CODE SHIFTED REFERENCE IMPULSE-BASED COOPERATIVE UWB COMMUNICATION SYSTEM

P a g e 1 CODE SHIFTED REFERENCE IMPULSE-BASED COOPERATIVE UWB COMMUNICATION SYSTEM Pir Meher Ali Shah Mohammed Abdul Rub Ashik Gurung This thesis is presented as part of Degree of Master of Science in Electrical Engineering Sweden September 2011 School of Engineering Department of Applied Signal Processing Supervisor: Muhammad Gufran Khan Examiner: Dr. Jörgen Nordberg

P a g e 2

P a g e 3 Abstract Ultra wideband (UWB) is a radio technology in which the transmission of information is done over a large bandwidth with very short pulses at low energy levels. UWB technology has gained a high popularity in the field of short-range wireless communications. UWB provides significant benefits like position location capability, reduced fading effects, and higher channel capacity. UWB technology is very desirable because of its certain characteristics like low power consumption, cost reliability and simple architecture. However, UWB systems face challenges regarding system design to achieve low complexity and low cost. UWB systems need high sampling frequencies and face problems while using digital signal processing technology. In this thesis, first, the comparison of transmitted reference (TR), multi-differential frequency shifted reference (MD-FSR) and code shifted reference (CSR) is done in terms of BER performance and system complexity. The simulation results validate that the CSR-UWB system has better BER performance and lower complexity than MD FSR-UWB system. Secondly, cooperative communication is implemented for CSR-UWB. The system BER performance of the CSR impulse-based cooperative UWB communication system is evaluated for different number of relays and different average distances between source node and destination node. The simulations are carried out under both line of sight (LOS) and non-line of sight (NLOS) environments. The simulation results show that the performance of the system decreases with the increase in average source-to-destination distance. We also observe that the system performs better under an environment of LOS channel than under an environment of NLOS channel. Finally, the results validate that the system performs better as the number of relay nodes increases until it reaches an adequately large number.

P a g e 4

P a g e 5 Acknowledgement First of all, we would like to express our warmest thankfulness to our supervisor Muhammad Gufran Khan. We are highly indebted towards him for giving us his valuable time, effort and guidance throughout our entire thesis period. We would really want to appreciate his clarity in his direction and high dedication as a key of our motivation towards our thesis. Without his help and support, this thesis would not have been possible. We would also like to convey our gratitude to our examiner, Dr. Jörgen Nordberg. Our thanks also goes to our teachers in. We would like to thank our classmates and friends for all their help and for making our life here enjoyable. We are really grateful to our parents, brothers and sisters for providing good education and good environment to us. Lastly, we would like to thank God for giving us the courage and strength to move ahead towards our destiny.

P a g e 6

P a g e 7 Table of Contents ABSTRACT....3 ACKNOWLEDGEMENT....5 CHAPTER ONE...9 INTRODUCTION....10 1.1 OVERVIEW 10 1.2 HISTORY AND BACKGROUND.10 1.3 DEFINITION OF UWB.11 1.4 FEATURES AND ADVANTAGES OF UWB..12 1.5 UWB CHALLENGES..14 1.6 THESIS CONTRIBUTION 15 CHAPTER TWO..16 UWB SYSTEMS, MODULATION AND MULTIPLEXING TECHNIQUES.17 2.1 UWB SYSTEM TYPES..17 2.1.1 MULTICARRIER UWB..17 2.1.2 IMPULSE RADIO UWB 17 2.2 UWB PULSE SHAPE.18 2.3 MODULATION TECHNIQUES USED IN UWB SYSTEMS..20 2.3.1 PULSE POSITION MODULATION.20 2.3.2 PULSE AMPLITUDE MODULATION 21 2.3.3 BINARY PHASE SHIFT KEYING..22 2.3.4 ON-OFF KEYING.22 2.4 MULTIPLE ACCESS TECHNIQUES USED IN UWB SYSTEM 23 2.4.1 TIME-HOPPING UWB.23 2.4.2 DIRECT SEQUENCE UWB.25 CHAPTER THREE..26 ULTRA WIDEBAND WIRELESS CHANNELS 27 3.1 PROPAGATION MECHANISMS AND CHANNEL CHARACTERISTICS..27 3.1.1 MULTIPATH..27 3.1.2 DELAY SPREAD 28 3.1.3 COHERENCE BANDWIDTH..29 3.2 CHANNEL FADING DISTRIBUTIONS. 29 3.2.1 GAUSSIAN CHANNEL.. 29 3.2.2 RAYLEIGH CHANNEL 30 3.2.3 RICEAN CHANNEL.30 3.3 ULTRA WIDEBAND CHANNELS 30 3.4 IEEE 802.15.4A UWB CHANNEL MODEL.. 31

