適用於多頻帶正交分頻多工系統之通道及多微網衍生干擾對抗技術

國

立 交 通 大 學

電信工程學系碩士班

碩士論文

適用於多頻帶正交分頻多工系統之通道

及多微網衍生干擾對抗技術

Channel and Multi-Piconet Induced Interference

Mitigation for Multi-band OFDM System

研 究 生：洪宗樺 Student:

Chung-Hua Hung

指導教授：李大嵩 博士 Advisor:

Dr.

Ta-Sung

Lee

適用於多頻帶正交分頻多工系統之通道

及多微網衍生干擾對抗技術

Channel and Multi-Piconet Induced Interference

Mitigation for Multi-band OFDM System

研 究 生：洪宗樺 Student:

Chung-Hua Hung

指導教授：李大嵩 博士 Advisor:

Dr. Ta-Sung Lee

國立交通大學

電信工程學系碩士班

碩士論文

A Thesis

Submitted to Institute of Communication Engineering

College of Electrical Engineering and Computer Science

National Chiao Tung University

in Partial Fulfillment of the Requirements

for the Degree of

Master of Science

in

Communication Engineering

June 2005

Hsinchu, Taiwan, Republic of China

適用於多頻帶正交分頻多工系統之通道

及多微網衍生干擾對抗技術

學生：洪宗樺

指導教授：李大嵩 博士

國立交通大學電信工程學系碩士班

摘要

由於超寬頻通訊系統具有取代家庭與辦公室內高傳輸速率之有線系統的能力 而成為極具潛能的新技術。超寬頻技術可提供每秒數億位元的傳輸速率而其涵蓋 的範圍可達數公尺。IEEE802.15.3a 多頻段正交分頻多工(Multiband OFDM, MB OFDM)系統是眾多超寬頻技術中較新的技術之一，該系統每次傳輸只利用 500 MHz 的頻寬並且將該頻寬分割成數個更小的副載波，然而 MB OFDM 之系統效能 會受到長延遲擴散通道及多微網間干擾的影響而大幅下降。在本論文中，吾人提 出一種具有對抗因長延遲擴散通道產生的載波間干擾(intercarrier interference)及避 免其它微網(piconet)干擾之接收機，雖然最小均方差等化器可以改善此 ICI 問題， 然而其複雜度過高不利於被MB OFDM 所採用，因此吾人提出「載波間干擾消除 輔助決策法」有效改善此一問題。另一方面，為了降低多微網間之干擾， MB OFDM 採用四組時頻交錯碼，吾人針對此問題提出「避免干擾傳輸法」，此方法略過被干 擾的頻槽(frequency slot)，而只在不受干擾的頻槽傳輸信號，藉以達到避免干擾之 目的。吾人藉由電腦模擬驗證上述架構在超寬頻環境中可有效改善位元錯誤率。

Channel and Multi-Piconet Induced Interference

Mitigation for Multi-band OFDM System

Student:

Chung-Hua

Hung

Advisor:

Dr.

Ta-Sung

Lee

Institute of Communication Engineering

National Chiao Tung University

Abstract

Ultra wideband (UWB) communication is a potential new technique for replacing high-speed data cables in homes and offices. UWB technology will be capable of transmitting hundreds of megabits per second over distances of several meters. Multi-band (MB) orthogonal frequency division multiplexing (OFDM) system is one of the most innovative UWB techniques. MB OFDM system utilizes only 500 MHz instantaneous bandwidth and dividing that frequency band into smaller simultaneously transmitted subcarriers. In this thesis, a MB OFDM receiver is proposed, which mitigates the intercarrier interference (ICI) induced by a long delay spread channel and avoids the interference from other piconet, respectively. Performance of a MB OFDM system is typically significantly degraded due to long delay spread channels. Although the minimum mean-square error (MMSE) equalizer can solve this problem, the matrix inverse computation is too complex for the MB OFDM system. To alleviate this problem, the decision-aided ICI canceller is designed to improve the influence of ICI. On the other hand, the four unique time-frequency interleaving codes are adopted to reduce the effect of cochannel interference (CCI) from simultaneously operating piconets (SOP). Furthermore, we proposed an interference avoidance transmission scheme which turns off the collided frequency slots and conveys signal in the interference-free frequency slots. Finally, we evaluate the performance of the proposed system and confirm that it works well in a UWB environment.

Acknowledgement

I would like to express my deepest gratitude to my advisor, Dr. Ta-Sung Lee, for his enthusiastic guidance and great patience. I learn a lot from his positive attitude in many areas. Heartfelt thanks are also offered to all members in the Communication Signal Processing (CSP) Lab for their constant encouragement. Finally, I would like to show my sincere thanks to my parents and friends for their inspiration and love.

Contents

Chinese Abstract

I

English Abstract

II

Acknowledgement III

Contents IV

List of Figures

VII

List of Tables

X

Acronym Glossary

XI

Notations XIII

1 Introduction

1

2 Overview of IEEE 802.15.3a Multi-band OFDM System

5

2.1 Review of OFDM ... 5

2.2 IEEE 802.15.3a Multi-band OFDM System... 10

2.2.1 PHY Frame Structure...11

2.2.2 Transmitter Architecture ... 13

2.3 Summary...22

3 Intercarrier Interference (ICI) Compensation in IEEE 802.15.3a

Multi-band OFDM System

34

3.1 Indoor UWB Channel Model...34

3.2 Receiver Architecture ...39

3.2.1 Synchronization ...39

3.2.2 Channel Estimation...44

3.3 Zero Padded Prefix (ZPP) OFDM System ...44

3.3.1 ZPP OFDM System Model for Long Delay Spread Channel...47

3.4 ICI Compensation for IEEE 802.15.3a MB OFDM System ...49

3.4.1 ICI Compensation Using MRC...50

3.4.2 ICI Compensation Using MMSE Equalizer ...51

3.4.3 ICI Compensation Using Decision-Aided ICI Canceller ...52

3.5 Channel Estimation Using Decision-Aided ICI Canceller ...53

3.6 Computer Simulations ...55

3.7 Summary ...57

4 Interference Avoidance Transmission Scheme for IEEE 802.15.3a

Multi-band OFDM 72

4.1 Review of Multiple Access Techniques...72

4.2 Multiple Access in Multi-band OFDM System ...76

4.2.1 Time Division Multiple Access (TDMA) for Intra-Piconet Interference Reduction ...77

4.2.2 Time-Frequency Interleaving Codes for Inter-Piconet Interference Reduction ...77

4.3 Simultaneously Operating Piconets (SOP) ...79

4.3.1 Collision Characteristics ...79

4.3.2 Interference Avoidance Transmission Scheme for SOP ...79

4.4 Computer Simulations ...83

5 Conclusion

91

List of Figures

Figure 1.1 UWB spectral mask for indoor communication systems. Emission level

is measured in 1 MHz bandwidth ...4

Figure 2.1 OFDM signal with cyclic prefix extension...23

Figure 2.2 A digital implementation of appending cyclic prefix into the OFDM signal in the transmitter ...23

Figure 2.3 Black diagrams of the OFDM transceiver. ...24

Figure 2.4 PLCP frame format of the MB OFDM system...24

Figure 2.5 Standard PLCP preamble format of the MB OFDM system ...25

Figure 2.6 Shortened PLCP preamble format of the MB OFDM system ...25

Figure 2.7 PHY header bit assignment of the MB OFDM system...26

Figure 2.8 The transmitter architecture of MB OFDM system...26

Figure 2.9 Scrambler/descrambler schematic diagram in MB OFDM system. ...27

Figure 2.10 Convolutional encoder: rate R = 1/3, constraint length K = 7...27

Figure 2.11 An example of the bit-stealing and bit-insertion procedure (R=1/2)...28

