Spread spectrum systems II: WCDMA • WCDMA basic properties • Channel mapping • Chip sequence processing • Soft handover • Power control
Jan 13, 2016
Spread spectrum systems II: WCDMA
• WCDMA basic properties
• Channel mapping
• Chip sequence processing
• Soft handover
• Power control
1. WCDMA basic properties
Issues / important concepts:
• Two duplex alternatives: UTRA FDD vs. UTRA TDD
• Spectrum allocation (UTRA FDD)
• Spreading in WCDMA
• DPDCH/DPCCH/DPCH channel bit rates
Two duplex alternatives: FDD vs. TDD
In UTRA FDD (Frequency Division Duplex), uplink and downlink are separated in frequency domain:
UL UL DLDL
frequency
In UTRA TDD (Time Division Duplex), uplink and downlink are separated in time domain:
time
UL UL DLDL UL UL DLDL ......
Two duplex alternatives: FDD vs. TDD
UTRA FDD will be more widely used in the near future, since UTRA TDD technology is more complex.
However, UTRA TDD offers some benefits:
More flexible UL/DL capacity allocation (in non-voice applications, DL usually demands more capacity than UL) Channel reciprocity (channel estimation in one direction could be used in the other direction)
No need for duplex filter.
1.
2.
3.
Spectrum allocation for UTRA FDD
1920 - 1980 MHz1920 - 1980 MHz 2110 - 2170 MHz2110 - 2170 MHz
Uplink Downlink
60 MHz
5 MHz
Chip sequnces are multiplexed in code domain and transmitted within a 5 MHz frequency slot.
The chip rate is always 3.84 Mchips/s.
Spectrum is allocated to operators at
this level
(Europe & part of Asia)
Spreading in WCDMA
Channel data
Channel data
Channelization code
Channelization code
Scrambling code
Scrambling code
Channel bit rate
Chip rate Chip rate
Usage of codeUsage of code UplinkUplink DownlinkDownlink
Channelization code
Scrambling code
User separation
User separation Cell separation
(always 3.84 Mchips/s)
QPSK
Spreading in WCDMA
Chip rate = SF x channel bit rateChip rate = SF x channel bit rate
Uplink: DPCCH SF = 256, DPDCH SF = 4 - 256Uplink: DPCCH SF = 256, DPDCH SF = 4 - 256
Downlink: DPCH SF = 4 - 256 (512)Downlink: DPCH SF = 4 - 256 (512)
SF = Spreading factor
One bit consists of 256 chips
One bit consists of 4 chips
Uplink DPDCH bit rates
256
128
64
32
16
8
4
SF Channel bit rate (kb/s)
User data rate (kb/s)
15
30
60
120
240
480
approx. 7.5
approx. 15
approx. 30
approx. 60
approx. 120
approx. 240
960 approx. 480
Uplink DPCCH bit rate
256
SF Channel bit rate
15 kb/s
How many control channel bits does one
time slot contain?
Each 10 ms radio frame (38400 chips long) is divided into 15 time slots (2560 chips long).