P a g e 8 3.4.1 PATH LOSS.31 3.4.2 SHADOWING 32 3.4.3 POWER DELAY PROFILE 32 3.4.4 SALEH AND VALENZUELA 33 3.4.5 DELAY DISPERSION..35 3.4.6 SMALL SCALE FADING 36 CHAPTER FOUR 37 IR-UWB RECEIVERS..38 4.1 INTRODUCTION.38 4.2 TRANSMITTED REFERENCE (TR) UWB RECEIVER..38 4.3 FREQUENCY-SHIFTED REFERENCE (FSR) UWB RECEIVER 39 4.4 CODE-SHIFTED REFERENCE (CSR) UWB RECEIVER 41 4.5 BER PERFORMANCE COMPARISON OF THE TR-UWB, THE FSR-UWB AND THE CSR-UWB SYSTEMS..45 CHAPTER FIVE 48 COOPERATIVE UWB COMMUNICTION SYSTEM.49 5.1 INTRODUCTION.49 5.2 WHAT IS COOPERATIVE COMMUNICATION?.49 5.3 COOPERATIVE COMMUNICATION PROTOCOLS: PROCESSING MODES OF RELAYS... 51 5.3.1 DECODE-AND-FORWARD...51 5.3.2 AMPLIFY-AND-FORWARD..51 5.3.3 COMPRESS-AND-FORWARD. 52 5.3.4 ESTIMATE-AND-FORWARD...52 5.3.5 CODED COOPERATIONS..52 5.4 COOPERATIVE UWB SYSTEM MODEL 53 5.5 PERFORMANCE EVALUATION FOR RELAY POSITIONING...55 5.6 PERFORMANCE EVALUATION OF THE COOPERATIVE CSR-UWB SYSTEM UNDER DIFFERENT CHANNEL CONDITIONS...57 5.6.1 Case I: 4M (LOS) VS. 7M (LOS) WITH 5 RELAYS 58 5.6.2 Case II: 4M (LOS) AND 7M (LOS) VS. 4M (NLOS) WITH 5 RELAYS.....59 5.6.3 Case III: 4M (LOS), 7M (LOS) AND 4M (NLOS) VS. 7M (NLOS) WITH 5 RELAYS....59 5.6.4 Case IV: 4M (LOS) VS. 7M (LOS) WITH 10 RELAYS.60 5.6.5 Case V: 4M (LOS) AND 7M (LOS) VS. 4M (NLOS) WITH 10 RELAYS...61 5.6.6 Case VI: 4M (LOS), 7M (LOS) AND 4M (NLOS) VS. 7M (NLOS) WITH 10 RELAYS....62 CHAPTER SIX..64 CONCLUSIONS..65 FUTURE WORK.66 REFERENCES 67

P a g e 9 CHAPTER ONE Introduction

P a g e 10 Chapter 1: Introduction 1.1 Overview The tremendous growth in wireless technology and the huge demand in achieving successful deployment of wireless communication have significant impact on our daily lives. Since 1990, wireless communication has come to rise in the field of communication technology throughout the whole world because of its undeniable applications. The cellular communication growth from analog to digital, the development of third and fourth generation radio systems and the transition of wired connection to Wi-Fi are enabling customers to access a broad range of information at any time and from anywhere [1]. The need of every customer is faster service, higher capacity, and more confidential wireless connections. The Ultra Wideband (UWB) technology fulfills those demands by introducing exciting new features in radio communications. 1.2 History and Background Ultra wideband (UWB) differs from other communication techniques because it uses tremendously narrow radio pulses for the communication between transmitters and receivers. The utilization of short-duration pulses as building units for communication can produce a very large bandwidth and provide many advantages [1] like immense throughput, robustness, along with the co-existence of current radio features [1]. Ultra wideband communications was first introduced by Guglielmo Marconi in the early 19 th century by employing spark gap radio transmitters for spreading Morse code sequences over the Atlantic Ocean [2]. However, the advantage of a wide bandwidth and the potential of implementation of multiuser systems presented by electromagnetic pulses were not taken in account at that moment. After about fifty years later, modern pulse based communication was introduced in the form of radars. The technology was limited to military and defense departments for confidential purposes like extremely secure communications. Development in the field of micro-processing