Figure 2.12 An example of the bit-stealing and bit-insertion procedure (R=5/8)...28

Figure 2.13 An example of the bit-stealing and bit-insertion procedure (R=3/4)...29

Figure 2.14 Guard subcarrier creation based on edge subcarriers of the MB OFDM symbol...29

Figure 2.15 Frequency of operation for the MB OFDM system...30

Figure 3.1 Simulation of passband system in terms of equivalent complex baseband system ...59

Figure 3.2 100 impulse responses based on the CM3 channel model (NLOS

up to 10 m with average RMS delay spread of 15 ns)...59

Figure 3.3 Average power decay profile for the channel model CM3 (NLOS up to 10 m with average RMS delay spread of 15 ns). ...60

Figure 3.4 Block diagram of the MB OFDM receiver...60

Figure 3.5 Block diagram of the cross-correlation packet detection...61

Figure 3.6 Block diagram of the symbol timing estimation...61

Figure 3.7 Block diagram of the frequency synchronization estimation. ...62

Figure 3.8 PSD plots for the OFDM system using CP prefix ...62

Figure 3.9 PSD plots for the OFDM system using ZPP...63

Figure 3.10 The ZPP OFDM system pick up larger noise at receiver ...63

Figure 3.11 Illustration of ISI and ICI due to long delay path ...64

Figure 3.12 Illustration of ISI (right) and ICI (left) channel matrix shapes...64

Figure 3.13 Magnitude of ICI matrix in frequency domain...65

Figure 3.14 Block diagram of the decision-aided ICI canceller ...65

Figure 3.15 Coded BER as a function of Eb/N0 for 53.3 Mbps data rate of the MB OFDM system in CM1-4 channels with parameters estimation...66

Figure 3.16 Coded BER as a function of Eb/N0 for 106.7 Mbps data rate of the MB OFDM system in CM1-4 channels with parameters estimation...66

Figure 3.17 Coded BER as a function of Eb/N0 for 200 Mbps data rate of the MB OFDM system in CM1-4 channels with parameters estimation...67

Figure 3.18 Coded BER as a function of Eb/N0 for 480 Mbps data rate of the MB OFDM system in CM1-4 channels with parameters estimation...67

Figure 3.19 Captured multipath energy as a function of ZPP length for CM1-4 channels ...68

Figure 3.20 Uncoded BER versus Eb/N0 with different iteration number in the CM4 channel ...68

Figure 3.21 Coded BER as a function of Eb/N0 for 480 Mbps data rate of the MB OFDM system in the CM4 channel ...69

Figure 3.22 Mean square estimation error of the CM4 channel frequency response….. ...69

Figure 3.23 Coded BER versus Eb/N0 for 480 Mbps data rate of the MB OFDM

system in the CM4 scenario with estimated channel impulse response….. ...70 Figure 4.1 The 802.15.3 piconet elements ...85 Figure 4.2 The 802.15.3 piconet superframe ...85 Figure 4.3 Pictorial representation of bandwidth expansion for the MB OFDM

system ...86 Figure 4.4 Collision property of two time–frequency interleaving codes for two

piconets ...86 Figure 4.5 Illustration of time-frequency slots in each frequency slot...87 Figure 4.6 Flow chart of the interference avoidance transmission scheme ...87 Figure 4.7 The ideal collision situation for no coordination of transmissions among two piconets ...88 Figure 4.8 Example for the 106.7Mbps mode of the MB OFDM system...88 Figure 4.9 Example of transmission scheme for the 106.7Mbps data rate mode of the MB OFDM system...89 Figure 4.10 Coded BER versus Eb/N0 for the 106.7 Mbps data rate mode of the MB

OFDM system in the CM 1 channel with different number time- frequency slots (SIR = 0 dB) ...89 Figure 4.11 Coded BER versus Eb/N0 for the 106.7 Mbps data rate mode of the MB

OFDM system in the CM 2 channel with different number time- frequency slots (SIR = 0 dB) ...90 Figure 4.12 Coded BER versus Eb/N0 for the 106.7 Mbps data rate mode of the MB

List of Tables

Table 2.1 Rate-dependent parameters of PHY header for the MB OFDM system ... ..31 Table 2.2 The data rate dependent modulation parameters of the MB OFDM

system ... 31 Table 2.3 Scrambler seed selection of PHY header for the MB OFDM system ... 32 Table 2.4 Modulation-dependent normalization factor KMOD for OFDM symbols.

... 32 Table 2.5 QPSK encoding table for OFDM symbols ... 32 Table 2.6 Time frequency interleaving codes and associated preamble patterns for the MB OFDM system... 33

Acronym Glossary

ADC analog-to-digital conversion CAP contention access period

CCI cochannel interference CDMA code division multiple access

CP cyclic prefix

CSMA/CA carrier sense multiple access with collision avoidance CTAP channel time allocation period

CTAs channel time allocations DAIC decision-aided ICI canceller

DFT discrete Fourier transform DS-UWB Direct-sequence UWB

FCC Federal Communications Commission FCS frame check sequence

FDMA frequency division multiple access FFT Fast Fourier Transform FIR finite impulse response

FSK frequency shift keying GPS global positioning system HCS header check sequence

ICI intercarrier interference ISI intersymbol interference IFFT Inverse Fast Fourier Transform LNA low noise amplifier

LOS line of sight

LS least square

LSB least significant bit

MAC Medium Access Control MB Multi-band

MCM multicarrier modulation

MCTAs management CTAs

MMSE minimum mean-square error MPCs multipath components (MPCs) NLOS non-LOS

OFDM orthogonal frequency division multiplexing PAPR peak-to-average power ratio

PHY Physical Layer

PLCP physical layer convergence procedure

PNC piconet coordinator

PSD power spectral density QoS quality of service

QPSK quadrature phase shift keying

RF radio frequency

SIR signal-to-interference ratio SNR signal-to-noise ratio

SOP simultaneously operating piconets S-V Saleh-Valenzuela

TDMA time division multiple access TFIC time-frequency interleaving code TSF Time Spreading Factor

UWB Ultra wideband

WLANs wireless local area networks WPANs wireless personal-area networks

Notations

NB number of bands

c

N number of subcarriers cp

N number of guard interval samples

DT

N number of data tones

sig

P _{power of desired signal}

int

P power of interference

R information data rate

Tc threshold of correlation

T_l delay of the lth cluster

s

T symbol duration

W effective bandwidth of transmitted signal

M modulation order X the lognormal shadowing

α multipath gain coefficient

,

k l

β fading associated with the kth _{ray of the l}th _cluster

Λ cluster arrival rate λ ray arrival rate

τ_κ delay of the kth multipath component

ρ correlation coefficient

2

n

σ noise power l

Chapter 1

Introduction

Since the Federal Communications Commission (FCC) approved the regulation for the commercial of Ultra Wideband (UWB) in February 2002, UWB has become a popular technology from commercial or civilian application and the development of UWB technology is drastically gaining momentum recently.