Since SF = 256, each time slot contains 10 control channel bits that can be used, for example, like this:
FBI TPCTFCIPilot
3GPP TS 25.211 Slot format 3
Downlink DPCH bit rate
256
128
64
32
16
8
4
SF Channel bit rate (kb/s)
User data rate (kb/s)
15
30
60
120
240
480
approx. 1-3
approx. 6-12
approx. 20-24
approx. 45
approx. 105
approx. 215
960 approx. 456
512
1920 approx. 936
User data rate vs. channel bit rate
Channel bit rate (kb/s)Channel bit rate (kb/s)
User data rate (kb/s)User data rate (kb/s)
Channel codingChannel coding
InterleavingInterleaving
Bit rate matchingBit rate matching
Interesting for user
Interesting for user
Important for system
Important for system
2. Channel mapping
Issues / important concepts:
• Physical channels
• Transport channels
• Logical channels
• DPDCH/DPCCH multiplexing in uplink
• DPCH user/control data multiplexing in downlink
Phy Phy Lower layers
Lower layers
Logical / transport / physical channels
RLCRLC RLCRLC
MACMAC
UE Base station RNC
MACMAC
Lower layers
Lower layers Phy Phy
Logical channels
Physical channels
Transport channels
: :
WCDMA
Logical / transport channel mapping
CCCHCCCH DCCHDCCH
PCHPCH DCHDCHDSCHDSCHFACHFACHBCHBCHDCHDCHCPCHCPCHRACHRACH
CTCHCTCHBCCHBCCHPCCHPCCH
Uplink Downlink
DTCHDTCH
Logical channels
Transport channels
Note the different possibilities for transmitting user data over transport channels
DCCHDCCH
DTCHDTCH
CCCHCCCH
Transport / physical channel mapping
PCHPCH DCHDCHDSCHDSCHFACHFACH BCHBCHCPCHCPCHRACHRACH
PRACHPRACH PCPCHPCPCH SCCPCHSCCPCH PCCPCHPCCPCHDPDCHDPDCH
DPCCHDPCCH
SCHSCHCPICHCPICH
AICHAICH
PICHPICH
CSICHCSICHPhysical channels
Transport channels
DPCHDPCH
CD/CA-ICH
CD/CA-ICH
Uplink Downlink
PDSCHPDSCH
DCHDCH
These channels are only for transport of information in the physical layer at the air (radio) interface
DPDCH / DPCCH structure in uplink
DataData
PilotPilot TFCITFCI FBIFBI TPCTPC
DPDCH (I-branch)DPDCH (I-branch)
10 ms radio frame (38400 chips)
0 1 2 14
Time slot containing 2560 chips
DPCCH (Q-branch)DPCCH (Q-branch)
Dual-channel QPSK modulation:
DPCH structure in downlink
QPSK modulation, time multiplexed data and control information:
TFCITFCI DataData TPCTPC DataData
0 1 2 14
PilotPilot
Time slot containing 2560 chips
10 ms radio frame (38400 chips)
3. Chip sequence processing
Issues / important concepts:
• Spreading, scrambling, multiplexing and modulation
• Uplink and downlink processing somewhat different
• Channelization codes vs. spreading codes
Uplink spreading
DPDCH
DPCCH
OVSF Code 1
The DPCCH is spread on the Q-branch using SF = 256.
In case of very high user bit rates, up to six DPDCH channels can be used in parallel by distributing the signals to the I and Q branches using additional OVSF codes.
OVSF Code 2
In the UE, the user data (DPDCH) and control data (DPCCH) signals are spread to the chip frequency of 3.84 Mchips/s using different channelisation codes, also called OVSF (Orthogonal Variable Spreading Factor) codes.
I branch
Q branch
Uplink multiplexing
++
DPDCH
DPCCH
I branch
Q branch
Weight 1
jWeight 2
The DPCCH signal and DPDCH signal (or up to 6 DPDCH signals) are synchronously combined, i.e. ”multiplexed in code domain”, to form the complex signal I+jQ.
Complex-valued signal I + jQ
Uplink scrambling
++
I + jQ
After scrambling, signals from different UEs can be separated at the base station, since each UE uses a different scrambling code.
Scrambling codes must have good correlation properties even when not synchronized (=> m-sequence or Gold codes).
Scrambling codeRe{S}
Im{S}
Complex-valued signal S
The complex signal I+jQ is multiplied by the complex-valued, UE specific scrambling code.
Uplink modulation
++
Re{S}
Im{S}
To RF part and UE antenna
The real and imaginary parts of the scrambled signal S are fed to the I and Q branches of the modulator and are modulated by sinusoids with a 90-degree phase shift to achieve the desired QPSK modulation.
The QPSK signal is transmitted from the UE antenna.