P a g e 11 and semiconductor technology has lifted UWB for commercial uses [1]. The demand for UWB raised and developers of UWB system approached the Federal Communication Commission (FCC) for an approval for commercial implementation. In 2002, FCC granted an approval to UWB for commercial purpose [1]. 1.3 Definition of UWB Ultra wideband is a radio signal that can be employed with a very low energy in a high bandwidth. FCC suggests a UWB system has a bandwidth that is larger than 500 MHz or it has a fractional bandwidth more than 20% of the center frequency [3]. Fractional bandwidth is defined as [3] f f f H L BW = (1) fc Here, f BW is the fractional bandwidth, f H is the highest cutoff frequency (at -10 db emission point) and f L is the lowest cutoff frequency (at -10 db emission point). f c is the center frequency that can be calculated as f c = (f H + f L )/2. According to FCC, the UWB range for unlicensed frequency is 3.1 GHz to 10.6 GHz for both outdoor and indoor environments [5]. Figure 1.1 shows the comparison between ultra wideband communication system and narrowband. Narrowband Power Spectral density UWB -10 db f L fc f H Frequency (Hz) Figure 1.1: Comparison between ultra wideband communication system and narrowband [4]

P a g e 12 1.4 Features and advantages of UWB The main advantages of UWB are as follows: Channel capacity improvement As mentioned earlier, UWB utilizes a very large frequency spectrum which improves the channel capacity C (bits per seconds) or data rate. An increase in radio frequency (RF) bandwidth also increases the capacity of band limited additive white Gaussian noise (AWGN) channel [6]. Prec C = BRF log 2 (1+ ), (2) B where C represents the capacity of the channel with its radio frequency (RF) bandwidth B RF. P rec is the received power signal and o is the noise of power spectral density (PSD). RF o Ability to work with low SNR The above channel capacity equation also denotes that it logarithmically depends on the signal to noise ratio. Hence, the UWB system is able to work in rough communication channels with low signal to noise ratio and provides better channel capacity which is the outcome of large bandwidth [1]. Accurate positioning and tracking or radar sensing Larger the bandwidth, finer is the resolution. One of the key features of UWB technology is to provide accurate positioning. This application is mostly used in radar sensing to detect the targets [7]. High Performance in Multipath Channels Multipath is an unavoidable phenomena in wireless communication channels. Multipath reflection of the transmitted signal can be caused when the transmitted signal gets reflected from several surfaces like trees, buildings, people, etc as shown in figure 1.2. Multipath fading is the variation in the attenuation of the signal caused when the signal reaches the destination through multiple paths [1]. The effect of multipath fading is low in UWB communication

P a g e 13 system because UWB has short duration pulses, thus the effect of reflected pulses on the signal is less degrading [1]. Figure 1.2: Multipath phenomenon in wireless transmission Simple Transceiver Architecture The transmission in UWB communication system is carrier-less. That means the data need not be modulated over the continuous waveform with any particular carrier frequency. Carrierless transmission needs smaller number of radio frequency components compared to carrierbased transmission [1]. Mixers and oscillators are not required in UWB transceiver to translate carrier frequency to required frequency band [1]. Because of these reasons, UWB transceiver system architecture has lower complexity compared to other narrowband transceivers and is less expensive to design.

P a g e 14 Resistance to jamming Processing gain (PG) refers to the resistance of a radio system against jamming. It is identified as the ratio of RF bandwidth to information bandwidth of the signal. RF Bandwidth PG= (3) Information Bandwidth UWB spectrum accommodates a wide range of frequencies and provides a high processing gain which in turn gives the UWB signals high resistance to jamming. 1.5 UWB Challenges There are many challenges faced in the UWB technology by using very short pulses for communications. Here we discuss only the important challenges observed in UWB communication system [1]. Channel estimation The estimation of channel performance is very sensitive issue for designing a receiver in a wireless communication system. Measuring the exact performance of each channel is not possible in the field of wireless communication. To estimate channel parameters, it is essential to employ training sequences such as delays and attenuations of the propagation path. Mostly UWB receiver associate received signal with predefined signal model, which is not possible without any prior knowledge of the wireless channel parameter. But, because of the large bandwidth and lowered signal energy, UWB pulses face harsh distortion which makes channel estimation very difficult [8]. Multiple Access Interference In a multiuser environment, multiple users send information independently and simultaneously through a shared transmission medium. On the receiving side, more than one receiver should be set to separate users and receive information from the particular users. The interference of multiple users leads to multiple-access interference (MAI). Increase in MAI