According to FCC definition, any signals that occupy more than 500 MHz bandwidth or have a fractional bandwidth of more than 20% are called UWB. Due to wide bandwidth the signals will be ultra-short waveforms in time domain. As such pulses with ultra-short duration have UWB spectral occupancy, UWB radios possess with some advantages for the radar and communications communities [1]:

1. Strengthened ability to pierce through obstacles

2. Supporting high precision ranging at the few centimeters

3. Providing very high data rates along with a increase in user capacity 4. Potentially small size and processing power

Currently, FCC has allocated 7500 MHz of spectrum for unlicensed use of UWB from 3.1 to 10.6 GHz frequency band and regulates power levels of UWB are very low, which allows UWB technology to overlay already available services such as the global positioning system (GPS) and the IEEE 802.11 wireless local area networks (WLANs) that coexist in the 3.6−10.1 GHz band. According to the

spectrum mask in Figure 1.1 [2], the power spectral density (PSD) measured in 1 MHz bandwidth must not exceed the specified -41.25 dBm [3]. It is a serious challenge to any UWB system about the restriction on PSD because other systems operated in the same band on licensed or unlicensed bands will have a much higher transmitted power than UWB system.

Although there are many challenges to UWB system, UWB is still emerging as a solution for the IEEE 802.15.3a standard [4] due to the properties of the UWB, like low power and high data-rate. The purpose of this standard is to provide a specification for a low complexity, low-cost, low power consumption, and high data-rate wireless connectivity among devices within or entering the personal operating space. The data rate must be high enough (greater than 110 Mbps) to satisfy a set of consumer multimedia industry needs for wireless personal-area networks (WPANs) communications. The standard also addresses the quality of service (QoS) capabilities required to support multimedia data types. Products compliant with this standard are envisioned to complement, not compete with, products compliant with IEEE 802.11 [5] which is a standard for local-area networks (LANs). The difference is similar to the differences between the Ethernet [6] LAN standard and the USB [7] or Firewire [8] standards that provide for connectivity to peripheral devices.

Currently, UWB systems are divided into main two groups: Direct-sequence UWB (DS-UWB) [9]-[11] and Multi-band (MB) OFDM [12]-[14]. In the MB OFDM, the spectrum is divided into several bands and transmits OFDM symbols in all bands. Additionally, in DS-UWB, the spectrum is divided into two sub-bands and the signals using a pseudorandom sequence for the spreading of information bits are transmitted in two sub-bands, respectively.

This thesis is organized as follows. In Chapter 2, we introduce the transmitter architecture of IEEE 802.15.3a MB OFDM system. In Chapter 3, the channel model of UWB and the receiver architecture of the IEEE 802.15.3a MB OFDM system are described. In addition, we introduce the equalization schemes to compensate intercarrier interference (ICI). In Chapter 4, the interference avoidance transmission scheme combining cross-correlator is proposed. In Chapter 5, we conclude this thesis and propose some potential future works.

Figure 1.1: UWB spectral mask for indoor communication systems. Emission level is measured in 1 MHz bandwidth.

Chapter 2

Overview of IEEE 802.15.3a

Multi-band OFDM System

MB OFDM system is one of the most innovative techniques. MB OFDM system involves utilizing only 500 MHz instantaneous bandwidth and dividing that frequency band into smaller simultaneously transmitted subcarriers. Such systems present high regulatory flexibility for universal operation because they enable independent control of portions of the emitted spectrum to adapt for different environments.

2.1 Review of OFDM

OFDM is a special case of multicarrier transmission, where a single data stream is transmitted over a number of low data rate subcarriers. OFDM can be thought of as a hybrid of multicarrier modulation (MCM) and frequency shift keying (FSK) modulation scheme. The principle of MCM is to transmit data by dividing the data stream into several parallel data streams and modulate each of these data streams onto individual subcarriers. FSK modulation is a technique whereby data is transmitted on one subcarrier from a set of orthogonal subcarriers in symbol duration. Orthogonality between these subcarriers is achieved by separating these subcarriers

by an integer multiples of the inverse of symbol duration of the parallel data streams. With the OFDM technique used, all orthogonal subcarriers are transmitted simultaneously. In other words, the entire allocated channel is occupied through the aggregated sum of the narrow orthogonal subbands.

The main reason to use OFDM systems is to increase the robustness against frequency-selective fading or narrowband interference. In a single carrier system, a single fade or interference can cause the entire link fail, but in a multicarrier system, only a small amount of subcarriers will be affected. Then the error correction coding techniques can be used to correct errors. The equivalent complex baseband OFDM signal can be expressed as

1 2 ₁ 2 2 2 ( ) 0 ( ) ( ) ( ) 0 Otherwise c c c c N N k k N _{k k T} k N k d t t T x t φ d φ t u t − − =− =−     _{≤ ≤}  _{ } =_ =    _{ } 

∑

_∑

_(2.1)

where Nc is the number of subcarriers, T is the symbol duration, dk is the transmitted

subsymbol (M-PSK or M-QAM), ( ) j2 f tk /

k t e π T

φ = is the kth subcarrier with the frequency /f_k =k T , and uT(t) is the time windowing function. Using the

correlator-based OFDM demodulator, the output of the jth branch can be presented as 1 2 ₂ * 0 0 2 1 ( ) ( ) c c N k j T T j t T j j k N k j y x t t dt d e dt T d π φ − ₋ =− = = =

∑

∫

∫

_(2.2)

By sampling x(t) with the sampling period Td=T/Nc, the discrete time signal xn

can be expressed as 1 2 2 2 1 0 1 ( ) IFFT{ } 0 Otherwise c c c d N k j n N k c N n t nT c k k d e n N x x t N d π − = =−   _{≤ ≤ −}  = =_ =  

∑

(2.3)

Note that xn is the Inverse Fast Fourier Transform (IFFT) output of the N input

data subsymbols. Similarly, the output of the jth branch can also be presented in the digital form 1 1 _{2 2} 0 2 1 FFT{ } [ ] c c c c N j N _{j n} N j n k k j N n c k y x x e x k j d N π δ − − ₋ = ₌₋ = =

∑

=

∑

− = (2.4)

In theory, the orthogonality of subcarriers in OFDM systems can be maintained and individual subcarriers can be completely separated by the Fast Fourier Transform (FFT) at the receiver when there are no intersymbol interference (ISI) and intercarrier interference (ICI) introduced by transmission channel distortions. However, it is impossible to obtain these conditions in practice. In order to eliminate ISI completely, a guard interval is imposed into each OFDM symbol. The guard interval is chosen larger than the expected delay spread, such that the multipath from one symbol cannot interfere with the next symbol. The guard interval can consist of no signals at all. However, the effect of ICI would arise in that case due to the loss of orthogonality between subcarriers. To eliminate ICI, the OFDM symbol is cyclically extended in the guard interval to introduce cyclic prefix (CP) as shown in Figures 2.1 and 2.2. This ensures that delayed replicas of the OFDM symbol always have an integer number of cycles within the FFT interval, as long as the delay is smaller than the guard interval. As a result, the delayed multipath signals which are smaller than the guard interval will not cause ICI. The complete OFDM signal with CP is given by 1 2 2 ( ) 2 1 0 1 0 Otherwise  c cp c c N k j n N N k c cp N n _{c k} d e n N N x N π − − =−   _{≤ ≤ + −}  =   

∑

(2.5)

where Ncp is the number of samples in CP. Due to CP, the transmitted OFDM symbol

symbols with the channel impulse responses will become a circular convolution one. Assuming the value of Ncp is larger than the channel length, the received data vector

can be expressed as = + y Hx η 0 1 0 1 0 0 1 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 c cp cp c c cp cp N N N N N N N h x h h h y x h h x h h y x h η η − − − − −                 _{   }    _{=   +}    _{  }                      _ _     y η x H # # # % # % # # # #   # % #   "   (2.6)

