10.1.1.63.7134.pdf

E-UTRA RACH within the LTE system

ROMAIN MASSON

Masters Degree ProjectStockholm, Sweden 2006-02-03

XR-EE-KT 2006:002

This page intentionally contains only this sentence.

Abstract

The thesis project has been carried out at the standardization department ofLG Electronic Mobile Comm. LG is a member of 3GPP (3rd Generation Part-nership Project), the organization which standardizes UMTS (Universal MobileTelecommunication Services). The third generation (3G) mobile telecommuni-cation systems are deployed and the need of 3G long term evolution (LTE) ispointed out to meet the future demand. This evolution is based on a new airinterface. Several new technology components such as OFDM are presented aspotential candidates. Members of 3GPP are currently considering the evolutionof the radio access technology in order to ensure their competitiveness. One fieldof research concerns the Physical Random Access Channel (RACH), used dur-ing the initial access. This project proposes to analyze the new requirements, tohighlight the new air interface candidates, to compare different channel designsand to provide a first examination of the subject.

3

Acknowledgements

First of all, I would like to thank Patrick Fischer and Dragan Vujcic for theiradvice, their time and for giving me the chance to perform my Master Thesisas part of the standardization team at LG Electronic Mobile Comm. I wouldalso like to show gratitude to the whole S&AT (Standardization and AdvancedTechnology) group for their warm welcome and hospitality. I was very pleasedto work in such a pleasant working environment. Last but not least, I wouldlike to thank my examiner at KTH, Erik Larsson, and my advisor Xi Zhang.

5

Contents

1 Introduction 11.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1.1 3GPP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1.2 LTE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1.3 RACH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.2 Problem definition . . . . . . . . . . . . . . . . . . . . . . . . . . 21.2.1 A new air interface . . . . . . . . . . . . . . . . . . . . . . 21.2.2 Several possible scenarios . . . . . . . . . . . . . . . . . . 31.2.3 Requirements and RACH purpose . . . . . . . . . . . . . 31.2.4 Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.3 Research approach . . . . . . . . . . . . . . . . . . . . . . . . . . 41.3.1 Multiple Access protocols . . . . . . . . . . . . . . . . . . 41.3.2 Air interface analysis . . . . . . . . . . . . . . . . . . . . . 41.3.3 Existing random access channels analysis . . . . . . . . . 41.3.4 List different scenarios . . . . . . . . . . . . . . . . . . . . 41.3.5 Implementation . . . . . . . . . . . . . . . . . . . . . . . . 41.3.6 Evaluation and Analysis . . . . . . . . . . . . . . . . . . . 5

2 Multiple Access Protocols 72.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.2 The pure Aloha protocol . . . . . . . . . . . . . . . . . . . . . . . 72.3 The Slotted Aloha protocol . . . . . . . . . . . . . . . . . . . . . 82.4 The CSMA protocol . . . . . . . . . . . . . . . . . . . . . . . . . 82.5 The Reservation protocols . . . . . . . . . . . . . . . . . . . . . . 102.6 Temporary conclusion . . . . . . . . . . . . . . . . . . . . . . . . 11

3 Air interface candidates for uplink access 133.1 Basic OFDM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

3.1.1 Properties and main advantages . . . . . . . . . . . . . . 133.1.2 The key elements: IFFT/FFT . . . . . . . . . . . . . . . 143.1.3 The need of a cyclic prefix . . . . . . . . . . . . . . . . . . 153.1.4 OFDMA: OFDM for multiple access . . . . . . . . . . . . 16

3.2 DFTs OFDMA . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173.3 IFDMA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183.4 General comparison . . . . . . . . . . . . . . . . . . . . . . . . . 19

7

8 Contents

4 Existing random access procedures 214.1 RACH procedure within UMTS . . . . . . . . . . . . . . . . . . . 21

4.1.1 RACH purpose . . . . . . . . . . . . . . . . . . . . . . . . 214.1.2 Collision handling . . . . . . . . . . . . . . . . . . . . . . 214.1.3 The power ramping technique . . . . . . . . . . . . . . . . 224.1.4 Format of the preambles . . . . . . . . . . . . . . . . . . . 234.1.5 Format of the message . . . . . . . . . . . . . . . . . . . . 244.1.6 RACH access procedure . . . . . . . . . . . . . . . . . . . 24

4.2 Ranging procedure within WiMax . . . . . . . . . . . . . . . . . 274.2.1 Wimax overview . . . . . . . . . . . . . . . . . . . . . . . 274.2.2 Ranging overview . . . . . . . . . . . . . . . . . . . . . . . 284.2.3 Ranging procedure within the WirelessMAN OFDMA PHY 284.2.4 Signatures . . . . . . . . . . . . . . . . . . . . . . . . . . . 314.2.5 Ranging signal . . . . . . . . . . . . . . . . . . . . . . . . 32

5 RACH design within LTE 335.1 Discussion on the RACH purpose . . . . . . . . . . . . . . . . . . 33

5.1.1 Timing adjustments . . . . . . . . . . . . . . . . . . . . . 335.1.2 Power adjustments . . . . . . . . . . . . . . . . . . . . . . 345.1.3 Resource request . . . . . . . . . . . . . . . . . . . . . . . 34

5.2 Discussion on the transmission method . . . . . . . . . . . . . . . 355.2.1 Signature and payload . . . . . . . . . . . . . . . . . . . . 355.2.2 Transmission band assigned to random and reservation

access channel . . . . . . . . . . . . . . . . . . . . . . . . 365.2.3 Transmission bandwidth allocated to a UE random access 375.2.4 Transmission duration and random access period . . . . . 37

5.3 Signal parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . 375.4 Power ramping . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395.5 Frequency hopping . . . . . . . . . . . . . . . . . . . . . . . . . . 405.6 Signature format . . . . . . . . . . . . . . . . . . . . . . . . . . . 415.7 Subcarriers mapping . . . . . . . . . . . . . . . . . . . . . . . . . 415.8 Random access procedure . . . . . . . . . . . . . . . . . . . . . . 42

5.8.1 UE procedure . . . . . . . . . . . . . . . . . . . . . . . . . 425.8.2 NodeB procedure . . . . . . . . . . . . . . . . . . . . . . . 43

6 Implementation and Simulation 456.1 System overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 456.2 Channel modeling . . . . . . . . . . . . . . . . . . . . . . . . . . 46

6.2.1 Frequency selective and Time varying channel . . . . . . . 466.2.2 Additive white noise . . . . . . . . . . . . . . . . . . . . . 486.2.3 Shadowing variance . . . . . . . . . . . . . . . . . . . . . 49

6.3 Detection analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 506.3.1 Time domain detection . . . . . . . . . . . . . . . . . . . 506.3.2 Frequency domain detection . . . . . . . . . . . . . . . . . 51

6.4 Simulation results . . . . . . . . . . . . . . . . . . . . . . . . . . 536.4.1 Threshold tuning . . . . . . . . . . . . . . . . . . . . . . . 536.4.2 Impact of the number of symbols . . . . . . . . . . . . . . 546.4.3 Allocated bandwidth to random access . . . . . . . . . . . 556.4.4 Subcarriers mapping analysis . . . . . . . . . . . . . . . . 556.4.5 Impact of the delay between two attempts . . . . . . . . . 56

Contents 9

6.4.6 Improvement with frequency hopping . . . . . . . . . . . 576.4.7 Power ramping . . . . . . . . . . . . . . . . . . . . . . . . 586.4.8 Impact of the shadowing variance . . . . . . . . . . . . . . 58

7 Conclusion and Future work 61

References 63

List of Tables

4.1 Hadamard codes of length 16 . . . . . . . . . . . . . . . . . . . . 24

5.1 System parameters for uplink transmission . . . . . . . . . . . . . 385.2 Particular scenario for random access . . . . . . . . . . . . . . . . 39

6.1 Typical Urban channel profile . . . . . . . . . . . . . . . . . . . . 46

11

List of Figures

2.1 The pure Aloha protocol . . . . . . . . . . . . . . . . . . . . . . . 72.2 The vulnerable period of the basic Aloha protocol . . . . . . . . 82.3 The vulnerable period of the slotted Aloha protocol . . . . . . . 82.4 The hidden terminal problem . . . . . . . . . . . . . . . . . . . . 92.5 The basic detection avoidance procedure . . . . . . . . . . . . . . 102.6 Illustration of the CSMA-CA protocol . . . . . . . . . . . . . . . 102.7 PRMA protocol in the case of 6 slots per frame . . . . . . . . . . 112.8 Reservation Aloha: example of transmission . . . . . . . . . . . . 112.9 Temporary solution: a subframe is dedicated to random access . 12

3.1 Principle of the OFDM method . . . . . . . . . . . . . . . . . . . 143.2 Basis functions in OFDM system . . . . . . . . . . . . . . . . . . 143.3 Basic OFDM transmission . . . . . . . . . . . . . . . . . . . . . . 153.4 Typical channel impulse response . . . . . . . . . . . . . . . . . . 153.5 Adding a cyclic prefix to a frame . . . . . . . . . . . . . . . . . . 163.6 Frequency multiplexing using OFDM . . . . . . . . . . . . . . . . 173.7 DFTs OFDMA transmitter . . . . . . . . . . . . . . . . . . . . . 173.8 Sub carriers allocation schemes . . . . . . . . . . . . . . . . . . . 183.9 IFDMA signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

4.1 Overview of the RACH procedure within UMTS . . . . . . . . . 224.2 Power ramping technique . . . . . . . . . . . . . . . . . . . . . . 234.3 Orthogonal Variable Spreading Factor (OVSF) codes . . . . . . . 254.4 MAC procedure for a RACH access with UMTS . . . . . . . . . 264.5 Physical procedure for a RACH access with UMTS . . . . . . . . 274.6 Network entry procedure . . . . . . . . . . . . . . . . . . . . . . . 284.7 WiMAX TDD mode . . . . . . . . . . . . . . . . . . . . . . . . . 294.8 Ranging opportunities . . . . . . . . . . . . . . . . . . . . . . . . 294.9 OFDMA PHY ranging procedure . . . . . . . . . . . . . . . . . . 304.10 The PRBS generator . . . . . . . . . . . . . . . . . . . . . . . . . 314.11 Initial ranging signal . . . . . . . . . . . . . . . . . . . . . . . . . 324.12 Initial ranging block scheme . . . . . . . . . . . . . . . . . . . . . 32