P a g e 15 tends to unavoidable noise that considerably degrades the UWB pulses and creates complications in detection [1]. Pulse Shape Distortion UWB pulses can be distorted considerably by the transmission path. According to Friis transmission formula, received signal power decreases when frequency is increased [1]. Because of the long range of frequencies of UWB spectrum, the received power immensely changes and distorts the shape of the UWB pulse. This degrades the performance of the UWB receivers [1]. Synchronization of High Frequency The synchronization of time is a very important challenge in UWB systems. But, as the UWB pulses are tremendously short, flawless synchronization is hard to achieve. In such a case, major issues can arise due to poor detection of the exact position of the received signal [1]. 1.6 Thesis Contribution This thesis presents introduction, back ground, features and challenges of UWB technology in chapter 1. Chapter 2 introduces types of UWB system such as Multicarrier UWB (MC-UWB), Impulse Radio UWB (IR-UWB) and description of various modulation techniques Chapter 3 gives brief outline about the different UWB channels and its mechanism as well as equations for path loss. In addition, IEEE 802.15.4a channel model for low data rate UWB systems is studied. Chapter 4 presents the structure and implementation of different IR-UWB receivers such as Transmitted Reference Shifting (TR) UWB, Frequency Shifted Reference (FSR) UWB and Code Shifted Reference (CSR) UWB and the simulation results. In chapter 5 we have discussed about Cooperative UWB communication system. We have compared and analyzed computer simulation results of cooperative CSR-UWB systems under different channels.

P a g e 16 CHAPTER TWO UWB Systems, Modulation and Multiplexing Techniques

P a g e 17 Chapter 2: UWB Systems, Modulation and Multiplexing Techniques 2.1 UWB system types UWB systems can be typically divided into two categories: one based on sending multiple simultaneous carriers named as Multicarrier UWB and the other based on sending very short duration pulses with relatively low energy called Impulse radio UWB [9]. These two systems are discussed in the following sections. 2.1.1 Multicarrier UWB (MC-UWB) The idea of using multiple carriers in sending UWB signal is to divide the channel bandwidth into a number of small sub-channels with adequately small bandwidth for the efficient utilization of the bandwidth of the system [10]. For multi-carrier transmission technique, OFDM (orthogonal frequency division multiplexing) is employed which allows the sub-carriers to overlap in frequency without interfering with each other that results an increase in spectral efficiency [9]. Such a system is called Multi-band Orthogonal Frequency Division Multiplexing (MB-OFDM) [11]. In such a technique, a spectrum is divided into further smaller subspectrums with a minimum bandwidth of 500 MHz each. The data is then interleaved on these small sub-spectrums and transmitted into the air using multi-carrier OFDM technique [11]. Thus, by using this system, multiple users can also be supported by the allocation a group of sub-channel to each user [9]. A high throughput can be obtained by a reliable communication system by the transmission of multiple streams of data in parallel on separate carrier frequencies [9]. 2.1.2 Impulse Radio UWB (IR-UWB) In impulse radio UWB systems, transmission is based on a series of discontinuous short pulses or a pulse wave form which is generally known as monocycle pulses that have relatively low

P a g e 18 energy level [11]. The monocycle waveform could be any function that fulfills the requirements of spectral mask regulations. Such common pulses comprise of Rayleigh, Laplacian, Gaussian or Hermitean pulses [12]. A Gaussian monocycle waveform is employed along with Binary Phase Shift Keying (BPSK) as a data modulation scheme in this thesis. In such systems, because of the short length of the pulses which is nearly in nanoseconds, the bandwidth of the transmitted signal is in gigahertz [9]. Such pulses have ultra-wide frequency domain features which do not need any carrier modulation for propagation in the radio channel [12]. This approach is usually employed for a single user, but it can also be employed on multiple users by using the techniques of time-hopping or direct sequence spreading [12]. For the purpose of attaining a proper processing gain which can be employed to handle noise and different interferences from the environment, a single symbol which has to be transmitted is stretched over number of monocycle pulses [12]. This processing gain can be expressed as PG = log ( ) (1) 1 10 10 With the help of pseudorandom (PR) time-hopping code, consecutive pulses are transmitted in air interface in a discontinuous time-hopped scheme which provides UWB communication resistance against severe multipath propagation [12]. The short pulse length and relatively lengthy pulse repetition time helps in reducing the inter-pulse interference. This allows the multipath components related to the transmitted pulse to be attenuated prior to sending the next pulse [12]. The inter symbol interference (ISI) can be avoided between the pulses by increasing the time in between the pulses so that it becomes greater than the channel delay spread [12]. 2.2 UWB Pulse Shape Usually, the pulse shapes implemented in UWB communications consist of Gaussian pulse, Gaussian monocycle and Gaussian doublet as shown in figure 2.1. A Gaussian pulse [12] is expressed as 1 (1/ 2)(( t µ ) / σ ) P Re( t) = e (2) 2πσ 2

