocation

+ Pilot

Insertion

IFFT

IFFT

DAC

DAC

Analog

Domain

Digital

Domain


PHY

• TDD and FDD

• Adaptive modulation and coding (CC with puncture & RS)

– subscriber by subscriber, burst by burst, uplink and downlink

– Optional Turbo-coding to increase coverage/capacity at the expense of latency

and complexity

• Point to multipoint

• Support for adaptive antennas and space-time coding

• Slot allocation and framing

• Dynamic frequency selection to detect and avoid interference

• 256 sub-carriers (192+28+27+8+1) for OFDM

• Configurable CP length of 1/4, 1/8, 1/16 or 1/32 depending on expected

delay

• Optional signaling support for Adaptive antenna

• Optional transmit diversity support (Space time block codes)


PHY Coding Rates

• If R bps is the input data rate, Nused of FFT, M Modulation order, ½

FEC, then each data subcarrier would carry {(R/Nused)* (2/1)*M} bit

rate

Modulation

Uncoded

Blocks (bytes) RS Code CC Code

Coded Blocks

(bytes)

Overall

Coding

BPSK 12 (12, 12, 0) 1/2 24 1/2

4-QAM 24 (32, 24, 4) 2/3 48 1/2

4-QAM 36 (40, 36, 2) 5/6 48 3/4

16-QAM 48 (64, 48, 8) 2/3 96 1/2

16-QAM 72 (80, 72, 4) 5/6 96 3/4

64-QAM 96 (108, 96, 6) 3/4 144 2/3

64-QAM 108 (120, 108, 6) 5/6 144 3/4


Latency

• Traffic delays through equipment due to processing and propagation

• Increased delay results in annoying voice echo

• Voice over IP applications

• Video conferencing

• Simulcast applications

• Time out issues for some data applications

• Issues to reliably controlling remote devices in real time

• Dynamically adjustment for certain protocols

• Limited alignment performed by BS prior to mobile handover

• Latency decreases as the symbol rate increases

• Latency increases for longer frame size

• Higher latency with interleaver

• Round trip total latency must be ≤100ms? (should be ≤ 20 ms round trip for VoIP

without echo canceller)

• Latency accumulates linearly with increased number of tandem back haul hops


Handoff (HO)

A

B

C

Operator „X‟

backbone network

Operator „Y‟

backbone network

Gateway

Backhaul

connection

Sector sw


Handoff (HO) Schemes

• Mobile WiMAX performs mobile communication but no mesh mode

– Hard handover (HHO) - mandatory

– Micro-Diversity handover (MDHO) - optional

– Fast BS switching (FBSS) - optional

RS

L

RS

L

BS1 BS2

No hysteresis

Rx Threshold

With

Hysteresis

Noise Floor


Hard Handoff (HHO)

• Handover allows MSs to handover between neighboring BSs while moving across the corresponding coverage areas

– This may also be triggered by BS to do an optimal traffic load balancing

• BS periodically broadcasts the neighbor advertisement message (MOB_NBR-ADV). Once the handover decision is made, handover process is carried out in two steps

• Handover preparation: MS or BS may initiate the handover by using the MOB_MSHO-REQ / MOB_BSHO-REQ, the serving BS replies with MOB_BSHO-RSP message containing recommended BSs after negotiation with candidate BSs

• Handover execution: MS sends MOB_HO-IND message to the serving BS and cuts all communication with serving BS. MS then switches the link and executes ranging with target BS. Then MS negotiates basic capabilities, performs authentication and finally registers with the target BS


Micro-diversity Handoff (MDHO)

• Multiple BS serve the MS within the same frame, i.e., Multiple BS transmit the same

packet to the MS within the same frame so that MS can perform the diversity

combining

• MS scans the neighbouring BS and maintain a set of BSs that are involved on

MDHO - the diversity set

• MDHO begins when an MS decides to transmit or receive unicast messages and

traffic from multiple BSs in the same time interval

• MS communicates with all BSs in the diversity set for UL and DL unicast messages

and traffic

– For DL MDHO, two or more BSs provide synchronized transmission of data to MS such

that diversity combining can be performed at the MS

– For UL MDHO, MS transmission is received by multiple BSs such that selection diversity

of the received information could be performed

– When the long-term CINR of a serving BS in diversity set is less than a threshold, the MS

shall send the MOB-MSHO-REQ to delete this BS and update the diversity set

• Allows for true soft-handover (make before break)