5.1 Timing offset caused by the round trip delay . . . . . . . . . . . 335.2 Resource request after being synchronized . . . . . . . . . . . . . 345.3 The resource request could be part of the RACH signal . . . . . 355.4 One to one correspondance between the sets of signature and the

available resources . . . . . . . . . . . . . . . . . . . . . . . . . . 355.5 Transmission band assigned to random acces . . . . . . . . . . . 36

13

14 List of Figures

5.6 The necessary guard time . . . . . . . . . . . . . . . . . . . . . . 375.7 Illustration of a basic transmission in system bandwidth of 10 MHz 385.8 Frequency hopping principle . . . . . . . . . . . . . . . . . . . . . 405.9 Impact of the mapping on the autocorrelation function . . . . . . 42

6.1 System overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 456.2 Multipath radio environment . . . . . . . . . . . . . . . . . . . . 466.3 The TU channel impulse response . . . . . . . . . . . . . . . . . 476.4 The scattering nature of the paths . . . . . . . . . . . . . . . . . 476.5 The Jakes coefficients used in a multipath scheme . . . . . . . . . 496.6 Additive white Gaussian noise . . . . . . . . . . . . . . . . . . . . 496.7 A signature is defined as a code and a frequency band . . . . . . 506.8 Correlators at the NodeB . . . . . . . . . . . . . . . . . . . . . . 516.9 The NodeB procedure when detecting in the frequency domain . 516.10 Averaging procedure . . . . . . . . . . . . . . . . . . . . . . . . . 526.11 Several sets of FFT/IFFT are performed to detect the presence

of a signature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526.12 Threshold tuning: missing probabilities . . . . . . . . . . . . . . 536.13 Threshold tuning: false alarm rate . . . . . . . . . . . . . . . . . 546.14 The impact of the number of OFDM symbols on the detection rate 546.15 The impact of the transmission bandwidth . . . . . . . . . . . . . 556.16 Timing estimation error variance for each mapping scheme . . . . 566.17 Impact of the delay between two attempts . . . . . . . . . . . . . 566.18 Gain of performance with frequency hopping at 3km/h, BW =

500 kHz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 576.19 Gain of performance with frequency hopping at 3km/h, BW =

1.25 MHz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 576.20 Gain of performance with frequency hopping at 3km/h, BW = 5

MHz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 586.21 Gain of performance with power ramping . . . . . . . . . . . . . 596.22 Impact of the difference between uplink and downlink shadow

fading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

List of Abbreviations

3GPP . . . . . . . . . . . . . . . . . . . . . . Third Generation Partnership Project

AACK . . . . . . . . . . . . . . . . . . . . . . . Acknowledgment packetADSL . . . . . . . . . . . . . . . . . . . . . . Asymmetric Digital Subscriber LineAI . . . . . . . . . . . . . . . . . . . . . . . . . Acquisition IndicatorAICH . . . . . . . . . . . . . . . . . . . . . . Acquisition Indication ChannelARIB/TTC . . . . . . . . . . . . . . . . Association of Radio Industries and Business /

Telecommunication Technology CommitteeATIS . . . . . . . . . . . . . . . . . . . . . . . Alliance for Telecommunications Industry Solutions

BBCH . . . . . . . . . . . . . . . . . . . . . . . Broadcast ChannelBPSK . . . . . . . . . . . . . . . . . . . . . . Binary Phase Shift KeyingBS . . . . . . . . . . . . . . . . . . . . . . . . . Base Station

CCA . . . . . . . . . . . . . . . . . . . . . . . . Collision AvoidanceCAZAC . . . . . . . . . . . . . . . . . . . . Constant Amplitude Zero Auto Correlation sequenceCCSA . . . . . . . . . . . . . . . . . . . . . . China Communications Standards AssociationCD . . . . . . . . . . . . . . . . . . . . . . . . Collision DetectionCDMA . . . . . . . . . . . . . . . . . . . . . Code Division Multiple AccessCID . . . . . . . . . . . . . . . . . . . . . . . . Connection IdentifierCP . . . . . . . . . . . . . . . . . . . . . . . . . Cyclic PrefixCSMA . . . . . . . . . . . . . . . . . . . . . Carrier Sense Multiple AccessCTS . . . . . . . . . . . . . . . . . . . . . . . Clear To Send packet

DDFT . . . . . . . . . . . . . . . . . . . . . . . Digital Fourier TransformDFTs OFDMA . . . . . . . . . . . . . DFT spread OFDMA

EEDGE . . . . . . . . . . . . . . . . . . . . . Enhanced Data Rates for Global EvolutionE-UTRA . . . . . . . . . . . . . . . . . . . Evolved UTRAETSI . . . . . . . . . . . . . . . . . . . . . . . European Telecommunications Standards Institute

F

15

16 List of Abbreviations

FDMA . . . . . . . . . . . . . . . . . . . . . Frequency Division Multiple AccessFFT . . . . . . . . . . . . . . . . . . . . . . . Fast Fourier Transform

HHSDPA . . . . . . . . . . . . . . . . . . . . High Speed Downlink Packet Access

IICI . . . . . . . . . . . . . . . . . . . . . . . . . Inter Channel InterferenceIEEE . . . . . . . . . . . . . . . . . . . . . . Institute of Electrical and Electronics EngineersIFDMA . . . . . . . . . . . . . . . . . . . . Interleaved Frequency Division Multiple AccessIFFT . . . . . . . . . . . . . . . . . . . . . . Inverse Fast Fourier TransformISI . . . . . . . . . . . . . . . . . . . . . . . . . Inter Symbol Interference

LLAN . . . . . . . . . . . . . . . . . . . . . . . Local Area NetworkLFSR . . . . . . . . . . . . . . . . . . . . . . Linear Feedback Shift RegisterLTE . . . . . . . . . . . . . . . . . . . . . . . Long Term Evolution

MMAC . . . . . . . . . . . . . . . . . . . . . . Medium Access Control

NNACK . . . . . . . . . . . . . . . . . . . . . Negative Acknowledgment

OOFDM . . . . . . . . . . . . . . . . . . . . . Orthogonal Frequency Division MultiplexingOFDMA . . . . . . . . . . . . . . . . . . . Orthogonal Frequency Division Multiple Access

PPAPR . . . . . . . . . . . . . . . . . . . . . . Peak to Average Power RatioPRBS . . . . . . . . . . . . . . . . . . . . . . Pseudo Random Binary SequencePRMA . . . . . . . . . . . . . . . . . . . . . Packet Reservation Multiple Access

QQAM . . . . . . . . . . . . . . . . . . . . . . Quadrature Amplitude ModulationQPSK . . . . . . . . . . . . . . . . . . . . . Quadrature Phase Shift Keying

RRACH . . . . . . . . . . . . . . . . . . . . . Random Access ChannelRNG-REQ . . . . . . . . . . . . . . . . . Ranging RequestRNG-RES . . . . . . . . . . . . . . . . . Ranging ResponseRTS . . . . . . . . . . . . . . . . . . . . . . . Request To Send

SSIM . . . . . . . . . . . . . . . . . . . . . . . . Subscriber Identity ModuleSNR . . . . . . . . . . . . . . . . . . . . . . . Signal to Noise RatioSS . . . . . . . . . . . . . . . . . . . . . . . . . Subscriber Station

List of Abbreviations 17

TTDMA . . . . . . . . . . . . . . . . . . . . . Time Division Multiple AccessTTA . . . . . . . . . . . . . . . . . . . . . . . Telecommunication Technology AssociationTTI . . . . . . . . . . . . . . . . . . . . . . . . Transmission Time IntervalTU . . . . . . . . . . . . . . . . . . . . . . . . Typical Urban

UUE . . . . . . . . . . . . . . . . . . . . . . . . . User EquipementUCD . . . . . . . . . . . . . . . . . . . . . . . Uplink Channel DescriptorUMTS . . . . . . . . . . . . . . . . . . . . . Universal Mobile Telecommunication SystemUTRA . . . . . . . . . . . . . . . . . . . . . UMTS Terrestrial Radio Access

Chapter 1

Introduction

1.1 Background

1.1.1 3GPP

The 3rd Generation Partnership Project (3GPP) is a collaboration agreementthat was founded in December 1998. Its a co-operation between ETSI (Eu-rope), ARIB/TTC (Japan), CCSA (China), ATIS (North America) and TTA(South Korea). Several hundred companies from around the world participateas Individual Members. The scope of 3GPP is to make a globally applicablethird generation (3G) mobile phone system specification. This organization isto produce a complete set of applicable Technical Specifications and Reportsfor:

UMTS: 3G systems based on the evolved GSM core network and theUniversal Terrestrial Radio Access (UTRA) in FDD and TDD mode.

GSM including evolved GSM radio access technologies (GPRS, EDGE).

3GPP prepares and maintains specifications for several technologies, espe-cially:

GSM, GPRS, EDGE W-CDMA -FDD (Wide band CDMA Frequency Division Duplex) TD-CDMA -TDD (Time Division CDMA Time Division Duplex)

The good aspect of 3GPP is the centralization of the standards. Indeed, asingle organization for these technologies ensures global interoperability.

1.1.2 LTE

With enhancements such as HSDPA and Enhanced Uplink, the 3GPP radio-access technology will be highly competitive for several years. However, toensure competitiveness for the next 10 years and beyond, a long-term evolutionof the 3GPP radio-access technology needs to be considered.

1

2 Chapter 1. Introduction

The standardization department at LG (3GPP member), where this projectis conducted, is currently working on the 3G Long Term Evolution (LTE).