P a g e 19 where σ is the length of the pulse and µ represents the centre of the pulse. The Gaussian pulse is practically not applicable in wireless communication systems because it consists of a DC term. But, the higher derivatives of the Gaussian pulse are free from such type of DC terms, and hence can be practically implemented in wireless communication systems. The first derivative of Gaussian pulse is regarded as Gaussian monocycle. It is commonly employed in impulse radio systems. A Gaussian monocycle in time domain can be expressed as 2 1 t µ 2 σ 2 1 t µ P G ( t) = 1 e (3) 2πσ σ For Gaussian monocycle, µ = 3.5σ and the effective time length T p = 7σ. The second derivative of Gaussian pulse is regarded as the Gaussian doublet. It contains two Gaussian pulses which are opposite in terms of amplitude. A Gaussian doublet in time domain can be expressed as 2 2 1 t µ 1 t µ T w 1 = 2 σ 2 σ P ( ) GD t e e (4) 2πσ where T w is the time gap between the maxima of each pulse. The effective time length is T = 14σ at T = 7σ. p w 1 Gaussian pulse Gaussian monocycle Gaussian doublet 0.5 Amplitude 0-0.5-1 -2-1 0 1 2 Time [s] X 10-9 Figure 2.1: Waveforms for Gaussian pulse, Gaussian monocycle and Gaussian doublet [12]

P a g e 20 2.3 Modulation Techniques Used in UWB Systems The process of changing the characteristics of periodic waveform with an external signal by changing its amplitude, phase, or frequency is known as modulation. In UWB communication system, different modulation techniques are employed. The most commonly used modulation schemes are pulse-position modulation (PPM), pulse amplitude modulation (PAM), On-Off Keying (OOK) and binary phase-shift keying (BPSK). These modulation schemes are discussed below. 2.3.1 Pulse Position Modulation (PPM) In Pulse Position Modulation (PPM), the selected bit that is to be transmitted controls the position of UWB pulse. PPM is concerned with the nominal pulse position. In PPM, two or more positions in time encode the information, which is shown in the figure 2.2 [12] [13]. Those pulses which are transmitted at nominal position are represented as 0 and the ones that are transmitted beyond the nominal position are represented by 1 [13]. In figure 2.2, 2-ary PPM modulation is shown in which one bit is encoded in every impulse [12] [13]. Adding more positions allows more bits per symbol. In general, the time delay in between the position of pulses is a fraction of a nanosecond, which is much shorter than the one in between the nominal positions. This aids in avoiding the interferences among the impulses [13]. For PPM signals, the signal model is generally expressed as + k= 0 s ( t) = p( t kt f ± T pk ) (5) where p(t) represents the UWB pulse and little shifts T pk in pulse position performs the data modulation [13].

P a g e 21 1 0 1 0 1 t Figure 2.2: 2-ary PPM signal [14] 2.3.2 Pulse Amplitude Modulation (PAM) Pulse Amplitude Modulation (PAM) is concerned with the transmission of data in a time sequence of electromagnetic pulses by changing the power amplitudes or the voltage of individual pulses. It can also be defined as a technique in which the data to be transmitted is encoded in the amplitude of a series of signal pulses. The 2-ary PAM signal is illustrated in figure 2.3 in which the pulse with higher amplitude is represented by 1 and the one with lower amplitude is represented by 0 [14]. The M-ary PAM signal with different amplitude levels of M that consists of sequences of modulated pulses is expressed as m kt f k= s ( t) = a ( k) p( t ) (6) where a m (k) is the amplitude of the k th pulse that depends on the M-ary information symbol m {0,1,,M-1}, T f is the frame interval and T p is the pulse duration [14]. 1 0 1 0 1 t Figure 2.3: 2-ary PAM Signal [14]

P a g e 22 2.2.3 Binary Phase Shift Keying (BPSK) In Binary Phase Shift Keying (BPSK) which is also known as bi-phase modulation scheme, the binary data is carried in the polarity of the pulses [14]. A positive polarity is given to the pulse that represents the information bit 1 and a negative polarity is given to the pulse representing the information bit 0, which is demonstrated in figure 2.4. BPSK is the simplest version of Phase Shift Keying (PSK) [14]. The Binary Phase Shift keying can be expressed as k= s ( t) = d( k) p( t kt f ) (7) where 1 if information bit is 1 d (k) = (8) 1 if information bit is 0 1 0 1 0 1 t Figure 2.4 BPSK Signal [14] 2.3.4 On-Off keying (OOK) On-Off keying is also occasionally known as non-return-to-zero (NRZ) encoding. It is a binary level modulation scheme that contains two symbols with equal probabilities [15]. It is a special case of PAM, where m belongs to {0,1} with pulse amplitude as a m (k) = m(k) [14].