Applying SVD on the channel response, we have H

=

H UΣV (2.7) where U and V are unitary matrices, and Σ is a diagonal matrix. Substituting

Equation 2.7 and the equalities of _x=V and _{X Y=}UH_{y into Equation 2.6, the} received data vector can be written as

N ( ) H H H H = = + = + = + U η Y U y U Hx η U HVX N ΣX N (2.8)

This means that the output Y can be expressed in terms of the product of Σ and X plus noise. When x_−i =x_{N i−} for i=1,...,N_cp, a more compact matrix form of the guard interval can be written as

0 1 1 0 1 0 0 0 0 1 1 0 1 1 1 0 0 0 0 0 0 0 0 0 0 cp cp cp c cp c c cp cp N N N N N N N N N h h h h h h h y x h h y _{h h h} x h h h η η − − − −           _{ }     _{= }  ₊    _{ }      _{ }                    " " % " % # # % % % % # % " # # # % " # # % % % % % " " (2.9)

transform (DFT) matrix with the lth entry as 2 1 j l_c N l c e N π − = Q (2.10)

As in Equation 2.8, the received data y can be transformed into Y N ( ) H H H H H = = + = + = + N Σ Y Q y Q Hx η Q HQ X Q η ΣX N   (2.11)

According to Equation 2.11, by adding CP to the OFDM symbol, the modulation in OFDM is equivalent to multiplying the frequency domain signals of the OFDM symbol with the channel’s frequency response Σ .

The block diagrams of the OFDM transceiver is shown in Figure 2.3, where the upper path is the transmitter chain and lower path corresponds to the receiver chain. In the center, IFFT modulates a block of input values onto a number of subcarriers. In the receiver, the subcarriers are demodulated by the FFT, which performs the reverse operation of the IFFT. In fact, the IFFT can be made using the FFT by conjugating input and output of the FFT and dividing the output by the FFT size. This makes it possible to use the same hardware for both transmitter and receiver. This complexity saving is only possible when the transceiver doesn’t have to transmit and receive simultaneously. The functions before the IFFT can be discussed as follows. Binary input data is first encoded by a forward error correction code. The encoded data is then interleaved and mapped onto QAM values. In the receiver path, after passing the radio frequency (RF) part and the analog-to-digital conversion (ADC), the digital signal processing starts with a training sequence to determine symbol timing and frequency offset. The FFT is used to demodulate all subcarriers. The FFT outputs are mapped onto binary values and decoded to produce binary output data. In order to successfully map the QAM values onto binary values, the reference phases and amplitudes of all subcarriers have to be acquired first.

In conclusion, OFDM is a powerful modulation technique that simplifies the removal of distortion due to the multipath channel and increases bandwidth efficiency. The key advantages of OFDM transmission scheme can be summarized as follows:

1. OFDM is an efficient way to deal with multipath. For a given delay spread, the implementation complexity is significantly lower than that of a single carrier system with an equalizer.

2. In relatively slow time-varying channels, it is possible to significantly enhance the capacity by adapting the data rate per subcarrier according to the signal-to-noise ratio (SNR) of that particular subcarrier.

3. OFDM is robust against narrowband interference because such interference affects only a small amount of subcarriers.

4. OFDM makes single-frequency networks possible, which is especially attractive for broadcasting applications.

2.2 IEEE 802.15.3a Multi-band OFDM

System

The MB OFDM system provides WPANs with data payload communication capabilities of 53.3, 80, 106.7, 160, 200, 320, and 480 Mbps [12]. The support of transmitting and receiving at data rates of 53.3, 106.7, and 200 Mbps are mandatory. The system uses a total of 122 sub-carriers that are modulated using quadrature phase shift keying (QPSK). Forward error correction coding (convolutional coding) is used with a coding rate of 1/3, 1/2, 5/8, and 3/4. The MB OFDM system utilizes a time-frequency interleaving code (TFIC) to interleave coded data over 3 frequency bands (called a band group). Four such band groups with 3 bands each and one band

group with 2 bands are defined. There are also four 3-band TFICs and two 2-band TFICs, which, when combined with the appropriate band groups provide the capability to define eighteen separate logical channels or independent piconets.

2.2.1 PHY Frame Structure

As shown in Figure 2.4, a complete PLCP frame format defined in the MB OFDM standard consists of the PLCP preamble, PLCP header, MAC frame body (frame payload plus FCS), Tail bits and Pad bits. In the Specification, the PLCP preamble includes two kinds of OFDM training signals, which are used for synchronization, carrier-offset recovery, and channel estimation in the receiver, respectively. The PLCP header of the PLCP frame is composed of several fields, and the information conveyed in these fields is processed in the receiver to aid the demodulation from the frame payload.

PLCP Preamble Field

There are two kinds of PLCP preamble: standard PLCP preamble and shortened PLCP preamble. The main function of the standard PLCP preamble field is for receiver synchronization, carrier-offset, and channel estimation. The structure of the standard preamble defined in the MB OFDM is shown in Figure 2.5. The standard preamble includes three distinct portions: packet synchronization sequence, frame synchronization sequence, and channel estimation sequence. The packet synchronization sequence is composed of successively 21 repetitions of a time-domain sequence. Each period of the timing synchronization sequence is constructed by appending 37 “zero samples” to the sequences. This portion of the standard preamble is used for packet detection, acquisition, coarse carrier frequency estimation, and coarse symbol timing. The frame synchronization sequence of the

preamble is composed of successively 3 repetitions of a time-domain sequence similar to the packet synchronization sequence. This portion of the standard preamble is used to synchronize the receiver algorithm within the preamble. Finally, the channel estimation sequence of the standard preamble is composed of successively 6 repetitions of the OFDM training symbol. This training symbol is generated by passing the frequency-domain sequence though the IFFT, and appending the output with 37 “zero samples” to the resulting time-domain output. This portion of the standard preamble is used to estimation the channel frequency response, fine carrier frequency estimation, and fine symbol timing.

For data rates of 200 Mbps and lower, all the packets use the standard PLCP preamble. However, for the data rates higher than 200 Mbps, the first packet use the standard PLCP preamble, while the remaining packets use the shortened PLCP preamble. The structure of the shortened preamble is shown in Figure 2.6.

PLCP Header

As show in Figure 2.4, The PLCP header composed of PHY header and MAC header is always sent at an information data rate of 53.3 Mbps. The bit assignment of PHY header is illustrated in Figure 2.7. The RATE field occupies the Bits 2-6 of the PHY header, and conveys the information about the data rate chosen to transmit the following MAC frame body. The mapping between the data rate and the content of the RATE field is shown in Table 2.1. By the information from the RATE field, the receiver can determine what coding rate of the FEC coding is used in MAC frame body. Bits 9-20 of the PHY header are the LENGTH field, and the information conveyed in this field indicates the number of octets in MAC frame body. The number of octets in the LENGTH field shall be an unsigned 12-bit integer and has the range from 1 to 4095, and the least significant bit (LSB) shall be transmitted first

in time. Bit 23-24 shall encode the initial state of the scrambler, which is used to synchronize the descrambler of the receiver. For the other bits of PHY header, these are reversed for future use. The Tail bits are added after the PHY header in order to return the convolutional encoder to the “zero state”. Because the information conveyed in the RATE, LENGTH and SCRAMBLER fields are required for coding MAC frame body, it is required that these fields can be decoded immediately after the reception of the Tail field. Pad bits are added to the end of the Tail bits in order to align the data stream in the OFDM symbol interleaver boundaries. The remainder of the PLCP frame is sent at the desired information data rate depended on modulation parameter listed in Table 2.2.