• Highly complex, requires synchronization and scheduling above BS layer


Micro-diversity Handoff (MDHO)

Area of

Neighboring

BSs

Diversity

SetActive BS

Neighbor

BS

Neighbor

BS

Active BS

Active BS

Anchor BS

Only RSL measurement

No Traffic

UL & DL Comm

including Traffic

MS


Fast BS Switching (FBSS)

• A state where the MS may rapidly switch from one BS to another

• Multiple BS are ready to serve the MS

• The Diversity Set is maintained as for MDHO

• MS communicates with single BS within given OFDMA frame

• An Anchor BS is defined within the Diversity Set that MS is registered,

synchronized, communicates with for all UL and DL traffic including management

messages

• MS continuously monitors the signal strength of the active BS and select one to be

the anchor BS

• A FBSS handover begins with a decision by a MS to switch to another Anchor BS

using the MOB_MSHO-REQ message

• The anchor BS can be changed from frame by frame. This means every frame can

be sent via different BS in Diversity Set

• Required synchronization among group of BS using a common timing source

• Allows for version of soft-handover (communication is never interrupted)


Fast BS Switching (FBSS)

Data are transmitted & received

but not processed in BS or MS

Area of

Neighboring

BSs

Diversity

SetActive BS

Neighbor

BS

Neighbor

BS

Active BS

Active BS

Anchor BS

Only RSL measurement

No Traffic

UL & DL Comm

including Traffic

MS


Handoff Summary

• BS informs neighbouring BSs via MAC Messages

• Handover initiated by MS & BS

• Process optimized for FBSS (fast BS switching) & MDHO (macro

diversity handoff)

• MS sync with other BSs to estimate associated channel conditions

• Handover process allows a MS to switch to another BS in order to

improve its QoS

• All quality of service and services access are maintained during

handovers

• Hard HOs use a break before make approach and are typically

sufficient for data services.

• Soft HOs, while complex to implement and administer, are

beneficial for applications that require low-latency such as VoIP


Idle Mode / Paging

• Allows the MS to traverse a cellular environment

and become periodically available for DL

broadcast without UL transmission

• For MS: save power and operation resources

• For BS: provide a simple and timely method for

alerting the MS to pending MS-directed DL traffic


Network Entry Process, SS

• SS network entry process

Ranging

Obtain UL

Parameters

Power ON

Scan for

DL Channel

Synchronize

with DL of

Serving BS

Obtain IP

Address

Negotiate

Basic

Capabilities

SS Authorization

and key

Exchanage

Register

with

Network

Network

Entry

Complete

Get Time of

Day

Transfer

Operational

Parameters

Establish

Provisioned

Parameters


Network Initialization, BS

• BS starts by sending beacon

• SS first listens for a beacon and then sends a ranging request in the

ranging period

• BS then sends a ranging response. In the ranging response, the BS

assigns the SS two connection-IDs called the primary CID and the basic

CID. The primary CID is used for further exchange of management

messages while the basic CID is used for further periodic ranging

exchanges

• Registration process is required prior to any connection formation. The

process involves a registration request from the SS, followed by a

registration response from the BS

• After registration, the SS can request for a connection. A connection

request from an SS to the BS elicits a connection response from the BS to

the SS

• The BS and SS are now ready to exchange data with each other


BS Scanning

• BS starts with a known channel. Scan all possible channels,

until a valid channel is found. PHY Sync is the first step.