Important parts of such a long-term evolution include reduced latency, higheruser data rates, improved system capacity and coverage, and reduced cost forthe operator. In order to achieve this, an evolution of the radio interface aswell as the radio network architecture will be considered. Considering a desirefor even higher data rates and also taking into account future additional 3Gspectrum allocations the long-term 3GPP evolution will include an evolutiontowards support for wider transmission bandwidth than 5 MHz. At the sametime, support for transmission bandwidths of 5 MHz and less than 5 MHz will beinvestigated in order to allow for more flexibility in whichever frequency bandsthe system may be deployed. The main objectives of LTE are then:

Significantly increased peak data rate e.g. 100 Mbps (downlink) and 50Mbps (uplink)

Significantly improved spectrum efficiency

Reduced latency

Scaleable bandwidth

5, 10, 20 and possibly 15 MHz

1.25 and 2.5 MHz: to allow flexibility in narrow spectral allocationswhere the system may be deployed

Reasonable system and terminal complexity, cost, and power consumption

1.1.3 RACH

The Random Access Channel is a contention-based channel for initial uplinktransmission, i.e. from UE (User Equipment) to NodeB (base station). Thischannel can be used for several purposes. The RACH function is differentdepending on the technology of the system. The RACH can be used to accessthe network, to request resources, to carry control information, to adjust thetime offset of the uplink, to adjust the transmitted power, etc. It can evenbe used to transmit small amounts of data. Contention resolution is the keyfeature of the random access channel. Many UE can attend to access a samebase simultaneously, leading to collisions.

1.2 Problem definition

1.2.1 A new air interface

One of the main changes in the LTE system compared to UMTS is the airinterface. The way to modulate the signal is completely different. In the thirdgeneration systems, WCDMA (Wideband Code Division Multiple Access) isthe most widely adopted technology. We propose to highlight the characteristicitems of WCDMA:

1.2. Problem definition 3

User information bits are spread over a wide bandwidth by multiplyingthe user data with a spreading code. The use of variable spreading factorallows a variation of the bit rate (up to 2Mbps).

The carrier bandwidth is approximately 5 MHz. The chip rate used is 3.84Mcps. A network operator can deploy multiple 5 MHz bands to increasecapacity.

The frame length is 10ms. During this phase, the user data rate is keptconstant. However the data capacity among the users can change fromframe to frame.

In the LTE system, this will be very different. The new system will presentan OFDM based structure. Different candidates are likely to be the new air in-terface. So far, OFDMA (OFDM Access) is very likely to be used on downlink.However, on uplink, the decision is not made by now. IFDMA (InterleavedFrequency Division Multiplexing Access) and DFTs OFDMA (DFT spreadOFDMA) seems to be the two best candidates. We will explain the detailsof these modulations in the section 5.

1.2.2 Several possible scenarios

The RACH design for LTE is just initiated. A few contributions have beenshared and are discussed in the 3GPP meetings. This project intends to give anoverview of the actual discussion and to investigate some possible scenarios. Bynow, there is no unique solution. Everything is under reflection. The subject isstill widely open, hence there is no use considering a particular random accessscheme. The purpose of the project is to analyze a number of design parametersin order to find out the more probable configurations.

1.2.3 Requirements and RACH purpose

The LTE requirements concerning the RACH are different from the UMTS ones.While the RACH as defined in the 3G systems is mainly used to register theterminal after power-on to the network, the LTE RACH will have to deal withnew constraints.

In an OFDM based system, orthogonal messages can be sent. This leads toa new way of designing the physical layer. A major challenge in such a system isto maintain uplink orthogonality among UEs. Hence both frequency and timesynchronization of the transmitted signals from the UEs are needed. Frequencysynchronization can be achieved by fixing the UE local oscillator to the downlinkbroadcast signal. The remaining frequency misalignment at the Node B is dueto Doppler, which cannot be estimated nor compensated and hence requires nofurther consideration. However, the timing estimation has to be performed bythe NodeB when measuring the received signal. This can be achieved during arandom access. The UE can then receive a timing advance command from theNodeB and adjust its uplink transmission timing accordingly.

Consequently one purpose of the LTE random access procedure is to obtainuplink time synchronization.

4 Chapter 1. Introduction

1.2.4 Design

The project proposes to give an idea of the RACH design within the LTE system.By considering the 3GPP contributions and by analyzing the new physical layerproperties, a first approach can be completed.

1.3 Research approach

1.3.1 Multiple Access protocols

In order to investigate the design of the channel, it is first proposed to studya number of multiple access protocols. Methods such as ALOHA, SlottedALOHA, CSMA, PRMA, and Reservation ALOHA are highlighted and exam-ined. Certain protocols are used in existing standards. Their feasibility mainlydepends on the system.

1.3.2 Air interface analysis

Obviously, it appears essential to examine the candidates for the new air in-terface. A theoretical analysis of each candidate for the uplink transmission(IFDMA and DFTs OFDMA) has been achieved and is presented. These twomodulation schemes have a particular frequency structure and it is interestingto see how the signal is created.

1.3.3 Existing random access channels analysis

An efficient way to learn how to design the new channel was to examine theexisting procedures in several systems. Much time has been spent to analyzethe UMTS RACH procedure and the WiMax ranging procedure. This phasewas an essential part of the study and was extremely helpful to understand theproblematic.

1.3.4 List different scenarios

According to the 3GPP contributions on the subject and the previous analysis,a set of different scenarios has been achieved. These configurations try to fitthe LTE requirements and the 3GPP members ideas. Each situation is testedand compared to the others. The goal is to obtain primary results to help thedesign of the channel.

1.3.5 Implementation

All the simulations are performed using Matlab 7.1 with the Signal Processingand the Communication System Toolboxes. The purpose of the program shouldnot be restricted to the RACH transmission. Indeed, it should be used for fur-ther researches by the Standardization team at LG MobileComm. The programcontains a GUI which allows the user to tune some parameters before launchinga simulation.

1.3. Research approach 5

1.3.6 Evaluation and Analysis

The last part of the project consists obviously in evaluating each scenario. Thedesign of a RACH implementation has to balance a number of factors e.g. de-tection probability, false alarm probability, time offset estimation and latency.Within the limitations set by the air interface it is necessary to optimize theseparameters to system requirements.

Chapter 2

Multiple Access Protocols

2.1 General

The RACH is a shared channel, i.e. several UEs are allowed to use it at thesame time. The key problem is that many UE can attend to access a samebase simultaneously, leading to collisions. To make a transmission successful,interference must be avoided or at least controlled.

There exist two types of multiple access protocols: the conflict-free and thecontention protocols. Conflict-free protocols are those ensuring a successfultransmission, whenever made, for the reason that it will not be interfered byanother transmission. Conflict-free transmission can be achieved by allocatingthe channel resources (in time or frequency slots) to the users. Examples ofconflict free protocols are TDMA (Time Division Multiple Access) and FDMA(Frequency Division Multiple Access). In a contention scheme, a transmissionis not guaranteed to be successful. Hence, the protocol must stipulate a way toresolve conflicts so that all messages are eventually transmitted successfully.

2.2 The pure Aloha protocol

t

t

Waits for acknowledgmentRandom

Figure 2.1: The pure Aloha protocol

The pure Aloha protocol is the most basic in the family of the Aloha pro-tocols. Whenever a UE has a packet to send, it transmits it via the channel,hoping for no collision. If the transmission is not successful, each colliding userschedules its retransmission to a random time in the future. The randomness

7

8 Chapter 2. Multiple Access Protocols

helps to ensure that the packets do not continue to collide indefinitely. To knowif a packet has been received or not, the UE waits for an acknowledgment fromthe NodeB.

The pure Aloha protocol is not efficient in the sense that the vulnerableperiod is two packet long (see Figure 2.2 ). For this reason, collisions are veryprobable and the capacity of the channel considerably reduced. A simple wayto improve the principle is to use the Slotted Aloha protocol.

t

if another packet is sentduring this period: collision

Figure 2.2: The vulnerable period of the basic Aloha protocol

2.3 The Slotted Aloha protocol

The slotted Aloha protocol is characterized by a slotted channel, i.e. time isdivided into time slots and the UEs are restricted to start transmission at thebeginning of a time slot. The slot size has to be larger than the packet duration.This method reduces by half the vulnerable period and leads to much betterresults. It is used within the UMTS system and it is very easy to implement.One can notice that synchronization between users is necessary.

t

if another packet is sentduring this period: collision

Figure 2.3: The vulnerable period of the slotted Aloha protocol

2.4 The CSMA protocol

The CSMA protocols (Carrier Sense Multiple Access) are used for examplein Ethernet, AppleTalk and Wireless LAN. They are characterized by sensingthe channel before transmitting i.e. the transmission is performed only if thechannel is sensed idle. If the medium is sensed busy, the station will defer itstransmission to a later time. These protocols are very effective when the channel

2.4. The CSMA protocol 9

is not highly loaded and lead to excellent delay performances. However one canobserve that the method does not prevent from collision. Indeed, if two userssense the channel idle at the same moment and decide to transmit, a collisionwill happen. There exist two types of CSMA protocols: CSMA-CD (CollisionDetection) and CSMA-CA (Collision Avoidance).

A B

station range

Figure 2.4: The hidden terminal problem: Station A senses the channel as freeand hence decides to transmit. A collision happens at the node

In CSMA-CD, the users are able to detect interference among severaltransmissions (including their own) while sending data and abort trans-mission of their collided packets. In that way, the duration of an unsuc-cessful transmission is reduced.

While the Collision Detection mechanism can be performed in a wiredLAN, it can not be used in a wireless environment. Indeed, it is impossibleto assume that all the stations can hear each other (see Figure 2.4).

Moreover, such a method would require a full duplex approach, capable oftransmitting and receiving at once. A solution to overcome these problemsis to use the Detection Avoidance procedure (used within 802.11):

Before any transmission, stations listen to the medium.

A station willing to send a packet transmits a Request-To-Send packet(RTS). This packet contains information about the length of the datato be sent Tdata.