2.2.2 Transmitter Architecture

As show in Figure 2.9, the transmitter architecture of MB OFDM is similar to that of a conventional OFDM system. In the following section we will introduce the transmitter architecture of MB OFDM in detail [13][14].

Data Scrambler

This is a method of removing long runs of 0’s or 1’s in the signal which would otherwise give the receiver a problem at the other end of the channel. This is achieved at the bit level. The MAC header, HCS, and MAC frame body are scrambled according to the scrambler seed identifier. However, the PLCP preamble, PLCP header, and tail bits shall not be scrambled. Because the initialization vector is determined from the seed identifier contained in the PLCP header of the received frame. The polynomial generator, g(D), for the pseudo random binary sequence (PRBS) generator shall be g(D) = 1 + D14 + D15, where D is a single bit delay

15 14 − − ⊕ = _{n n} n x x x (2.12)

where “⊕” denotes modulo-2 addition. This is realized, as shown in Figure 2.9, with a shift register. The following sequence defines the initialization sequence, xinit,

which is specified by the parameter “seed value” in Table 2.3. ]

[ _ni ₁ i_{n 2 n}i ₁₄ i_{n 15}

init x x x x

x = _{− −} " _{− −} (2.13)

where x_ni_−krepresents the binary initial value at the output of the kth delay element.

The scrambled data bits, sn, are obtained as follows:

n n

n b x

s = ⊕ (2.14)

where bn represents the unscrambled data bits. The de-scrambler at the receiver shall

be initialized with the same initialization vector, xinit, used in the transmitter

scrambler. The 15-bit seed value shall correspond to the seed identifier as shown in Table 2.3.

Convolutional Encoder

Convolutional coding operates at the bit level rather than block level such as RS coding. This has the advantage of the generator not having to store a whole block of data in expensive memory prior to performing the coding. The input stream is fed into shift register stages which have intermediate output taps after each stage. The input stream and various output taps are modulo two added. The MB OFDM system actually uses 6 shift register stages and the rate R = 1/3 with the generator polynomials are shown as,

G0 = 133Oct, G1 = 145Oct and G2 = 175Oct

The architecture ,as shown in Figure 2.10, can be considered a state machine. Since there are 6 stages in the MB OFDM implementation, this gives rise to 64 states. It is the change from one state to another based on which we can draw the trellis diagram to be applied in the Viterbi decoding algorithm.

The various coding rates are derived from the rate R = 1/3 convolutional code

by employing “puncturing”. Puncturing is a procedure for deleting the selected encoded bits in the transmitter (thus the number of transmitted bits are reduced and the coding rate is increacsed) and inserting a dummy “zero” metric into the convolutional decoder on the receive side according to the omitted bits pattern. The puncturing patterns are illustrated in Figures 2.11- 2.13. In each of these cases, the tables shall be filled in with encoder output bits from the left to the right. For the last block of bits, the process shall be stopped at the point at which encoder output bits are exhausted, and the puncturing pattern applied to the partially filled block. The PHY header, tail bits, MAC header, HCS, tail and pad bits shall be coded with a rate

R = 1/3. The encoder shall be reset to the all-zero state following this. Next, the

MAC frame body and tail bits shall be coded with a rate R = 1/3, 1/2, 5/8, or 3/4, corresponding to the desired data rate.

Bit interleaving

Bit interleaver is also performed to basically spread out the errors and so make the convolutional coding more effective. Bit interleaving provides robustness against burst errors. The bit interleaving operation is performed in three stages: (i) symbol interleaving across the OFDM symbols, (ii) intra-symbol tone interleaving, and (iii) intra-symbol cyclic shifts. The symbol interleaver permutes the bits across OFDM symbols to exploit frequency diversity across the sub-bands, while the tone interleaver permutes the bits across the data tones within an OFDM symbol to exploit frequency diversity across tones and provide robustness against narrow-band interferers. The length of the symbol interleaver is determined by the Time Spreading Factor (TSF) defined in Table 2.2. The symbol interleaver shall interleave among (6/TSF)*NCBPS coded bits, where NCBPS is the number of coded bits per OFDM

symbol. Following this, the symbols shall each be cyclically shifted by a different amount as described further in this section. This is done to exploit frequency diversity, especially in the modes that do not employ time spreading.

For the bit interleaving operation, the coded bits shall first be grouped together into blocks of (6/TSF)*NCBPS coded bits. Each group of coded bits shall then be

permuted using a block interleaver of size (6/TSF)*NCBPS. Let the sequences {U(i)}

and {S(i)}, where i = 0,…, (6/TSF)*NCBPS−1, represent the input and output bits of

the symbol block interleaver, respectively. The input-output relationship of this interleaver shall be given by:

(

)

( ) Floor 6 / * Mod( , _CBPS) CBPS i S i U TSF i N N      _      _{ }  = _{  }_+ _         (2.15)

where the function Floor(･) returns the largest integer value less than or equal to its argument value, and where the function Mod(･) returns the remainder after division of i by NCBPS.

The output of the symbol block interleaver is then passed through a tone block interleaver. The outputs of the symbol block interleaver are grouped together into blocks of NCBPS bits and then permuted using a regular block interleaver of size NTint

× 10, where NTint = NCBPS/10. Let the sequences {S(i)} and {T(i)}, where i = 0,…,

NCBPS−1, represent the input and output bits of the tone interleaver, respectively. The

input-output relationship of the tone block interleaver is given by:

int int ( ) Floor 10 * Mod( , _T ) T i T i S i N N      _      _{ }  = _{ } _+ _         (2.16)

where the function Mod(･) returns the remainder after division of i by NTint .

The output of the tone interleaver is then passed through the last stage, which consists of a different cyclic shift of each block of NCBPS bits within the span of the

0,1,…,NCBPS−1, represent the input and output sequences, respectively, of the cyclic

shift for the bth block. Then,

(

)

( , ) , Mod( ( ), _CBPS)

V b i =T b i+A b N (2.17)

For conjugate symmetric modes, NCBPS = 100 : A(b) = b*33, b = 0,1,2. For

non-conjugate symmetric modes with time spreading (TSF = 2), NCBPS = 200 : A(b) =

b*66, b = 0,1,2. For non-conjugate symmetric modes with no time spreading

(TSF=1), NCBPS=200: A(b) = b*33, b = 0,1,2,…,5.

Modulation

The OFDM subcarriers shall be modulated using QPSK. The conversion shall be performed according to the Gray-coded constellation mappings. The output values, d, are formed by multiplying the resulting (I + jQ) value by a normalization factor of

KMOD, as described in the following equation:

MOD

( )

d = I +jQ ×K

The normalization factor, KMOD, depends on the base modulation mode, as prescribed

in Table 2.4. For QPSK, b0 determines the I value, and b1 determines the Q value, as

illustrated in Table 2.5.

Frequency-domain Spreading

For information data rates of 53.3 and 80 Mbps, the stream of complex symbols is divided into groups of 50 complex numbers. We shall denote these complex numbers cn,k, which corresponds to subcarrier n of OFDM symbol k, as follows:

, 50* * ( 50), (49 ) 50* 0,1,..., 49, 0,1,..., 1 n k n k SYM n k n k c d n k N c d + + − + = = = − = (2.18)

where NSYM denotes the number of OFDM symbols in the MAC frame body, Tail bits,

For information data rates of 110, 160, 200, 320, 400 and 480 Mbps, the stream of complex numbers is divided into groups of 100 complex numbers. We shall denote these complex numbers cn,k, which corresponds to subcarrier n of OFDM symbol k,

as follows:

, 100* 0,1,..., 99, 0,1,..., 1

n k n k SYM

c =d ₊ n = k = N − (2.19)

where NSYM denotes the number of OFDM symbols in the MAC frame body, Tail bits,

and Pad bits.