MAC acquires channel control parameters for DL i.e.; DL

channel descriptor (DCD) containing BS ID, modulation,

coding, interval. Obtain UCD information containing back off,

modulation, coding and message length


BS MAC Layer Sequence of Steps

• Downlink Period

– BS prepares the UL-MAP and allocates the slots to different SSs by keeping in

mind the scheduling policy

– BS sends a beacon which contains the UL-MAP along with the preamble, UCD

and the DCD

– BS sends any pending ranging, registration or connection responses

– BS inspects its four different queues (one each for UGS, ertPS, rtPS and nrtPS)

and sends packets one by one until the DL period finishes

– The packets are sent in order of their priority i.e., UGS followed by ertPS, rtPS

and nrtPS

– The incoming packets from the link layer are added to the queues according to

the flow type

• Uplink Period

– BS receives the packets sent to it by the SS and passes it on to the upper link

layer. The BS does not have any other task to perform in this period


SS MAC Layer Sequence of Steps

• Downlink Period

– SS receives packets sent to it by the BS. Since the packets are

broadcasted it checks for the destination in the packet header

– Incoming packets from the link layer are added to the queues according

to the flow type. These packets are sent in the UL frame whenever the

SS is allocated a slot.

• Uplink Period

– SS checks if any of the ranging, registration and or connection requests

are still pending

– SS reads the UL-MAP and identifies the slots assigned to it

– SS starts sending the packets in the slots assigned to it in the order of

the priority of the packets. This is done by inspecting the four different

queues (one for UGS, ertPS, rtPS and nrtPS)


UL-MAP Preparation

• BS accesses the queues of all the SSs for four different flows

and hence gets to know about the requirements of all the

SSs

– Accessing the queues of the SSs provides information on current

piggybacking and the BW requirements

• BS starts filling the UL-MAP as per the bandwidth

requirements of the SS

• For UGS, ertPS & rtPS flows: the slots are assigned equal to

the number of slots required if the total UL slots are not over.

UGS flows are given the highest priority

• For nrtPS flows: left over slots are divided equally among the

SSs which have bandwidth requirement for nrtPS kind of

traffic


Network Reference Model, Typical

Private

IP

Service

Tunneling

MS R1

Another ASN

Internet or

any IP

Network

R4

R3

R2

R5

R6

ASN

BS

BS

BS

BS

R8

R8

R8

ASN

Gateway

ASN

Gateway

Internet or

any IP

Network

R6

R6

R6

R2

NAP NSP

Visiting

NSP

CSN

Home

NSP

CSN

R

o

a

m

i

n

g


Multi-operator Roaming Framework


WiMAX Reference Point

• Logical reference interfaces between WiMAX network equipment

– R1: MS-ASN

• Implements the air-interface specifications and management plane protocol

– R2: MS-CSN

• Authentication, service authorization, IP host configuration and mobility management

– R3: ASN-CSN

• QoS policy enforcement, mobility management

– R4: ASN-ASN

• Roaming between ASNs

– R5: CSN-CSN

• Roaming between CSNs

– R6: BS-ASNGW

• Mobility tunnel management, intra-ASN path and inter-ASN tunnels

– R7: ASNGW-DP & ASNGW-EP

• An optional protocol for coordinating between two groups identified in R6

– R8: BS-BS

• Control plane protocol between BSs to ensure fast and seamless HO


ASN, Access Service Network

• Gate way equipment between the BS and the Internet

• AAA proxy: transfer of device, user, and service credential to selected NSP

AAA and temporary storage of user profiles

• Provides fast a& efficient radio resource management, QoS policy

enforcement and applications per specific subscriber basis

• Provides Mobility related functions such as handover, location

management, paging within ASN and support for mobile IP with foreign-

agent functionality

• May include redundancy and load-balancing among several ASN-GWs

• Relay functionality for establishing IP connectivity between the MS and the

CSN

• Admission control functions

• Cache SS profiles and encryption keys

• Establishes mobility tunnels with BS and other resources


CSN, Connectivity Service Network

• IP address allocation to the MS for user sessions

• AAA proxy or server for user, device and services authentication, authorization and accounting

• Policy and QoS management based on the SLA/contract with the user

• Subscriber billing and inter-operator settlement

• Inter-CSN tunneling to support roaming between NSPs

• Connectivity infrastructure and policy control for such services as Internet access, access to other IP networks, ASPs, location-based services, peer-to-peer, VPN, IP multimedia services, law enforcement, and messaging