The receiver replies with a control packet called Clear-To-Send (CTS)including the same duration information Tdata. If the sender does notreceive the CTS packet, it does not transmit the data.

Every station hearing a RTS remains quiet during TCTS + Tdata.

Every station hearing a CTS remains quiet during Tdata.

CTS and RTS are very short packets. The probability of collision is hencevery small. In addition, this method reduces the probability of collisionwith a station which is hidden from the transmitter as described in theFigure 2.6.


RTCTdata1

CTSTdata1

t

t

Station A

NodeB

Station B

DATA

Tdata1

RTCTdata1

tDefer access

Figure 2.5: The basic detection avoidance procedure

A B

station range

CTS range

RTC range

Figure 2.6: The station A, which is not able to hear the signal from B, receivesthe CTS packet. The collision is avoided.

2.5 The Reservation protocols

Reservation schemes are designed to have the advantages of both the Aloha andthe TDMA approaches. They require a continuous broadcast of information(from the node) and a good synchronization between the different stations.Two types of reservation methods exist: the implicit reservation and the explicitreservation.

PRMA (Packet Reservation Multiple Access) is an implicit reservationprotocol. The principle, which combines contention and reservation, isvery simple. The frame is divided into equal time slots. First all terminalshave to use the slotted Aloha protocol for contention with other terminals.Once a station obtains a slot successfully (i.e. without collision), the slot isautomatically assigned to this station in the following frames. As soon asthe station stops using the slot, other stations can compete to reserve it. Incase of successful access, the NodeB has to broadcast the acknowledgment.Before transmitting, a station knows which slots are available by listeningto the broadcast information. An illustration of the procedure is provided

2.6. Temporary conclusion 11

in Figure 2.7.

R8 R2 R7

R5R7 R4R1 R2

R8 R2 R6 R3slot s1 s3s2 s4 s5 s6

frame i

frame i+1

frame i+2

active terminals

active terminals

8 4,5 2 7 - -

- 1 2 7 4 5

Available slot

R3 Reserved slot

4,5 requesting terminals

Figure 2.7: PRMA protocol in the case of 6 slots per frame

The reservation Aloha protocol is an explicit method of scheduling. In thesame way as PRMA, it combines contention and reservation. Here, theframe is divided into 2 distinct sections:

A contention phase which contains short time slots. In this period,terminals compete for reservation. The slotted Aloha technique isused to reduce the collision probability.

A reservation phase which consists of longer time slots. If a UEhas successfully obtained a slot during the contention phase, it cantransmit during the corresponding reserved slot.

A reservation list is created and every terminal has to keep it consistent. Afine synchronization is needed in order to avoid collisions. An illustrationof the procedure is provided in Figure 2.8

t

Reservationphase

Reservationphase

Contentionphase

Contentionphase

Figure 2.8: Reservation Aloha: example of transmission

2.6 Temporary conclusion

In an OFDM based system, it is achievable and very useful to introduce fre-quency multiplexing. The designer can consequently have the benefit of anadditional degree of freedom when trying to avoid collision probability. In thisway, several options are under considerations. Now slots are not only definedin the time domain. A slot is characterized by a time index and a frequency


band. As illustrated in Figure 2.9, a possible method to adopt is to reserve asub frame at regular intervals for random access. Then either several or all thefrequency sub bands are dedicated to the RACH access. Further considerationswill be detailed in a following section

Radio frame

subframe dedicatedto random access

RACH bursts Data

used either for contentionor data transmission

Figure 2.9: Temporary solution: a subframe is dedicated to random access

Chapter 3

Air interface candidates foruplink access

In a basic communication system, the data are modulated onto a single carrierfrequency. The available bandwidth is then totally occupied by each symbol.This kind of system can lead to inter-symbol-interference (ISI) in case of a fre-quency selective channel. The basic idea of OFDM is to divide the availablespectrum into several orthogonal sub-channels so that each narrowband sub-channel experiences almost flat fading. With OFDM, it is possible to haveoverlapping sub-channels in the frequency domain, thus increasing the trans-mission rate. OFDM systems have gained an increased interest during the lastyears. It is largely used by standards such as IEEE 802.11 (WiFi) or IEEE802.16 (WiMax), as well as in wired environment such as asymmetric digitalsubscriber lines (ADSL).

3.1 Basic OFDM

3.1.1 Properties and main advantages

The modulation technique used in an OFDM system helps to overcome theeffects of a frequency selective channel. A frequency selective channel occurswhen the transmitted signal experiences a multipath environment. Under suchconditions, a given received symbol can be potentially corrupted by a number ofprevious symbols. This effect is commonly known as inter-symbol interference(ISI). To avoid such interference, the symbols duration, T, has to be much largerthan the delay spread Tm (maximum amount of time between the first and lastmultipath signal at the receiver). But this would lead to poor efficiency in termsof transmission speed. In an OFDM system, the information is transmittedamong N different subcarriers, each with a transmission interval time multipliedby N (see Figure 3.1).

In this way, the system is N times more robust against ISI while the overalltransmission rate remains the same.

Also, the spectral efficiency of the OFDM modulation technique is excel-lent since the sub channels are overlapping. Indeed, in basic FDM (FrequencyDivison Multiplexing) a guard-band is left between the sub-channels to pre-

13

14 Chapter 3. Air interface candidates for uplink access

x


e2.pi.f1.t

e2.pi.f2.t

e2.pi.fN .t

R

N

In this way, the system is N times more robust against ISI while the overalltransmission rate remains the same.

Also, the spectral efficiency of the OFDM modulation technique is excellentsince the sub channels are overlapping. Indeed, in basic FDM (Frequency Divi-son Multiplexing) a guard-band is left between the sub-channels to prevent fromInter Carrier Interference (ICI). In order to improve the bandwidth efficiency,orthogonal frequency division multiplexing was proposed. Sub channels are over-lapping in the frequency domain but still are orthogonal, the basis functions arerepresented in Figure 3.1.

Chapter 2 Black TEAM

Demodulation. Symbols are transformed back to bits. The inverse of the estimated channelresponse is used to compensate the channel gain.

Deinterleaver (Interleaving inverse operation). The stream of bits fills the matrix columnby column. Then, the bits leave the matrix row by row.

Convolution decoder. The decoder performs the Viterbi decoding algorithm to generatetransmitted bits from the coded bits.

2.2 OFDM System

This section introduces OFDM and key system aspects are considered.

2.2.1 Evolution of OFDM

Frequency Division Multiplexing (FDM)

Frequency Division Multiplexing (FDM) has been used for a long time to carry more than onesignal over a telephone line. FDM divides the channel bandwidth into subchannels and transmitsmultiple relatively low rate signals by carrying each signal on a separate carrier frequency. Toensure that the signal of one subchannel did not overlap with the signal from an adjacent one,some guard-band was left between the different subchannels. Obviously, this guard-band led toinefficiencies.

Orthogonal Frequency Division Multiplexing (OFDM)

In order to solve the bandwidth efficiency problem, orthogonal frequency division multiplexingwas proposed, where the different carriers are orthogonal to each other. With OFDM, it is pos-sible to have overlapping subchannels in the frequency domain, thus increasing the transmissionrate. The basis functions are represented in Figure 2.3. This carrier spacing provides optimalspectral efficiency.!"#$%&'()(*(+"&,'-((

(

./

(

!"#( $ $"# % %"# & &"# ' '"# #(!"'(

(!"%(

!(

!"%(

!"'(

!")(

!"*(

$(

(

(

(

(

(

(

(

(

(

(

(

(

!"#$%&'()*+',%-./#/012'314"4'5$06-"/04'"0'10',!78'494-&:'

(

!"!# $%&'()*+,(-#

(

01(,'2&'(%,(3#41%#41(,'%",5,1#64%-(%"',75"(%"&(%'#1834884,1(,9&'(%"&(:"#11&6(#(:-:64:($'&;4,68(,;(%"&(;'#3&(#12($7%(#(:,$-(,;(%"&3(41(%"&(

>&5411415(,;( %"&( ;'#3&=(+"48(@466( #68,(3#?&( %"&(,7%$7%( ;',3( %"&( 0AB+($&'4,24:=(+"&(8451#6(

@466(1,@(#$$'(#8(8",@1(>&6,@=(

(

(

( ( ( (CD ;

&0?,68(,;(%"&(;'#3&(#12($7%(#(:,$-(,;(%"&3(41(%"&(

>&5411415(,;( %"&( ;'#3&=(+"48(@466( #68,(3#?&( %"&(,7%$7%( ;',3( %"&( 0AB+($&'4,24:=(+"&(8451#6(

@466(1,@(#$$'(#8(8",@1(>&6,@=(

(

(

( ( ( (CD ;

&0?,68(,;(%"&(;'#3&(#12($7%(#(:,$-(,;(%"&3(41(%"&(

>&5411415(,;( %"&( ;'#3&=(+"48(@466( #68,(3#?&( %"&(,7%$7%( ;',3( %"&( 0AB+($&'4,24:=(+"&(8451#6(

@466(1,@(#$$'(#8(8",@1(>&6,@=(

(

(

( ( ( (CD ;

&0?,68(,;(%"&(;'#3&(#12($7%(#(:,$-(,;(%"&3(41(%"&(

>&5411415(,;( %"&( ;'#3&=(+"48(@466( #68,(3#?&( %"&(,7%$7%( ;',3( %"&( 0AB+($&'4,24:=(+"&(8451#6(

@466(1,@(#$$'(#8(8",@1(>&6,@=(

(

(

( ( ( (CD ;

&0?,68(,;(%"&(;'#3&(#12($7%(#(:,$-(,;(%"&3(41(%"&(

>&5411415(,;( %"&( ;'#3&=(+"48(@466( #68,(3#?&( %"&(,7%$7%( ;',3( %"&( 0AB+($&'4,24:=(+"&(8451#6(

@466(1,@(#$$'(#8(8",@1(>&6,@=(

(

(

( ( ( (CD ;

&0?