An OFDM symbol rdata,k(t) is defined as

, , 0 2 MOD( ,127) 2 ( ) exp( 2 ( ) ( )) exp( 2 ( )) SD ST ST N data k n k F CP n N n F CP k n N r t c j M n t T p P j n t T π π = =− = ∆ − + ∆ − ∑ ∑ (2.20)

where NSD is the number of data subcarriers, NST is the number of total subcarriers,

and pMOD and Pn together describe the contribution of the pilot and guard subcarriers,

as further defined in Pilot Subcarriers section. The function M(n) defines a mapping from the indices 0 to 99 to the logical frequency offset indices –56 to 56, excluding the locations reserved for the pilot subcarriers, guard subcarriers, and the DC subcarrier, as shown below:

                      = − ≤ ≤ − ≤ ≤ − ≤ ≤ − ≤ ≤ − ≤ ≤ − ≤ ≤ − ≤ ≤ − ≤ ≤ − ≤ ≤ − ≤ ≤ − ≤ ≤ − ≤ ≤ − = − = 99 43 98 90 44 89 81 45 80 72 46 71 63 47 62 54 48 53 50 49 49 46 50 45 37 51 36 28 52 27 19 53 18 10 54 9 1 55 0 56 ) ( n n n n n n n n n n n n n n n n n n n n n n n n n n n n n M (2.21)

Pilot Subcarriers

In each OFDM symbol following the PLCP preamble, twelve of the subcarriers are dedicated to pilot signals in order to make coherent detection robust against frequency offsets and phase noise. These pilot signals shall be put in subcarriers numbered –55, –45, –35, –25, –15 –5, 5, 15, 25, 35, 45, and 55. The contribution due to the pilot subcarriers for the kth OFDM symbol is given by the inverse Fourier Transform of the sequence Pn,k below, which includes BPSK modulation by a

pseudorandom binary sequence, pl (defined further below), to prevent the generation

of spectral lines. MOD( ,127) 1 15,45 2 1 5,25,35,55 2 1..., 4,, 6,..., 14, 16,..., 24, 0 26,..., 34, 36,..., 44, 46,..., 54, 56 n _k j n j P p n n  +  ₌    − −  = _ =   _{= ± ± ± ± ± ±}   _{± ± ± ± ± ± ±}  (2.22)

For modes with data rates less than 106.7 Mbps:

55 , 45 , 35 , 25 , 15 , 5 , * , , =P− n=− − − − − − P_{nk nk} (2.23)

For 106.7 Mbps and all higher rate modes: 55 , 45 , 35 , 25 , 15 , 5 , , , =P− n=− − − − − − P_{nk nk} (2.24)

In this numbering, k = 0 shall correspond to the first OFDM symbol following the

PLCP preamble. The length 127 pseudo-random sequence, pl, which modulates the

pilot subcarriers is defined below:

p0…126 = {1, 1, 1, 1, -1, -1, -1, 1, -1, -1, -1, -1, 1, 1, -1, 1, -1, -1, 1, 1, -1, 1, 1, -1, 1, 1, 1, 1, 1, 1, -1, 1, 1, 1, -1, 1, 1, -1, -1, 1, 1, 1, -1, 1, -1, -1, -1, 1, -1, 1, -1, -1, 1, -1, -1, 1, 1, 1, 1, 1, -1, -1, 1, 1, -1, -1, 1, -1, 1, -1, 1, 1, -1, -1, -1, 1, 1, -1, -1, -1, -1, 1, -1, -1, 1, -1, 1, 1, 1, 1, -1, 1, -1, 1, -1, 1, -1, -1, -1, -1, -1, 1, -1, 1, 1, -1, 1, -1, 1, 1, 1, -1, -1, 1, -1, -1, -1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1}.

Guard Subcarriers

In each OFDM symbol ten of the subcarriers at the edges of the occupied frequency band shall be termed guard subcarriers. There are five guard subcarriers on either edge of the OFDM symbol occupied band, located in subcarriers with indices –61, –60, …, –57, and 57, 58, …, 61. These guard subcarriers shall be created by copying over the five outermost data-bearing subcarriers from the nearest edge of the OFDM symbol as shown in Figure 2.14 (note the intervening pilot subcarrier is not copied). The guard subcarrier symbol definition for the nth subcarrier of the kth symbol shall be given as follows:

, , , , , 0,1,2, 3, 4; 57 ; 95 , 0,1,2, 3, 4; 61 ; n k m k n k m k P c l n l m l P c l n l m l = = = + = + = = = − + = (2.25)

In this numbering, k = 0 shall correspond to the first OFDM symbol following the

PLCP preamble (i.e., the first OFDM symbol following the channel estimation symbols).

Time-domain Spreading

For data rates of 53.3, 80, 110, 160 and 200 Mbps a time-domain spreading operation shall be performed with a spreading factor TSF = 2, in order to improve frequency diversity and Simultaneously Operating Piconets (SOP) performance. The time-domain spreading shall consist of transmitting the same information over two OFDM symbols. The kth original OFDM symbol, represented as Sk(l), shall be

generated as specified in Modulation section. The repeated version of this OFDM symbol, represented asS l , shall be obtained in the time domain as follows: _k' ( )

( )

{

{

}

{

}

}

Mod(k+6,127)

'

Mod(k+6,127)

Im ( ) Re ( ) no conjugate symmetry

( ) with conjugate symmetry

k k k k S l j S l p S l S l p  ₊  =    (2.26)

where k = 0 shall correspond to the first OFDM symbol following the PLCP preamble, i.e., the first OFDM symbol following the channel estimation symbols, and the values of the index k are OFDM symbol numbers before time spreading. Also, the values for pk are selected from the same pseudo-random sequence used to

scramble the pilot subcarriers.

Operating Band Frequency

The MB OFDM system defines a unique numbering system for all channels that have a spacing of 528 MHz and lie within the band 3.1-10.6 GHz. The relationship between center frequency and band number is

Band center frequency = 2904+528×n_b,n_b =1...14(MHz) (2.27) as shown in Figure 2.15 and based on Equation 2.27, five band groups are defined,

consisting of four groups of three bands and one group of two bands as shown in Table 2.6.

2.3 Summary

The Specification of MB OFDM system has been introduced in this chapter. Although the transmitter architecture is similar to conventional OFDM system, some properties of channel are induced by the ultra wide bandwidth waveforms and some receiver function blocks should be modified. The following details will be discussed in Chapter 3.

Figure 2.1: OFDM signal with cyclic prefix extension.

Figure 2.2: A digital implementation of appending cyclic prefix into the OFDM signal in the transmitter.