• Inter-ASN mobility management and mobile IP home agent functionality


Authentication

• Support for device, user, and mutual authentication between MS/SS and

the NSP, based on PKMv2

• Support for authentication mechanisms, using variety of credentials,

including shared secrets, subscriber identity module (SIM) cards, universal

SIM, universal integrated circuit card, removable user identity module, and

X.509 certificate as long as they are suitable for EAP methods satisfying

RFC 4017

• Support for global roaming between home and visited NSPs in mobile

scenario, including support for credential reuse and consistent use of

authorization and accounting through the use of RADIUS in the ASN and

the CSN

• Accommodation of mobile IPv4 and IPv6 security associations

management

• Support for policy provisioning at the ASN or the CSN by allowing for

transfer of policy related information from the AAA to the ASN or CSN


Validation & Interoperability

• IEEE P802.16c

– Published in Jan’03

– Specifies particular combinations of options

– Used as basis of compliance testing

• MAC Profile: ATM and Packet

• PHY Profile: 1.25-20, 25 & 28 MHz, TDD & FDD

– Test Protocols: IEEE Standards 802.16/Conformance-0X

• PICS

• Test Suite Structure & Test Purposes

• Radio Conformance Tests

– Two levels of mobile certifications (Wave 1 & wave 2)

• Wave 1 includes basic PHY and MAC functions

• Wave 2 includes MIMO operation


Interoperability Conformance

• PHY Tests

– Emission, spectral mask, power control and accuracy

– Interference tolerance at CCI & ACI

– Relative constellation errors (RCE) vs. symbol & sub-carriers

– Spectral flatness, crest factor, peak, average & min EVM

– Error rates, RSSI, SNR, CINR, PCINR, SINR and Rx threshold

– Frequency error, DynFF


IOT, Release 1 Mobile PHY Profile and

CertificationRelease 1 PHY Profile Function Wave 1 Wave 2 Comments

PUSC ✓ ✓

PUSC w/ All Subchannels ✓ ✓

FUSC ✓ ✓

AMC 2x3 ✓ Required in Wave 1 for Band Class 3

PUSC ✓ ✓

AMC 2x3 ✓ Required in Wave 1 for Band Class 3

Initial Ranging ✓ ✓

Handoff Ranging ✓ ✓

Periodic Ranging ✓ ✓

Bandwidth Request ✓ ✓

Fast-Feedback 6-bits ✓ ✓

Repetition ✓ ✓

Randomization ✓ ✓

Convolutional Coding (CC) ✓ ✓

Convolutional Turbo Coding (CTC) ✓ ✓

Interleaving ✓ ✓

Preamble ID ✓ ✓

DCD, UCD ✓ ✓

Packing ✓ ✓

Fragmentation ✓

PHS ✓

IPv4 ✓

IPv6/IPv4 with ROHC ✓

BS Initiated ✓ ✓

SS Initiated ✓ ✓

H-ARQ Chase Combining ✓ ✓ Required in Wave 1 for Band Class 3

BS-BS Time/Freq Synchronization N/A N/A

BS-BS Frequency Synchronization N/A N/A

MSS Synchronization ✓ ✓

Closed-loop Power Control ✓ ✓

Open-loop Power Control ✓ ✓

Power, Frequency error ✓ ✓

Trasmit constellation error, Spectrum ✓ ✓

Synchronization

Power Control

Transmitter Measurements

BS Configuration

MAC PDU Manipulation

Service Flow Initiation

DL Subcarriers Allocation

UL Subcarriers Allocation

Ranging & Bandwidth Request

Channel Coding


IOT, Release 1 Mobile PHY Profile and

Certification (conti.)Release 1 PHY Profile Function Wave 1 Wave 2 Comments

Physical CINR Using Preamble ✓ ✓

Physical CINR Using Pilots ✓ ✓

Effective CINR Using Pilots ✓ Required in Wave 1 for Band Class 3

RSSI Measurements ✓ ✓

Ping Support ✓ ✓

Ack/Nack Support ✓ ✓

AWGN ✓ ✓

RF Amplitude ✓ ✓

DL 4-QAM ✓ ✓

DL 16-QAM ✓ ✓

DL 64-QAM ✓ ✓

UL 4-QAM ✓ ✓

UL 16-QAM ✓ ✓

UL 64-QAM (Optional) ✓ ✓

Normal MAP ✓ ✓

Compressed MAP ✓ ✓

Sub DL-UL-MAP ✓ ✓

UGS ✓ ✓

erPS ✓ ✓

rtPS ✓ ✓

nrtPS ✓ ✓

Best Effort ✓ ✓

2nd Order Matrix A/B ✓ Required in Wave 1 for Band Class 3