3.1. Basic OFDM 15

IFFT CP channel FFTCP

X(n) x(k) s(k) r(k) y(k) Y(n)

Figure 3.3: Basic OFDM transmission (CP stands for Cyclic Prefix)

x(k) =1N

N1n=0

X(n)e2pijknN k = 0..N 1 (3.1)

At the receiver side, the data is recovered by performing FFT on the receivedsignal,

Y (n) =1N

N1k=0

y(k)e2pijknN n = 0..N 1 (3.2)

3.1.3 The need of a cyclic prefix

In an typical wireless environment, the channel has a finite impulse response.We note tm the maximum delay of all reflected paths of the OFDM transmittedsignal, see Figure 3.4.

Chapter 2 Black TEAM

An N-point FFT only requires Nlog(N) multiplications, which is much more computationallyefficient than an equivalent system with equalizer in time domain.

2.2.4 Cyclic Prefix

In an OFDM system, the channel has a finite impulse response. We note tmax the maximumdelay of all reflected paths of the OFDM transmitted signal, see Figure 2.5.

channel impulse response

t

0 tmax

Figure 2.5: Channel impulse response

Cyclic prefix is a crucial feature of OFDM to combat the effect of multipath. Inter symbolinterference (ISI) and inter channel interference (ICI) are avoided by introducing a guard intervalat the front, which, specifically, is chosen to be a replica of the back of OFDM time domainwaveform.

Figure 2.6 illustrates the idea.

FRAMECP

t

tc > tmax0

Figure 2.6: Adding a cyclic prefix to a frame

From above expressions the subcarrier waveforms are now given by

s(k) =

{x(k + N) M k < 0x(n) = 1

N

N1k=0 X(k).e

j2pik nN 0 k < N 1 (2.3)

The idea behind this is to convert the linear convolution (between signal and channel re-sponse) to a circular convolution. In this way, the FFT of circulary convolved signals is equivalentto a multiplication in the frequency domain. However, in order to preserve the orthogonalityproperty, tmax should not exceed the duration of the time guard interval. As shown below, oncethe above condition is satisfied, there is no ISI since the previous symbol will only have effect

Page 13

h(t)

Figure 3.4: Typical channel impulse response

The cyclic prefix is a crucial feature of OFDM to combat the effect of mul-tipath. Inter Symbol Interference (ISI) and Inter Channel Interference (ICI)are avoided by introducing a guard interval at the front, which, specifically, ischosen to be a replica of the back of OFDM time domain waveform. Figure 3.5illustrates the idea.

The idea behind this is to convert the linear convolution (between signal andchannel response) to a circular convolution. In this way, the FFT of circularyconvolved signals is equivalent to a multiplication in the frequency domain andthe data can be recovered properly (Y (n) depends only on X(n) and not onX(i) with i 6= n).

Y (n) = DFT (y(k)) = DFT (IDFT (X(n)) h(k) + e(k)) (3.3)= X(n).DFT (h(k)) +DFT (e(k)) (3.4)


FrameCP

t0 tcp > tmax

Figure 3.5: Adding a cyclic prefix to a frame

= X(n)H(n) + E(n) (3.5)(3.6)

However, in order to preserve the orthogonality property, tm should notexceed the duration of the guard interval. As shown below, once the abovecondition is satisfied, the linear convolution associated with the channel impulseresponse becomes a circular convolution. If we note x(n) the signal with CPand M the length of the channel impulse response, we get:

y(k) = x(n) h(k) (3.7)

=Mn=1

h(k)x(k n) (3.8)

=Mn=1

h(k)x(k n)N (3.9)

= x(k) h(k) (3.10)

Thus, there is no ISI since the previous symbol will only have effect oversamples within [0, tmax] and orthogonality is maintained so that there is noICI.

3.1.4 OFDMA: OFDM for multiple access

A possible usage of the OFDM technique is a frequency multiplexing amongseveral users. By assigning different sets of sub carriers to different users, trans-missions will be orthogonal to each other. Then, before transmitting any data,each user has to know what frequency band he is allowed to use and map itsdata symbols onto the corresponding sub carriers. The Figure 3.6 illustrates theidea.

Several transmission schemes are related to the OFDM modulation. Forthe LTE system, not only the basic OFDMA scheme is likely to be chosen forthe uplink transmission. Two other schemes are potential candidates: DFTsOFDMA and IFDMA.

3.2. DFTs OFDMA 17

IFFT

user 1

modulation spreading S/P

00000000

IFFT

user 2

modulation spreading S/P

00000000

Figure 3.6: Frequency multiplexing using OFDM

3.2 DFTs OFDMA

The structure of DFT spread OFDMA transmitter is represented in Figure 3.7.

IFFTmodulation FFTS/P mapping

Figure 3.7: DFTs OFDMA transmitter

The data symbols are spread by performing an FFT before being mapped tothe sub carriers. This means that each sub carrier carries a portion of superposedDFT spread data symbols. Various mappings are possible, which is illustratedin Figure 3.8. The allocation of the spread data can be either localized, pseudorandomized or equidistant.


0

IFFT

FFTS/P

IFFTFFTS/P FFTS/P IFFT

localized equidistant randomized

Figure 3.8: Sub carriers allocation schemes

3.3 IFDMA

Whereas a DFTs OFDMA transmitter builds the signal in the frequency domain,the IFDMA signal is realized in the time domain. The process is very simple.If N is the number of used sub carriers among Ntot sub carriers, then the Nsymbols data are repeated Nrep = Ntot/N times. This creates a comb-shapedspectrum as illustrated in Figure 3.9. The users are separated by assigning adifferent phase shift to each one.

We can notice that the process is equivalent to a DFTs OFDMA processwith equidistant mapping. Indeed, if we note xs(n) the input symbols, X(m)the signal after FFT, Xmap(l) the signal after mapping and x(k) the signal afterIFFT, we get:

x(k) =Ntot1l=0

Xmap(l)e2pijlkNtot (3.11)

=N1m=0

Xmap(mNtotN

)e2pijmkNtotNtotN (3.12)

=N1m=0

X(m)e2pijmkN for k = 0..Ntot 1 (3.13)

= xs(k) for k = 0..N 1 (3.14)

and since e2pijmkN = e

2pijm(k+N)N , we get :

x(k + iN) = xs(k) for k = 0..N 1 and for i = 0..NtotN

1

3.4. General comparison 19

s1 s2 sN

sKN symbols

s1 s2 sN s1 s2 sN s1 s2 sN s1 s2 sN

Nrep repetitions

Time domain

S

f

Frequency domain

Nrep -1 free subcarriers N used subcarriers

modulation N/Ntot repetitionsfrequency

shift CPN Ntot

Figure 3.9: IFDMA signal

3.4 General comparison

In this section, we propose to highlight some characteristics of the mentionedschemes. One of the parameters to take into account is the Peak to AveragePower Ratio (PAPR). Reducing the PAPR of the uplink transmission is a ma-jor question, since the effective usage of power amplifier is one of the importantfactors on the UE side.

IFDMA:

Simple transmitter implementation Provides wide band diversity High inter-user interaction Low PAPR

DFTs OFDMA:

Localized allocation:Poor frequency diversity

Less inter-user interference (edges only interact)

PAPR OFDMA > PAPR DFTs OFDMA localized > PAPR IFDMA

Randomized allocation:


The PAPR is as high as with a basic OFDMA

Provides frequency diversity

Equidistant allocation:See IFDMA

Chapter 4

Existing random accessprocedures

4.1 RACH procedure within UMTS

This section proposes to describe the RACH procedure within UMTS. Under-standing the mechanism in the current working system is very useful in order toacquire a practical knowledge on the subject. Even if the air interface is totallydifferent, some aspects are still exploitable.

4.1.1 RACH purpose

Within a UMTS system, RACH is used for initial access, i.e. the procedurewhere the UE sends a first message to the network. It is characterized bythe small amount of transmitted data. The initial access is associated with aconnection request message containing the reason of the request. There existseveral reasons for sending a connection request, as defined in [1]. The differenttypes of calls are listed below:

Originating call: the UE wants to setup a connection after power-on to thenetwork. It can be a request for speech connexion or for data streamingfor example.

Terminating call: the UE answers to paging. The downlink Paging Chan-nel (PCH) is used to reach a mobile station which is not currently main-taining a connection with the network. Hence, if the UE receives a requestto set a connection over this channel, it answers over the RACH and setsup a connection.

Registration: the UE wants to register only to location update. In thisway, the network will, after the process, be aware of the mobiles position.

4.1.2 Collision handling

The RACH is a common channel. Due to simultaneous access of several users,collisions may occur such that the initial access message cannot be decoded by

21

22 Chapter 4. Existing random access procedures

the network. In order to prevent from collision on the message of interest, aprocedure using preambles is applied. A preamble is a short signal which is sentbefore the transmission of the higher layer message (a preamble is about tentimes shorter than a request message). In other words, a RACH access consistsof two steps: the transmission of a set of preambles and the transmission of amessage. Contention can only happen on the preambles. The UE persists insending preambles until it receives an Acquisition Indicator (AI) from the NodeBon the AICH (Acquisition indicator channel) indicating that the network hascorrectly detected the preamble. Subsequently, the transmission of the requestmessage is contention free (except the rare case when 2 UEs would have sentthe same signal simultaneously). The details of the procedure will be explainedlater.

The method is depicted in the Figure 4.1.

timing offset

Access slot5120 chips

1,33 ms

Preamble4096 chips

Acquisition Indicator4096 chips

message(10 or 20 ms)

Figure 4.1: Overview of the RACH procedure within UMTS

The used access method is a type of slotted Aloha protocol. The RACHtransmission, both preambles and request message, has to start at the beginningof a slot. Moreover, in case the whole procedure fails, i.e. the maximum numberof preambles has been sent without receiving an acknowledgment message, another attempt is performed after a back-off delay.