Serial to parallel converter Input data Symbols IDFT 2 1 Nc d ₋ 2 2 Nc d ₋ 2 Nc d₋ 2 1 Nc d_{− +} .... .... ... .... . 0 x 1 x 1 c N x − Parallel to serial converter D/A OFDM signal .... .... . c cp N N x − .... .... . .... .... . X₀ X₁ X_N-1 X_N-1 X_N-Ncp Cyclic Prefix N_cp Useful Part N

Complete OFDM Signal

N+N_cp X₀ X₁ X_N-1 X_N-1 X_N-Ncp Cyclic Prefix N_cp Useful Part N

Complete OFDM Signal

Figure 2.3: Black diagrams of the OFDM transceiver. PLCP Preamble PHY Header MAC Header HCS Tail Bits Tail Bits Pad Bits Frame Payload Variable Length: 0 − 4095 bytes

Pad Bits Tail Bits FCS 53.3 Mb/s PLCP Header RATE

5 bits LENGTH12 bits Scrambler Init2 bits Reserved

2 bit Reserved2 bit Reserved2 bit Reserved3 bit

53.3, 55, 80, 106.7, 110, 160, 200, 320, 400, 480 Mb/s

Figure 2.4: PLCP frame format of the MB OFDM system.

Binary Input Data

Binary Output Data

Coding Interleaving QAM Mapping

Pilot Insertion

Serial to Parallel

Decoding _interleavingDe- _MappingQAM _CorrectionChannel Parallel to_Serial

IFFT (TX) FFT (RX) Parallel to Serial Serial to Parallel Add Cyclic Extension

and Windowing DAC

RF TX

Remove Cyclic Extension Timing and Frequency

Synchronization ADC

Packet Sync Sequence 21 OFDM symbols

Channel Est Sequence 6 OFDM symbols

9.375 µs

Frame Sync Sequence 3 OFDM symbols 0 ... 0 C₀ C₁ ... C₁₂₇ 0 0 0 0 0

PS₀ PS₁ PS₂₀ FS₀ FS₁ FS₂ CE₀ CE₁ CE₅

0 ... 0 −C₀ −C₁ ... −C₁₂₇ 0 0 0 0 0

Figure 2.5: Standard PLCP preamble format of the MB OFDM system

Packet Sync Sequence 9 OFDM symbols

Channel Est Sequence 6 OFDM symbols

5.625 µs

Frame Sync Sequence 3 OFDM symbols 0 ... 0 C₀ C₁ ... C₁₂₇ 0 0 0 0 0

PS₀ PS₁ PS₈ FS₀ FS₁ FS₂ CE₀ CE₁ CE₅

0 ... 0 −C₀ −C₁ ... −C₁₂₇ 0 0 0 0 0

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 LSB MSB S1S2 LENGTH (12 bits) SCRAMBLER (2 bits) R: Reserved Transmit Order (from left to right)

R R1 R2 R3 R4 RATE (5 bits) R 18 19 20 21 22 R 23 24 25 26 27 R R R R R5 R R

Figure 2.7: PHY header bit assignment of the MB OFDM system

Figure 2.8: The transmitter architecture of MB OFDM system

DAC Scrambler Convolutional_Encoder Puncturer _InterleaverBit Constellation_Mapping

IFFT Insert Pilots Add CP & GI Time-Frequency Code exp(j2πfct) Input Data

x

14

_x

13

_…

_x

2

_x

1

x

15

Data in

Scrambled

data out

Figure 2.9: Scrambler/descrambler schematic diagram in MB OFDM system

D D D D D D Input Data Output Data A Output Data B Output Data C

X₀ B₀ C₀ A₀ B₀ C₀ y₀ Stolen Bit

Inserted Dummy Bit A₀

Source Data

Encoded Data

Bit Stolen Data (sent/received data)

Bit Inserted Data

Decoded Data

A0 C0

Figure 2.11: An example of the bit-stealing and bit-insertion procedure (R=1/2)

X₄ X₃ X₂ X₁ X₀ A₄ A₃ A₂ A₁ B₄ B₃ B₂ B₁ B₀ C₄ C₃ C₂ C₁ C₀ A₄ A₃ A₂ A₁ A₀ B₄ B₃ B₂ B₁ B₀ C₄ C₃ C₂ C₁ C₀ y₄ y₃ y₂ y₁ y₀ Stolen Bit

Inserted Dummy Bit A₀

Source Data

Encoded Data

Bit Stolen Data (sent/received data)

Bit Inserted Data

Decoded Data

A0 C0 C1 A2 C2 A₃ A4 C4

X₂ X₁ X₀ A₂ A₁ B₂ B₁ B₀ C₂ C₁ C₀ A₂ A₁ A₀ B₂ B₁ B₀ C₂ C₁ C₀ y₂ y₁ y₀ Stolen Bit

Inserted Dummy Bit A₀

Source Data

Encoded Data

Bit Stolen Data (sent/received data)

Bit Inserted Data

Decoded Data

A₁ C₀ C₁ A₂

Figure 2.13: An example of the bit-stealing and bit-insertion procedure (R=3/4)

5 15 25 35 45 55 0 DC c50c53p5c54c62p15c63c71p25c72c80p35c81c89p45c90c98p55c99c95 c99 c4 c0p5c1 c9p15c10c18p25c19c27p35c28c36p45c37c45p55c46c49 c0 -55 -45 -35 -25 -15 -5 -61 61 P P P P copy copy

Figure 2.14: Guard subcarrier creation based on edge subcarriers of the MB OFDM symbol

Figure 2.15: Frequency of operation for the MB OFDM system f 3432 MHz 3960 MHz 4488 MHz 5016 MHz 5544 MHz 6072 MHz 6600 MHz 7128 MHz 7656 MHz 8184 MHz 8712 MHz 9240 MHz 9768 MHz Band #1 Band #2 Band #3 Band #4 Band #5 Band #6 Band #7 Band #8 Band #9 Band #10 Band #11 Band #12 Band #13

GROUP 1 GROUP 2 GROUP 3 GROUP 4

Band #14 10296 MHz GROUP 5 f 3432 MHz 3960 MHz 4488 MHz 5016 MHz 5544 MHz 6072 MHz 6600 MHz 7128 MHz 7656 MHz 8184 MHz 8712 MHz 9240 MHz 9768 MHz Band #1 Band #2 Band #3 Band #4 Band #5 Band #6 Band #7 Band #8 Band #9 Band #10 Band #11 Band #12 Band #13

GROUP 1 GROUP 2 GROUP 3 GROUP 4

Band #14 10296

MHz

Table 2.1: Rate-dependent parameters of PHY header for the MB OFDM system Rate (Mb/s) R1 – R5 53.3 00000 55 01000 80 00001 106.67 00010 110 01010 160 00011 200 00100 320 00101 400 00110 480 00111 Reserved 01001, 01011–11111

Table 2.2: The data rate dependent modulation parameters of the MB OFDM system

Data Rate (Mb/s) Modulation Coding rate (R) Conjugate Symmetric Input to IFFT Time Spreading Factor Overall Spreading Gain Coded bits per OFDM symbol (NCBPS) 53.3 QPSK 1/3 Yes 2 4 100 80 QPSK ½ Yes 2 4 100 106.7 QPSK 1/3 No 2 2 200 160 QPSK ½ No 2 2 200 200 QPSK 5/8 No 2 2 200 320 QPSK ½ No 1 1 200 400 QPSK 5/8 No 1 1 200 480 QPSK ¾ No 1 1 200

Table 2.3: Scrambler seed selection of PHY header for the MB OFDM system Seed identifier (b1, b0) Seed value (x14 … x0)

0,0 0011 1111 1111 111

0,1 0111 1111 1111 111

1,0 1011 1111 1111 111

1,1 1111 1111 1111 111

Table 2.4: Modulation-dependent normalization factor KMOD for OFDM symbols

Modulation KMOD

QPSK 1/ 2

Table 2.5: QPSK encoding table for OFDM symbols Input bit (b0 b1) I-out Q-out

00 -1 -1

01 -1 1

10 1 -1

Table 2.6: Time frequency interleaving codes and associated preamble patterns for the MB OFDM system

Chapter 3

Intercarrier Interference (ICI)

Compensation in IEEE 802.15.3a

Multi-band OFDM System

Due to its wide bandwidth, a new UWB channel model is developed by the IEEE 802.15.3a standard and is described in this chapter. For highly dispersive channels, some conventional OFDM receiver algorithms are not suitable anymore. In this chapter, the indoor UWB channel model and conventional synchronization techniques for the IEEE 802.15.3a MB OFDM system will be introduced first. Then, the zero padded prefix (ZPP) OFDM system will be introduced in the following section. In addition, the phenomenon and the equalization scheme for long delay spread channels will be described. Finally, the performance simulations are shown in Section 3.5.