P a g e 23 In OOK, a pulse or signal is transmitted only when the information bit is equal to 1 as shown in figure 2.5. No pulse or signal is transmitted when the information bit is equal to 0. On-Off keying is mathematically expressed as: k= s ( t) = m( k) p( t kt f ) (9) where m(k) is the pulse amplitude and T f is the frame time [14]. 1 0 1 0 1 t Figure 2.5 OOK Signal [14] 2.4 Multiple Access Techniques used in UWB System In single band UWB systems, multiple users share a single UWB spectrum simultaneously. For accommodating those multiple users, a suitable multiple access technique is required [16]. There are two common multiple access schemes: Time hopping (TH) and Direct Sequence (DS) spreading, which are used to allow the users in a single band UWB system [16]. The difference between the two systems is that the TH technique is concerned with the randomization of the location of the transmitted UWB impulse in time, whereas the DS technique is concerned with the continuous transmission of pulses comprising a single data bit [17]. The TH-UWB and DS-UWB are explained in detail in the subsections below. 2.4.1 Time-Hopping UWB (TH-UWB) In TH-based system, the information verified by the TH sequence is transferred with a time offset for each pulse [16]. TH-UWB makes use of low duty cycle pulses, where users are time

P a g e 24 multiplexed by spreading the time spreading in between the pulses [16]. Every frame duration is split into multiple smaller segments of which only one carries the transmitted monocycle of the user [16]. The user is assigned a unique code known as TH sequence to designate the segment employed for transmission in every frame interval [16]. As the position of every impulse is verified by a pseudo-random (PR) code, extra energy is added to the symbol because of which the range of the transmission is increased [17]. In such a way, the identification of different users is done by their unique TH-code that allows them to be transmitted at the same time [17]. TH-UWB for the j th users for different modulation schemes of UWB can be expressed as follows [17] [18] For PAM modulation: For PPM modulation: s s ( j) ( j) 1 s ( t) = p( t kt lt c T ) d (10) k= l= 0 1 s k= l= 0 s f ( j) l ( j) l c ( j) k ( j) k ( t) = p( t kt lt c T d δ ) (11) s In these equations d k is the k-th data bit of j th user, s is the number of impulses transmitted for every information symbol, T s is the total symbol transmission time that is divided into s frames each of duration T f and each frame is itself subdivided into slots of duration T c [17]. The PR TH code sequence c l (unique for the j-th user) determines the position of one impulse in each frame to be encoded as shown in the figure 2.6 [17]. Because of the blank transmission in case of 0 th bit, OOK cannot be employed in TH spreading [17]. f c T s t T c T p T f Figure 2.6: TH-UWB Signals [17]

P a g e 25 2.4.2 Direct Sequence UWB In Direct Sequence UWB system, data is carried by multiple pulses whose amplitudes are based on a certain spreading code [16]. DS-UWB uses a train of high-duty-cycle pulses whose polarities are based on pseudo-random code sequences [16]. Each user in the system is specifically assigned a pseudo-random sequence that regulates pseudorandom inversions of the UWB pulse train [16]. This sequence of UWB pulses uses a data bit to modulate. This results in the transmission of continuous UWB pulses whose number depends on the length of the pulses itself and a system defined bit rate [17]. DS-UWB scheme is only applicable in PAM, OOK and PSM modulation. It is not suitable for PPM as it is a time-hopping technique [17] [16]. For PAM and OOK modulation, DS-UWB can be expressed as following [17][18], s ( j) 1 s ( t) = p( t kt lt ) c d ) (13) k= l= 0 s where d k is the k-th data bit, c l is the l-th chip of the PR code, p(t) is the pulse waveform of duration T p, T c is the chip length which is equal to T p as shown in figure 2.7, s is the number of pulses per data and j is the user index [17]. The PR sequence has values in {-1,+1} and length of the bit is T s = s T c [17]. c ( j) l ( j) k T s t T c = T p Figure 2.7: DS-UWB Signals [17]