A frame in a UMTS system is 10 ms long (38400 chips). The time axis forthe RACH and AICH is divided into time intervals: the access slots. Thereare 15 access slots per two frames. An access slot hence lasts 1.33 ms whichcorresponds to 5120 chips. Every preamble (1ms long or 4096 chips) is sentat the beginning of an access slot. From a particular UE point of view, notevery access slot is available for random access. Depending on its priority level(stored in the UE SIM card), the network decides whether the UE is allowedto use certain access resources. For instance, it can be desirable to prevent UEfrom sending requests in a case of emergency. Therefore, the UE is informedover a broadcasting channel (BCH) which access slots it is allowed to use. Also,the timing offsets to use between two preambles and between the last preambleand the message are signaled by the NodeB.

4.1.3 The power ramping technique

In order to enhance the success probability of a preamble retransmission andhence reduce the access delay, the transmission power is increased after everyunsuccessful attempt. This technique is called power ramping. This method

4.1. RACH procedure within UMTS 23

permits also to acquire information about the power to apply regarding therequest message. The procedure is depicted in Figure 4.2.

message(10 or 20 ms)

P

Pm p

Figure 4.2: Power ramping technique

Before any RACH access, the downlink power level is measured from theBCH. The initial power level is then computed from the measurement. Aftereach attempt, the UE updates its transmission power according to:

Pi+1 = Pi +P (4.1)

where P is the power ramp step in dB.When the UE receives an AI, the random access message is sent 3 or 4 access

slots after the last preamble (depending on information received over BCH). Themessage part is transmitted with a power equal to:

Pmessage = Plastpreamble +PMP (4.2)

Where PMP is a power step applied to ensure a correct reception of themessage of interest.

4.1.4 Format of the preambles

A preamble is a short signal which consists of 4096 chips. It does not containthe identity of the user. It is simply a sequence of 256 repetitions of Hadamardcodes of length 16. Hence, there exist 16 different preamble signatures (one foreach code). The preamble signatures are listed in Table 4.1.4:Also a scrambling code related to the cell identity is applied to the preamble.Before any attempt, the UE selects randomly a signature and sends the relatedpreamble. If the base station detects successfully the preamble, it can sendback an AI. This indicator contains a replica of the preamble so that the UEcan be aware of its target. The AI can take 2 values (positive acknowledgmentACK or negative acknowledgment NACK) depending on whether the messagetransmission is allowed or not. Since a NodeB is able to detect two differentsignatures in the same access slot, a problem occurs only when two UEs haveselected the same signature and the same access slot which is a very unlikelycase.


Preamble Value of nsignature 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15P0(n) 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1P1(n) 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1P2(n) 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1P3(n) 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1P4(n) 1 1 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1P5(n) 1 -1 1 -1 -1 1 -1 1 1 -1 1 -1 -1 1 -1 1P6(n) 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 1 1P7(n) 1 -1 -1 1 -1 1 1 -1 1 -1 -1 1 -1 1 1 -1P8(n) 1 1 1 1 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1P9(n) 1 -1 1 -1 1 -1 1 -1 -1 1 -1 1 -1 1 -1 1P10(n) 1 1 -1 -1 1 1 -1 -1 -1 -1 1 1 -1 -1 1 1P11(n) 1 -1 -1 1 1 -1 -1 1 -1 1 1 -1 -1 1 1 -1P12(n) 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 1 1P13(n) 1 -1 1 -1 -1 1 -1 1 -1 1 -1 1 1 -1 1 -1P14(n) 1 1 -1 -1 -1 -1 1 1 -1 -1 1 1 1 1 -1 -1P15(n) 1 -1 -1 1 -1 1 1 -1 -1 1 1 -1 1 -1 -1 1

Table 4.1: Hadamard codes of length 16

4.1.5 Format of the message

The message can be either 10ms or 20ms. Spreading and scrambling are ap-plied to the message part. The message part scrambling code has a one-to-onecorrespondence to the scrambling code used for the preamble part. Spreadingis performed using Orthogonal Variable Spreading Factor (OVSF) codes. Thecode-tree is showed in the Figure 4.3.

Spreading factors from 256 to 32 have been defined to be possible for themessage part. A direct mapping between the preamble signature and the mes-sage part spreading code node is applied. Each preamble signature points to oneof the 16 nodes in the code-tree that corresponds to SF = 16 (Spreading Factor).The sub-tree below the specified node is used for spreading of the message part.Explicitly, the message is spread by the channelization code Cch,SF,m where SFis the spreading factor and m = SF. s16 (s is the signature index).

4.1.6 RACH access procedure

This section describes step by step the procedure for the RACH access in theUMTS system. We first detail the MAC approach, and then the physical pro-cedure is analyzed. Before the initiation of the RACH procedure, the followinginformation is received from the higher layers:

The preamble scrambling code

The message length (either 10ms or 20ms)

The number of slots between the last preamble and the message (3 or 4)

4.1. RACH procedure within UMTS 25

C1,0 = (1)

C4,0 = (1,1,1,1)

C2,1 = (1,-1)

C2,0 = (1,1)

C4,1 = (1,1,-1,-1)

C4,2 = (1,-1,1,-1)

C4,3 = (1,-1,-1,1)

SF = 1 SF = 2 SF = 4 SF = 8

Figure 4.3: Orthogonal Variable Spreading Factor (OVSF) codes

The set of available signatures (among the total set of 16), the set of avail-able access slots, the persistence value Pi (used in the MAC procedure).

The power ramping step P in dB

The maximum number of preamble retransmission Nmax The maximum number of preamble cycles Mmax The initial preamble power computed from the downlink power fading.

The power offset Pmp between the last preamble and the message in dB

The backoff interval when a NACK is received (in number of 10ms inter-vals)

From the MAC (Medium Access Control) point of view, the procedure canbe illustrated by the chart flow in Figure 4.4:

Here each cycle corresponds to the transmission of a set of preambles. Acycle can end either because a positive or negative acknowledgment is received,or because the whole set has not been detected. If an ACK is received, theprocedure ends, otherwise, the procedure restarts from the beginning after acertain amount of time.

Different backoff timers are applied. For instance, a random timer (innerloop) is used after an unsuccessful cycle, which is part of the slotted Alohaprotocol.

In order to examine the RACH procedure in detail, the physical randomaccess procedure is performed as illustrated in Figure 4.5.

One can notice that the NodeB can send a negative acknowledgement evenif the transmission was successful. This negative acknowledgement is used whena congestion situation occurs in the network at the present time. In this casethe physical procedure ends and a new cycle is programmed some time


start

get RACH parameters

Nmax, Mmax, NB

M=0

info

send message

end

M=M+1

M < Mmax

Physical RACH

procedure

enddraw random integer k andwait k x 10ms

wait NB x 10ms

draw random integer k andwait k x 10ms

NACKNo ACK

ACK

Figure 4.4: MAC procedure for a RACH access with UMTS

4.2. Ranging procedure within WiMax 27

start

get RACH parameters

Nmax, Pinit , ...

N=0

N=N+1

N < Nmax reportNo ACK

info

Report ACK

Transmit a preamble in the next available

access slot

NACK

P = P + P

No ACK

ACK

Report NACK

P = Pinit

Figure 4.5: Physical procedure for a RACH access with UMTS

4.2 Ranging procedure within WiMax

4.2.1 Wimax overview

WiMAX (Worldwide Interoperability for Microwave Access) is a technology forwireless broadband. It is a suite of standards for fixed, portable and mobilepoint-to-multipoint wireless access. Theses standards emerge from IEEE 802.16Work Group. Regarding its general aspects, WiMAX has a range up to 30 milesand typically covers the 10 to 66 GHz range. It is an OFDM based system andcan support variable bandwidth sizes between 1.25 and 20 MHz.

The IEEE 802.16 standard was initially created to provide a flexible low-costalternative for the last mile broadband connectivity which is currently handledby wired solutions (cables or DSL). An evolved version, IEEE 802.16e, intendsto add mobility to this new service. Since it employs the same modulationtechnique as the LTE system and can provide mobility for high speed commu-nication, it represents an interesting base for the project.

This section proposes to highlight the MAC and physical layer of the 802.16estandard regarding the ranging procedure. As defined in the standards, themobile user will be called Subscriber Station (SS).


4.2.2 Ranging overview

Within 802.16e, a ranging procedure is supported to synchronize the SS withthe base station. Ranging, such as the RACH access within UMTS, intends toresolve contention. The process aims to align the SSs transmissions with theBS receive frame and acquire power adjustments. Initial ranging is part of thenetwork entry procedure. A general chart flow is depicted in Figure 4.6.

Registration

Obtain UL parameters

Scan for DL

channel

DL sync established

UL parameters

acquired

Ranging

BS

SS wave tript

t

Obtain the available slots for ranging

Obtain the timing offset, power adjustment and basic CID

Obtain the basic capabilities, secondary CID, set up the connexion

Figure 4.6: Network entry procedure

Before joining the network, the SS first scans for a downlink signal from thebase station and synchronize to it, i.e. the beginning of the slots and framesare identified with an offset corresponding to the signal trip delay. Then, thebroadcasted uplink channel descriptor (UCD) is detected to obtain the trans-mission parameters and the index of the ranging contention slots. After that,the actual ranging procedure can take place. The aims of this procedure are toalert the BS to the presence of the SS, to obtain a time and a power adjustmentand to gain a basic Connection Identifier (CID). Subsequently, the registrationprocedure is performed to obtain the secondary CID and set up the connection.

After the connection is established, the channel conditions may change, aswell as the distance between the base and the SS and periodic ranging is neces-sary to update the synchronization parameters.

4.2.3 Ranging procedure within theWirelessMANOFDMAPHY

There exist several candidates regarding the physical layer in the 802.16e stan-dard. This section focuses on the OFDMA PHY given that it is very similar tothe LTE system.

Firstly, a dedicated ranging channel is defined. Collision between data trans-missions and random accesses can not occur, i.e. ranging access does not in-terfere with data. To illustrate this concept, the frame structure for the TDDmode is depicted in Figure 4.7.