3.1 Indoor UWB Channel Model

All wireless systems must be able to deal with the challenges of operating over a multipath propagation channel, where objects in the environment can cause

multiple reflections to arrive at the receiver. The different multipath components (MPCs) are characterized by different delays and attenuations. The correct modeling of the parameters describing the MPCs is the art of channel modeling [15][16].

For narrow-band systems, these reflections will not be resolvable by the receiver when the narrow-band system bandwidth is less than the coherence bandwidth of the channel. When there are a large number of arriving paths at the receiver within its resolution time, the central limit theorem is commonly invoked in order to model the received envelope as a Rayleigh random variable. However, in UWB systems, the large bandwidth of UWB waveforms significantly increases the ability of the RX to resolve the different reflections in the channel. This large bandwidth can give rise to two effects. First, the number of reflections arriving at the receiver within the period of a very short impulse becomes smaller as the duration of the impulse gets shorter and shorter, so the central limit theorem used to justify a Rayleigh distribution for the receiver signal envelop is no longer applicable. Second, the multipath components may be resolved on a very fine time scale (proportional to the inverse of the signal bandwidth), and the time of arrival of the multipath components may not be continuous. This phenomenon could explain the “clustering” of multipath components. For a realistic performance assessment, a UWB channel model like the 802.15.3a standard model has to include all those effects.

Three main indoor channel models were considered: the tap-delay line Rayleigh fading model [17], the Saleh-Valenzuela (S-V) model [18], and the ∆-K model described in [19]. These models use a statistical process to model the discrete arrivals of the multipath components. However, the S–V model is unique in its approach of modeling arrivals in clusters, as well as rays within a cluster. This extra degree of freedom yielded better matching of the model to the channel characteristics gathered from measurement data. As a result, the IEEE 802.15.3a standards body selected the

S–V model, which then needed to be properly parameterized in order to accurately reflect the unique characteristics of the measurements.

3.1.1 Saleh-Valenzuela Model

The S-V model models the multipath of an indoor environment for wideband channels. In order to capture this effect, The S-V model distinguishes between “cluster arrival rates” and “ray arrival rates,” where the first cluster starts by definition at time t = 0, and the rays within the cluster arrive with a rate, given by a Poisson process with a start time relative to the cluster arrival time.

Though, the original S-V model has the characteristic that the amplitude statistics sufficiently match the Rayleigh distribution, the power of which is controlled by the cluster and ray decay factors. However, in UWB channels the amplitudes do not follow a Rayleigh distribution. Rather, either a lognormal or Nakagami distribution can fit the data equally well, which has been verified using Kolmogorov-Smirnov testing with a 1 percent significance level. According to these results, the S-V model was modified for the IEEE model by prescribing a lognormal amplitude distribution. The model also includes a shadowing term to account for total received multipath energy variation that result from blockage of the line-of-sight path. The impulse response of multipath model is described as

( ) _,

(

_,

)

0 0 L K i i i i i k l l k l l k h t X α δ t T τ = = =

∑∑

− − (3.1) where i_, k l

α are the multipath gain coefficient, i l

T is the delay of the lth cluster, i_, k l

τ

is the delay of the kth multipath component relative to the lth cluster arrival time i l

T ,

i

X represents the lognormal shadowing, and i refers to the ith realization.

By definition, we have τ = . The distribution of cluster arrival time and the _0,l 0 ray arrival time are given by the independent interarrival exponential probability

density function

(

l | l 1

)

exp

(

l l 1

)

, 0 p T T₋ = Λ _−Λ T −T₋ _ l > (3.2) ( )

(

k l, | k 1 ,l

)

exp

(

k l, (k 1 ,)l

)

, 0 p τ τ ₋ =λ _−λ τ −τ ₋ _ k >   (3.3)

where Λ is the cluster arrival rate, and λ is the ray arrival rate, i.e., the arrival rate of a path within each cluster. The channel coefficients are defined as follows:

, , ,

k l pk l l k l

α = ξ β (3.4)

where ξ reflects the fading associated with the lth cluster and _l β corresponds to _{k l,}

the fading associated with the kth ray of the lth cluster, the small-scale amplitude statistics were modeled as a lognormal distribution rather than the Rayleigh distribution, whichwas used in the original S–V model, which is reflected in the following equations

(

)

(

2 2

)

( , 1 2)20 1 2 , , , 20 log10 Normal , or 10 k l n n l k l k l l k l µ ξ β _∝ µ σ ₊σ ξ β ₌ + + (3.5) where

(

2

)

1 Normal 0, 1 n ∝ σ and

(

2

)

2 Normal 0, 2

n ∝ σ are independent and correspond to the fading on each cluster and ray, respectively. σ is standard ₁

deviation of cluster lognormal fading term (dB). σ is standard deviation of ray ₂

lognormal fading term (dB).

The behavior of the averaged power delay profile is

, 2 _/ / 0 , l k l T l k l E _{ξ β}  = Ωe− Γe−τ γ     (3.6)

where Tl is the excess delay of bin l and Ω is the mean energy of the first path of ₀

the first cluster, and p_k,l is equiprobable ±1 to account for signal inversion due to reflections. The µk,l is given by

2 2 0 , 1 2 , 10 ln( ) 10 / 10 / ( )ln(10) ln(10) 20 l k l k l T τ γ σ σ µ = Ω − Γ − − + (3.7)

variable in order to capture shadowing effects in the channel. the lognormal shadowing of the total multipath energy is captured by the term, X , the total energy _i

contained in the terms i_, k l

α is normalized to unity for each realization. This shadowing term is characterized by the following

2

20 log10( )X_i ∝Normal(0,σ_x) (3.8)

Note that, a complex tap model was not adopted here. The complex baseband model is a natural fit for narrowband systems to capture channel behavior independently of carrier frequency, but this motivation does not work for UWB systems where a real-valued simulation at RF may be more natural. Figure 3.1 illustrates the equivalent model for simulation of passband system in terms of complex baseband system. Therefore the real-valued passband multipath channel response is simplified as follow

( )

(

)

1 P i i i h t α δ t τ = =

∑

− (3.9)

where α is the real-values channel coefficient. The equivalent baseband multipath _i

channel response is described by

( ) 2

(

)

i

(

)

1 c i P P j f i i i i i i h t _αe− π τ_δ t _{τ α δ} t _τ = =

∑

− =

∑

−  _(3.10)

The UWB model parameters were designed to fit measurement results, and Table 3.1 provides the results of this fit for four kinds different channel scenarios (LOS refers to line of sight, NLOS to non-LOS).

(1). CM1 is based on LOS (0 – 4 m) channel measurements. (2). CM2 is based on NLOS (0 – 4 m) channel measurements. (3). CM3 is based on NLOS (4 – 10 m) channel measurements.