Collaborative Spacial Multiplexing ✓ Required in Wave 1 for Band Class 3

Fast Feedback on DL ✓ Required in Wave 1 for Band Class 3

Mode Selection Feedback w/ 6-bits ✓ Required in Wave 1 for Band Class 3

MIMO DL-UL Chase ✓ Required in Wave 1 for Band Class 3

PUSC w/ Dedicated Pilots ✓ Required in Wave 1 for Band Class 3

AMC 2x3 w/ Dedicated Pilots ✓ Required in Wave 1 for Band Class 3

UL Sounding 1 (Type A) ✓ Required in Wave 1 for Band Class 3

UL Sounding 2 ✓ Required in Wave 1 for Band Class 3

CINR Measurement (group Indication) ✓ PUSC, Required in Wave 1 for Band Class 3

MIMO Computation Feedback Cycle ✓ Required in Wave 1 for Band Class 3

MAP Support

MIMO (IO-MIMO for BS)

AAS/BS (IO-BF for BS)

Receiver Measurement

Modulation

Data Delivery Methods

Impairments


Interoperability MAC Conformance

• MAC Test:

–802.3 Frame format

–Protocol

–Scheduling

–Admission control

–QoS

–MIMO

–Link adaptation


IOT, Others

• SS Access Point and MSS Access Point:

– SS/MS connectivity, provisioning and admission control

– Over the air and end to end security

– Mobility management

– Device management

– UL and DL data exchange

– Authorization and tunneling for specialized IP services

– Application layer end to end signaling

– Power management, compression and data reliability

• CN1: Control, data and management plane between the RANs and operator‟s core network

• CN2: control , management and service planes to ASP networks

• RNSN: Control, data and management plane interfaces between two RNSNs

• RNSNAP: Control, data and management plane interfaces between an AP and an RNSN

• Mobility Management: Provisioning, multi-sector handover and end to end mobility management


System Synchronization

• For TDD system, the Tx and Rx time frames among BSs/SSs must be

synchronized to avoid interference and the SS transmission do not overlap each

others as they arrive to BS

• Timing and frequency offset can influence the performance

– Mitigated by reserved pilots & increased CP duration

• Frequency offset can influence orthogonality of sub-carriers

• Loss of orthogonality can lead to inter-carrier interference

• Loss of synchronization causes hits during handover

• Tracking and estimating the position of the frame is necessary for reliable data

delivery

• Timing sync through GPS (cost effective solution but difficult to access open sky if

in the basement)

• For interference mitigation, system-wide synchronization is essential when using

TDDtFRAME1

BS-1 DL-TX UL-RX DL-TX UL-RX DL-TX UL-RX

time

tFRAME2

BS-2 DL-TX UL-RX DL-TX UL-RX DL-TX

time


Network Synchronization

• For TDD system, the Tx and Rx time frames among BSs/SSs must

be synchronized to avoid interference

• Long FEC coding, interleaving and frame structure lead to jitter and

wander accumulation

• Clock accuracy, timing and synchronization is essential for reliable

MS handover, MS/SS operation and to minimize interference affect

in MIMO configuration

• GPS timing to aid in synchronizing the network

• IEEE 1588 timing over IP/Ethernet backhaul

– Synchronization distributed from IEEE1588 master clock in the network

– Less accurate than the GPS

• WiMAX network is entirely IP and there is no option of recovering

timing signal as there is with TDM application


Q & A

• Thank you for your attention!

• Your feedback and comments are greatly appreciated

Introduction to WiMAX Presentation

Documents

bs scheduler

access open

psh fcd0

acquired burst

fast bs switching

ul channel

actual phy

dl channel