Pulse shaping
Pulse shaping
Pulse shaping
Pulse shaping
cos(t)
-sin(t)
At the receiver side
At the transmitter side, signal formats and processing details are standardised (see 3GPP TS 25.213).
At the receiver side, base station manufacturers are free to implement any receiver structure they wish.
In general terms, the code processing is in the reverse order (demodulation, despreading, demultiplexing ...) and makes use of a Rake receiver able to resolve and despread separate multipath replicas of the transmitted signal.
Channel estimation and phase synchronisation is based on pilot bits transmitted in the DPCCH signal.
Downlink spreading
Serial/parallel conversion is applied to two consecutive channel bits. The bits in the I and Q branches are then spread using the same OVSF (channelisation) code.
++S/PAny downlink physical channelexcept SCH
OVSF Code n
OVSF Code n j
Complex-valued signal I + jQ
Downlink scrambling
Scrambling code
++
The spreaded, complex-valued signal is chipwise multiplied with a complex-valued scrambling code.
Scrambling codes are selected from a base station specific code set. A scrambling code can be shared among several physical channels.
Adjacent base stations use different (sets of) scrambling codes.
Downlink multiplexing
Weight n
Other downlink physical channels
SCH signal(s)
Before the spreaded and scrambled physical channels are ”multiplexed in code domain”, signal powers are adjusted to the appropriate levels determined by the downlink closed loop power control (on a channel-by-channel basis).
Note that all channels are multiplexed synchronously.
Re{S}
Im{S}
Complex-valued signal S
Downlink multiplexing
Weight n
Other downlink physical channels
SCH signal(s)
Synchronisation channels (SCH) are spread using special code sequences (i.e. no OVSF codes are involved).
SCH is first multiplexed with CCPCH in time domain. The composite signal is then added to the other channels in code domain (see 3GPP TS 25.211).
Re{S}
Im{S}
Downlink modulation
++
To RF part and
base station antenna
Pulse shaping
Pulse shaping
Pulse shaping
Pulse shaping
cos(t)
-sin(t)
Re{S}
Im{S}
The real and imaginary parts of the multiplex signal S are fed to the QPSK modulator like in uplink.
Note that this signal contains information for many UEs.
Synchronous / non-s. chip sequences
Sequences start here
Sequences end here
One sequence starts here
Another sequence starts here
Two synchronous chip sequences
Two non-synchronous chip sequences
Chips
Chip Sequence = encoded bit/symbol
Synchronous / non-s. chip sequences
Synchronous chip sequences
Non-synchronous chip
sequences
Channelization (Hadamard-Walsh)
codes
Scrambling (m-sequence, Gold)
codes
No interference (sequences are all
orthogonal)
Little interference (sequences are
near orthogonal)
Little interference (sequences are
near orthogonal)
Large interference
Different effect on different
types of codes:
4. Soft handover
Issues / important concepts:
• Serving RNC, Drift RNC, SRNS Relocation
• Micro/macrodiversity combining
• Soft handover in uplink
• Soft handover in downlink
Serving RNC and Drift RNC in UTRAN
Core network
Iu
Iur
Iub
Iub
DRNC
SRNC
UEUE
BSBS
BSBS
RNCRNC
RNCRNC
Concept needed for:Soft handover between base stations belonging to different RNCs
SRNS Relocation
Core network
Iu
Iur
Iub
Iub
DRNC
SRNC
UEUE
BSBS
BSBS
RNCRNC
RNCRNC Iu
SRNC
SRNC provides: 1) connection to core network 2) macrodiversity combining point
Micro- / macrodiversity combining
Iu
Iur
Iub
Iub
DRNC
SRNC
UEUE
BSBS
BSBS
RNCRNC
RNCRNC Macrodiversity combining point in
SRNC
Core network
Rake receiver
Multipath propagation
Microdiversity combining point in base station
(uplink)
Diversity combining, soft handover
Microdiversity combining: multipath signal components are processed in Rake branches and combined (MRC = Maximum Ratio Combining)
Macrodiversity combining: bit sequences carrying the same signal (but with different bit error positions) are either combined at SRNC (bit-by-bit majority voting), or best quality signal is selected.