Figure 4.7: WiMAX TDD mode

The ranging channel is composed of one or more groups of six adjacent subchannels. In an OFDMA based structure, a slot is not only defined in timeas within UMTS but is also characterized by a frequency band. Therefore, aranging request opportunity (or ranging slot) is identified by:

its time interval index the group of sub channels (as illustrated in Figure 4.8).

6 subchannels

0 1 2

3

6

9

4

7

10

5

8

11

empty

t

f

ranging allocation

ranging slot number

Figure 4.8: Ranging opportunities

The bandwidth of a ranging signal corresponds to one group of 6 sub chan-nels. In other words, the number of orthogonal opportunities within one timeinterval is equal to the number of groups. In addition, in the same way asHadamard codes are used within UMTS, a set of pseudo-random ranging codesis specified. A BS is then supposed to be able to detect two ranging attemptsif they differ either in time, frequency or code.

The steps of the ranging procedure are illustrated in Figure 4.9:

After acquiring downlink synchronization and transmission parameters,the SS selects randomly a ranging slot and a ranging code and sends it to


wait for initial ranging opportunity

Send an initial ranging code in a ranging slot

Acquire downlink synchronization

and transmission parameters

wait for anonymous request response

noresponse response

success continue

adjust parameters

adjust parameters

wait for ressources

random backoff

send RNG request with the proper

adjustments

Figure 4.9: OFDMA PHY ranging procedure


the BS. The request is anonymous in the sense that the SS identity is notpresent in the signal.

In case of successful detection, the BS broadcasts a ranging response mes-sage that announces the received ranging code and the ranging slot wherethe code has been detected. In this way, the SS can identify the responsewhich corresponds to its request. Timing and power correction are alsoincluded in the response, as well as a status notification. The status canbe either success or continue:

Success status: the BS subsequently provides bandwidth allocationfor the SS to send a Ranging Request (RNG-REQ) with the appropriatetiming and power adjustments.

Continue status: The SS adjusts its timing offset and transmissionpower based on the response information, selects a new ranging slot andranging code and starts again the ranging process.

After sending the RNG-REQ, the SS waits for the RNG-RSP (response)containing its primary CID.

The procedure can end if the maximum transmission power or the maximumnumber of attempts has been reached.

4.2.4 Signatures

The codes are binary pseudo random codes generated by a PRBS (Pseudo Ran-dom Bit Sequence) generator. This kind of signature provides good cross corre-lation properties. The polynomial for the PRBS generator is 1+x4+x7+x15. Itis initialized by the seed 0, 0, 1, 0, 1, 0, 1, 1, s0, s1, s2, s3, s4, s5, s6 where s6 is themost significant bit, and s6 : s0 = UL IDcell (cell identity which is part of thetransmission parameters broadcasted by the BS). The Figure 4.10 representsthe PRBS in the case UL IDcell= 0.

ck

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

00 1 0 1 0 1 1 0 0 0 0 0 0 0Initializationsequence

Figure 4.10: The PRBS generator

The length of each code is 144 bits. For instance, the first 144 bit code isobtained by clocking 144 times the pseudo noise generator. With UL IDcell =0,the first bits of the code are 00110000010001... Then, the next ranging code isachieved by taking the next 144 outputs (145th to 288th) of the PRBS, etc.

The number of available signatures is 256. They are divided into 3 distinctgroups which correspond to 3 different types of request:


Initial ranging Periodic ranging Bandwidth request

4.2.5 Ranging signal

The initial ranging signal consists of two consecutive and identical OFDM sym-bols. In order to prevent from discontinuity between the two OFDM symbols, amethod using post and cyclic prefix is performed. A ranging frame is illustratedin Figure 4.11.

ranging symbol k CP ranging symbol k CP

Figure 4.11: Initial ranging signal

Each signal is 144bits long and is modulated using BPSK (one symbol perbit). The symbols are mapped over 6 adjacent sub channels (one group of subchannels). As defined in [2], a sub channel is composed of 24 sub carriers. Thetotal number of used sub carriers is hence 144 and the block scheme at thetransmitter side is depicted in Figure 4.12.

BPSKPRBSS/P

24 subcarriers

6 subchannels

mapping IFFTex:2048

144 bits

144 subcarriers

144 symbols

P/S

Figure 4.12: Initial ranging block scheme

Chapter 5

RACH design within LTE

This section intends to highlight the current discussion topics regarding therandom access in the future LTE system. As the design investigation is justinitiated, several solutions are proposed by the 3GPP members and there is notany exclusive solution yet.

5.1 Discussion on the RACH purpose

5.1.1 Timing adjustments

3GPP members seem to agree on the fact that one of the main purposes ofthe random access is to obtain fine time synchronization. The synchronizationprocedure prior to the random access only gives access to the slot and framesynchronization in the downlink. In other words, the mobile receives from abroadcast signal the start and the end of slots and frames but the transmissiondelay implies a time shift between the transmission and the reception of thebroadcast signal. Hence, the mobile can not estimate when to send its data sothat the NodeB receives them at the beginning of a slot.

The RACH procedure could be a solution to adjust the timing offset at theUE side by informing the UE how to compensate for the round trip delay.

t

UEwave trip

NodeB

t

sync signal RACH signal

timing offset

Figure 5.1: Timing offset caused by the round trip delay

After a successful random access procedure, the NodeB and the UE shouldbe synchronized within a fraction of the uplink cyclic prefix. In this way, the

33

34 Chapter 5. RACH design within LTE

subsequent uplink signals could be correctly decoded and would not interferewith other users.

5.1.2 Power adjustments

In WCDMA, the power adjustment is performed using power ramping. Implic-itly, the power to apply relies on the last preamble power. There is no feedbackfrom the NodeB to inform the UE about its power adjustments. In the E-UTRArandom access, the use of power ramping is still under consideration. This willbe explained under section 4.1.3. In case power ramping is not performed, onecan imagine a power adjustment procedure during the random access in orderto obtain the uplink power to apply for the very next transmission:

The UE either estimates a suitable transmission power from the open-looppower control or transmits with maximum power.

In case of successful detection, the NodeB estimates the power adjustmentsto apply to the uplink signal and includes this estimation in its acquisitionresponse.

A proper power adjustment would permit an optimal amount of transmittedenergy and hence a better battery life.

5.1.3 Resource request

Prior to sending any data, the UE has to obtain resources for transmission. TheNodeB acts as a scheduler and provides the UEs with scheduling information.The E-UTRA random access could be an approach to request bandwidth andtime resources. One can imagine several possible solutions:

The RACH procedure is only used to acquire timing and/or power adjust-ments. The only purpose is to be synchronized within a fraction of thecyclic prefix. Then a subsequent resource request has to be performed inorder to acquire resource reservation. The resource request procedure isstill contention-based but the system is synchronized.

NodeB

UE

broadcastsignal

(slots and frames)RACH signal

Response with timing

adjustments

Resource request

Resource assignment

unsynchronized synchronized

Figure 5.2: Resource request after being synchronized

The random access signal is composed of two distinct and independentparts: the RACH signature and the RACH message. The signature is

5.2. Discussion on the transmission method 35

a pseudo-random code. It allows several decodable random accesses inthe same frequency band and time slot. On the contrary, the messagepart contains information. For instance, this information can include aresource request (amount of data to transmit). At the NodeB side, onlythe signature is used for detection. In case of successful detection, themessage part is decoded and analyzed. If the network presents availableresources, the NodeB will send a resource grant to the UE. The signaturepart can also be used as reference symbols (pilots) for the demodulationof the message part.

Signature(a code)

Message(with resource

request)

Random access signal

Figure 5.3: The resource request could be part of the RACH signal

The random access signal is only composed of a signature but a signatureimplicitly points to a certain resource request. One can envisage severalsets of signatures, with each set corresponds to a particular frequencyband and time interval to transmit the following data. Then, in case ofsuccessful detection, both the UE and the NodeB implicitly know whichresource will be subsequently used.

set 2

set 3set 1

set n-1

set n

1

4

3

2

8

7

6

5

t

f

resources fordata transmission

Signatures

Figure 5.4: One to one correspondance between the sets of signature and theavailable resources

5.2 Discussion on the transmission method

5.2.1 Signature and payload

At least a RACH signal should contain a code sequence, or signature, whichidentifies the random access attempt. As was suggested in the previous section,a payload could be included. This payload could be composed of the user


identity, the desired amount of resources, the reasons of request, the nature ofthe access (initial or periodic), etc. In case the signature and the message aretransmitted over the same subcarriers, the NodeB can use the signature as apilot to decode the message.

5.2.2 Transmission band assigned to random and reserva-tion access channel

It has been agreed that orthogonal contention based and scheduled access chan-nels are used. Two options are envisaged when designing the transmission band-width allocation to random and reservation access. The first one consists in as-signing the whole given system bandwidth to the random access during specifiedsub frames. In the second option, only a part of the given system bandwidth isallocated to RACH access. This is depicted in the following figure:

t

RACHaccess

(a) (b)

Figure 5.5: Transmission band assigned to random acces

In either of the two options, the random access should not exceed the dura-tion of one TTI (Transmission Time Interval). A TTI corresponds to 0.5 ms. Aframe is 10 ms and is hence composed of 20TTI. The selection of the assignmentmethod should be decided regarding the given system bandwidth. Indeed, whenthe given bandwidth is 5 MHz or narrower, the second option is not achievableand the first one seems appropriate. Meanwhile, in case of wider system band-width, i.e. 10 or 20 MHz, the option (b) is achievable. The bandwidth of arandom access should be large enough to allow a good detection rate. Indeed,a wider transmission bandwidth allows having a diversity benefit. However,when the traffic load is high, the plurality of the available frequency bands forrandom access can permit fewer collisions. An adaptive assignment would bea clever but complex solution. The NodeB could estimate the traffic load andthen broadcast the assignment scheme to the UE willing to access the network.For instance, in a given system bandwidth of 10MHz, the NodeB could decideto allocate:

5 MHz to each random access in case of low traffic load (2 available bandsper RACH sub frame).

1.25 MHz to each random access in case of high traffic load (8 availablebands per RACH sub frame).

5.3. Signal parameters 37

5.2.3 Transmission bandwidth allocated to a UE randomaccess

With the aim of allowing for an accurate timing estimation, the signal band-width has to be chosen wide enough. In order to recover the following signalsemitted by the UE, an accuracy of less than a CP is needed. We will see thata CP corresponds to a duration order of 1 s. This leads to a required signalbandwidth in the order of 1 MHz. Hopefully, this can be supported by thenarrowest spectrum allocation in the E-UTRA (1,25 MHz).