P a g e 26 CHAPTER THREE Ultra Wideband Wireless Channels

P a g e 27 Chapter 3: Ultra Wideband Wireless Channels 3.1 Propagation Mechanism and Channel Characteristics The design and analysis of UWB communication systems are based on the propagation features of UWB radio channels [18]. Reflection, diffraction and scattering are the main propagation mechanism in communication [24]. Reflection occurs when a propagating signal impinges on an object which has comparatively large dimensions than the propagating signal. Reflection may occur at surfaces of the floor, walls and buildings [24]. Diffraction occurs when the radio path between the transmitter and the receiver is obstructed by objects with sharp edges. This results in the bending of signal around the obstacle, even when the line of sight exists between transmitter and receiver. Scattering occurs when the signal passes through a medium which contains objects that have very small dimensions compared to the wavelength, and when the number of obstacles per unit volume is quite large [24]. In this chapter, we will discuss about UWB wireless channels. The natures of these UWB wireless channels are very important and helpful while designing UWB communication system to predict the coverage of the signal, to reach maximum data rate, to find optimal location of antennas and for efficient modulation [27]. 3.1.1 Multipath The purpose of any communication system is to convey the message from transmitter to receiver. A transmitted signal in wireless communication takes multiple paths to reach the receiver which causes multipath effects because of reflections from objects like mountains, buildings, water bodies, etc. Multipath effect includes constructive and destructive interference at the receiver antenna and phase shifting of these multipath components of the signal causes multipath fading [27]. Fading is a common problem that occurs in a propagating wireless signal. Any fluctuation in the received signal is referred to as fading. If fading occurs due to multipath then it is referred as multipath induced fading [19] [20]. Fading is classified into two types: slow fading and fast fading. Slow fading arises when the coherence time of the channel is

P a g e 28 relatively larger than the delay constraints of the channel and fast fading arises when the coherence time of the channel is relatively small than the delay constraint of the channel [21]. The fast fading transmitter takes the advantage of both channel variation and channel conditions by the use of time diversity that helps to increase the strength of the communication signal. Whereas in case of slow fading, time diversity cannot be used as an advantage due to single realization of the channel within its delay constraint [22]. 3.1.2 Delay Spread When a signal transmits via a time-dispersive multipath channel, the signal arrives to the receiver from different paths. This is the cause of delay spread. Delay spread depends on the distance and the position of objects near the transmission path. Delay spread can be interpreted as the difference between the arrival time of the first and last multipath components [23] [19]. Delay spread leads to inter symbol interference (ISI). ISI is a form of distortion in the communication channels. In practice, communication channels have limited bandwidth, hence the transmitted pulses spread during transmission. This spreading of pulses causes overlap over the adjacent time slot that causes errors at the receiver. This phenomena is referred to as inter symbol interference which is shown in figure 3.1. Figure 3.1: Inter Symbol Interfernce (ISI) in digital transmission

P a g e 29 3.1.3 Coherence Bandwidth Coherence Bandwidth is a range of frequencies that are allowed to pass through the wireless channel without any distortion. It can be regarded as a statistical measurement of the frequency ranges on the channel which can be assumed as flat [24]. Alternately, coherence bandwidth can be suggested as an approximate highest bandwidth at which two frequencies of the signal will possibly go through similar or correlated amplitude fading [25]. Multipath interference can be avoided by decreasing the signal bandwidth so that it is less than the coherence bandwidth [19]. BW C = 1 2πT (1) d 3.2 Channel Fading Distributions Channel fading distribution refers to the factors or conditions that distort the signal when it transmits from source to destination through the channel. The performance of the channels plays a very vital role in transmission [19]. 3.2.1 Gaussian Channel The Gaussian channel is particularly used in modeling the noise produced at the receiver when the transmission path is ideal [19]. This model is moderately correct in few cases like wired communications transmissions and space communications. This channel model is appropriate for channels with single transmitter and single receiver. A condition when the information is sent through a channel that can be subjected to an additive white Gaussian noise is demonstrated in figure 3.2. X i Channel Encoder + Channel Decoder Y i =X i +Z i Continuous values Noise (Z i ) Figure 3.2: Gaussian channel model

P a g e 30 Here, Y i is the output of the channel, X i is the input of the channel and Z i is the noise which is assumed to be independent of X. Z i is zero mean Gaussian with variance : Z i N (0, ). 3.2.2 Rayleigh Channel Rayleigh channel is a transmission channel which has a fading envelope in the form of Rayleigh probability density function (pdf). Rayleigh fading occurs in an environment where there are several obstacles that scatter the transmitted signal before it reaches the receiver [26]. Rayleigh fading distribution is generally used to describe the statistical time. It varies in the nature of the received envelope of flat fading. The envelope is a sum of two quadrature Gaussian noise signal which follows the Rayleigh distribution [24]. 3.2.3 Ricean Channel Ricean channel is a transmission channel that has a line of sight (LOS) propagation path along with a small scale fading envelope distribution [24]. In this channel, the signal reaches the receiver at different angles or paths that result in multipath interference. Ricean fading takes place when one LOS path signal is stronger than the others. The strong signal arriving with several weak multipath signals results Ricean distribution [24]. Complex signals resemble noise signals that have enveloped in Rayleigh channel. Ricean distribution degenerates to Rayleigh distribution when the dominant signal fades away [26]. 3.3 Ultra Wideband Channels As mentioned in chapter 1, the UWB systems provide a promising technological application across several commercial fields and military applications including radar, communication, and medical instruments. This technology offers very high data rates to several users during short range communication channels by allocating a large bandwidth. UWB channels demonstrate two significant effects: pulse distortion and multipath propagation. Particularly, in the course of propagation, each waveform can be discarded by any object that results in multipath propagation. Pulse distortion is rather concerned with the variation in the original UWB pulse. The main difference between narrow band channel and UWB channel is different radiation