Hard handover: slow, complex signalling
Soft handover: fast selection in SRNC is possible due to macrodiversity combining
(uplink)
Microdiversity combining, soft handover
Soft handover: same signal is transmitted via several base stations
Advantage: number of multipath components is increased
Draw-back: in downlink, soft handover decreases capacity
(downlink)
BSBS
BSBS
Rake receiver
UEUE
Different code
sequences
5. Power control
Issues / important concepts:
• Near-far problem
• Uplink SIR expression; what means Target SIR?
• Open loop power control
• Inner loop (closed loop) power control
• Outer loop (closed loop) power control
Why is power control needed?
Strong signal dominates
Near-far problem arises in uplink when all UEs use the same transmit power:
BSBS
UEUE
UEUE
Weak signal will be drowned
Rather, UEs should adjust their transmit power levels so that the received power levels are approximately the same at the base station.
Uplink SIR expression
Ps . SF
(N-1).Ps + Pn
SIR
In uplink, in case of same received power levels (and ignoring interference from UEs located in other cells) the signal-to-interference ratio for the k:th user is
This simple rule-of-thumb expression
• is useful for estimating capacity in uplink
• is the basis of admission control in uplink
• explains “Target SIR” used in power control.
Analysis of uplink SIR expression
Ps . SF
(N-1).Ps + Pn
SIR
Signal-to-interference ratio (SIR) is a very important parameter in a DS-CDMA system.
SIR describes the situation after despreading in the CDMA receiver.
The corresponding ratio before despreading is called CIR (carrier-to-interference ratio).
Analysis of uplink SIR expression
Ps . SF
(N-1).Ps + Pn
SIR
SF = Spreading Factor.
We assume here that the power of the desired signal (of k:th user) after despreading is SF times the power of interfering signal (of another user) after despreading if the powers before despreading are the same.
In other words, this is a crude model for estimating the level of cross-correlation in the CDMA receiver.
Analysis of uplink SIR expression
Ps . SF
(N-1).Ps + Pn
SIR
It is assumed that the received signal power (before despreading) of all N active users in the cell is Ps.
Pn is the thermal noise power in the receiver.
(In case there are users with different bit rates - and thus different spreading factors - this expression must obviously be modified)
Analysis of uplink SIR expression
Ps . SF
(N-1).Ps + Pn
Target SIR
The SIR for the k:th user must be larger than a certain value, Target SIR. In other words, the total interference in the system must remain below a certain target level.
>
First, we see that the best case is when Ps >> Pn . In other words, CDMA systems are interference limited, not noise limited.
Second, the inequality above is valid for values of N up to a maximum value Nmax , the capacity of the cell.
Analysis of uplink SIR expression
The target SIR value depends on various issues, such as required Bit Error Ratio (BER) or Frame Error Ratio (FER), user/channel bit rate of k:th user, etc.
BER
SIR
High BER means low target SIR
Low BER means high target SIR
Open loop power control
UE estimates the average path loss in downlink ...
This simple and inaccurate power control scheme must be used during the random access process at the beginning of a connection until more accurate control information is available.
... and adjusts the uplink transmission power accordingly
BSBSUEUE
(Note: uplink / downlink fading in UTRA FDD is not the same)
Inner loop power control
UE transmits initial signal
Is measured SIR larger (smaller) than
Target SIR?
BSBSUEUE
Inner loop power control (also called fast power control) is used both in uplink (shown in this figure) and downlink.
If answer is yes: decrease (increase) power
This loop is performed 1500 times per second
UE decreases (increases) transmit power
Outer loop power control
Is signal quality (BER)
ok?
BSBSUEUE
Outer loop power control is used both in uplink (shown in this figure) and downlink.
Inner loop power control uses new Target SIR value
RNCRNC
If not, increase or decrease Target SIR