5.2.4 Transmission duration and random access period

A random access signal is transmitted during a specified sub frame, i.e. a RACHsub frame, which lasts 0.5 ms. To avoid interferences with the next following subframe, guard interval is required, i.e. the random access signal is made shorterthan 0.5 ms. The goal is to prevent the NodeB from receiving a RACH signalduring a sub frame dedicated for data transmission. The duration of the guardinterval should depend on the cell size. A larger cell size implies a longer roundtrip delay between the downlink broadcasted signal and the uplink transmissionand this time shift has to be taken into account.

RACH signal

1 TTI

guard interval > maximal round trip delay

t

subframe dedicated to data tansmission

Figure 5.6: The necessary guard time

For instance, a UE-NodeB distance equal to 10km corresponds to a roundtrip delay of 66.67 s (10e323e8 ). A typical guard time would be hence in theorder of 70 s.

The LTE system is also supposed to be operational for very large cells, i.e.up to 100 km. So in the worse case, the round trip delay is equal to 0.66mswhich is even longer than a TTI. A simple solution to overcome this problemis to make sure the sub frame next to the RACH sub frame is not used for anytransmission. In this situation, the NodeB scheduler does not allocate the verynext sub frame to any synchronized UE. Consequently this sub frame acts as aguard interval.

5.3 Signal parameters

As mentioned in [3], the parameters for uplink transmission scheme within LTE(SC FDMA and OFDMA concepts) are specified in Table 5.3.

The FFT size is adapted to the given system bandwidth. In this way, thesub-carrier spacing is kept constant in any system bandwidth, which was part ofthe LTE requirements. f is always equal to 15 kHz. Also the time duration ofan OFDM symbol is also constant regardless of the system bandwidth: 66.67s.


System 1,25 MHz 2,5 MHz 5 MHz 10 MHz 15 MHz 20 MHzBW

Subframe 0,5 msduration

Subcarrier 15 kHzspacing

Sampling 1,92 MHz 3,84 MHz 7,68 MHz 15,36 MHz 23,04 MHz 30,72 MHz

frequency ( 12 3, 84 (2 3, 84 (4 3, 84 (6 3, 84 (8 3, 84

MHz) MHz) MHz) MHz) MHz)

FFT size 128 256 512 1024 1536 2048Number of

occupied 76 151 301 601 901 1201subcarriers

Table 5.1: System parameters for uplink transmission

Meanwhile the transmission bandwidth for a specified signal can be varied bychanging the number of used sub carriers. A simple illustration of a basictransmission is given in Figure 5.7.

S/P

P/S

map IFFT1024FFTNc CP

D/A15,36 MHz

1 OFDM symbol

0,66 stf

Nc x 15 kHzbaseband

signal

15,36 MHz0

CP

Figure 5.7: Illustration of a basic transmission in system bandwidth of 10 MHz

The design of the random access has to take into consideration these pa-rameters. For instance, a RACH signature must be a finite number of OFDMsymbols. Also, a CP may be added at the beginning of each symbol. The useof the CP will be discussed in the section 6.3. In a typical urban environment,the order of the maximum delay spread is of 5 s. Therefore, if a CP prefix isused, about 77 (= 5 106 15.36 106) chip symbols will be copied from theend of the OFDM symbol and added at the beginning.

An efficient way to understand the concept is to describe a particular simplescenario:

The random access signal can last 400 s and an OFDM symbol is 66.67

5.4. Power ramping 39

System bandwidth 10 MHzTransmission bandwidth for random access 1,25 MHzUE-NodeB distance 10 kmGuard interval 100 sAvailable interval for random access 400 sInsertion of CP NoBurst composition Signature only, no message part

Table 5.2: Particular scenario for random access

s long. With no CP insertion and no message part, a signature composedof 6 OFDM symbols can be transmitted.

The transmission bandwidth allocated to a random access is 1.25 MHz.As specified in Table 5.3, the number of used sub carriers for each OFDMsymbol should be 76 (in this way, the actual bandwidth is 76*15 kHz =1.14 MHz which is to prevent from leakage).

5.4 Power ramping

In WCDMA, the random access is carried out in the same frequency bandand time slots as the uplink data transmission. This prevents from assigning re-sources for random access but leads to interference. Therefore, a power rampingmethod, as mentioned in section 4.1.3, is performed to control the interferencecaused by the UE. A maximum power transmission would cause too much inter-user interference and would result in a deterioration of data reception, which isto be highly avoided. As mentioned earlier, E-UTRA random access is madeorthogonal to data transmission. In this way, no special procedure is neces-sary to control interference. Power ramping could hence be avoided to allowfor faster detection. However, a trade off between detection delay and energyconsummation has to be found. Transmitting every random access with highpower would obviously result in shorter latency but the risk would be to waste aprecious amount of energy. Also, in case of narrow transmission bandwidth (i.e.1.25 MHz), power ramping could be a solution to prevent from being stuck in adeep fading dip. Moreover, power ramping is an efficient method to overcomea saturation due to high traffic load. Indeed, in case of high traffic load, manyUE are likely to use the same frequency band at the same time interval. Then,if every UE tries to access the network by sending a powerful signal, the SIR(Signal to Interference Ratio) will become very low and the detection rate willdecrease. The use of the power ramping method would permit a lower interfer-ence level among the unsynchronized users. Finally, power ramping could alsobe useful to provide different priority classes among users. For example, onecan imagine that an emergency call is allowed to perform power ramping whilea simple call is not. The differentiation could as well be made by allocatingdifferent step size values to different user categories.

As a proposal, the following solution is given:

1.25 MHz transmission bandwidth: power ramping is used to overcomedeep localized fading. The used step size is function of the priority class


of the UE.

Wider bandwidth: Power ramping is an option. The NodeB informs theUE by broadcast whether the method has to be performed or not.

5.5 Frequency hopping

Frequency hopping is a simple solution to avoid deep frequency fading. It con-sists in choosing randomly a different frequency band for each successive attempt(see Figure 5.8).

t

f

Figure 5.8: Frequency hopping principle

Obviously, this can only be performed if multiple RACH sub channels aredefined on separated frequency bands. For instance, in the narrowest givensystem bandwidth, i.e. 1.25 MHz, frequency hopping is not achievable.

In case of frequency selective channel, this method results in better detectionrate by providing frequency diversity in the random access procedure. Theimprovement offered by frequency hopping is mainly noticeable when the UE isnot moving rapidly and/or in case two attempts are separated by few TTIs, i.e.when the channel remains almost the same between two attempts. Dependingon the coherence time of the environment, the performance enhancement canbe very high or insignificant.

The time varying aspect of the environment is defined by the maximumDoppler frequency shift which is function of the velocity:

fD = fC vc

where c is the velocity of light.

The time correlation of the channel response H can be written as:

H(t) = E[H(f, t).H(f, t+t)] = 22J0(2pifDt)

where 22 is the average gain.

The coherence time corresponds to a correlation function equal to 0.5 and isgiven by:

TC =9

16pifD

5.6. Signature format 41

For instance, if the UE is moving with a velocity equal to 120 km/h, themaximum Doppler frequency shift is 222 Hz (typical system with a carrier fre-quency of 2 GHz) and the coherence time is then equal to 0.8 ms. Consequentlywe can consider in this case that frequency hopping is useless because the envi-ronment changes significantly every 2 TTIs. Meanwhile, with a velocity equalto 3 km/h (pedestrian case), the coherence time is in the order of 32 ms andthen frequency hopping can provide great improvement.

5.6 Signature format

The signatures are pseudo noise codes. The goal is to obtain good cross correla-tion properties in order to achieve high detection rate and low false alarm rate.Ideally, the autocorrelation and cross correlation of the codes should be:

i,i(n) = E[si(k)si (k + n)] ={

E for n = 00 otherwise (5.1)

i,j(n) = E[si(k)sj (k + n)] = 0 n (5.2)Also it must be easy to reproduce the signature at the receiver side. Hencea noise-like waveform is not under consideration. Several techniques exist toobtain pseudo random codes. A common method is to generate a maximumlength sequence (m-sequence) by using a Linear Feedback Shift Register (LFSR).For instance, the WiMAX signatures are generated this way. The sequences areimplemented based on the recursion formula given by:

s(n) = Li=1

gis(n i) (5.3)

where gi belongs to [0,1], and L corresponds to the memory length of theregister.

The output is periodic with a period equal toN = 2L1. Hence, if the LFSRis designed to give a long enough m-sequence, the signatures can be chosen to bepieces of this sequence. In this way, no apparent relation exists between them.The bits are then modulated using QPSK (16 QAM could also be possible).

This remains a very simple method and obviously other techniques can beused to generate the pseudo random signatures.

5.7 Subcarriers mapping

Once a complex pseudo random code is generated, the next step consists inmapping the values to the sub-carriers. Here again, several possibilities couldbe adopted. In the DFTs OFDMA case, an FFT may be applied prior tothe mapping. Since the input is already pseudo random, this will not changethe performances significantly. However, this would permit to use the sametransmission chain as for data communication.

Regarding the mapping procedure, the 3 mentioned schemes (localized, ran-domized and equidistant) can be used. Let us co

10.1.1.63.7134.pdf

Documents

rach access procedure

rach procedure

uplink access

initial access

rach design

rach purpose335

multiple access protocols41

radio access technology