P a g e 31 bandwidth ranges. The narrowband is used to cover less than 20 MHz of bandwidth, where as UWB channels can cover more than 10 GHz of bandwidth [27] [29]. The standard UWB channel models are designed by IEEE 802.15 groups: SG3a and SG4a which are known as IEEE 802.15.3a and IEEE 802.15.4a channel models, respectively. 3.4 IEEE 802.15.4a UWB Channel Model IEEE 802.15.4a UWB channel model provides a broad range of channel environments such as industrial, residential, office and outdoor. It covers a frequency range from 2 GHz to 10 GHz. [28]. This model suggests data rates from 1Kbps to several Mbps. The important features of IEEE 80215.4a are: 3.4.1 Path Loss Path loss is the ratio between transmit and receive signal power. In an end-to-end wireless communication system, a transmitter communicates with a receiver by sending a signal over the wireless medium. The signal strength attenuates when it travels through the medium. Thus it becomes poorer or weaker as the propagation distance increases. The signal beyond a certain distance becomes unacceptable. Then at a regular interval to reactivate the signal strength, we need a booster or repeater. More challenging problems will occur when there are multiple receivers in the communication and more complexity will arise when distance from transmitter to receiver is varying [20]. Basically, path loss can be defined as: P t P L = (2) Pr where P t is the transmitted power and P r is the received power. Path loss in narrow band can be defined as: E{ PRX ( d, fc)} PL( d) = (3) PTX where PTX is the transmit power and PRX is the receive power, d is the distance between the transmitter and receiver, fc is the center frequency and E{} is expectation to averaging shadowing and small scale fading [30].

P a g e 32 3.4.2 Shadowing Shadowing is the effect that arises upon the received signal power when it is attenuated because of the obstacles in the propagation path in between the transmitter and the receiver. The nature of shadowing in UWB communication is similar to that of narrowband systems [29] [30]. The average path loss evaluated over a small scale fading in db is given as [30]: d PL ( d) = PL + n + S d 0 10 log 10 (4) 0 where, S denotes the shadowing Gaussian noise distributed random variable with zero mean and the standard deviation s. 3.4.3 Power delay Profile Power delay profile demonstrates the quality of the received signal passing through a multipath channel as a function of time delay. The time delay is the difference of travel time with multipath arrivals. Power delay profile can be described as the squared magnitudes of impulse response by spatial averaging along a local area [24], 2 PDP ( τ ) = h( t; τ ) (5) where, h (t; ) is the modulus value of the impulse response of the signal. With the help of this impulse response, we can get the received signal power as [27], 1 PDP( τ n ) = E{ h( t) } = n= 2 0 2 α δ ( t τ ) n n (6)

P a g e 33 Generally, the paths which come at the later stage in the power delay profile go through more attenuation. Consequently, the power delay profile is usually a descending function of the excess delay. Figure 3.3 shows multipath components with different delays and attenuations. Amplitude Delay Figure 3.3: Multipath components with different delays and attenuations 3.4.4 Saleh and Valenzuela Saleh-Valenzula is a simple multipath model developed for indoor propagation measurements. The basic postulation of this model is the arrival of multipath components (MPC) in the form of clusters. The MPC amplitudes are independent random Rayleigh variables with variance. The variance decays exponentially with both the cluster and excess delays within a cluster. The forming of clusters is concerned with building structure. The components inside a cluster are made from multiple reflections from objects. The clusters and MPC within the cluster that can be derived according to Poisson arrival processes with different rates have exponentially distributed inter-arrival times [24]. IEEE 802.15.4a model is based on Saleh-Valenzuela (SV) model which is shown in figure 3.4. In complex baseband, the impulse response based on SV model is defined as [31]. h discr = L 1 K 1 ( t) α k, l exp( jφk, l ) δ ( t Tl τ k, l ), l= 0 k= 0 (7)

P a g e 34 where, α is the tap weight of the k th component in the cluster, T k, l l is the arrival time of the l th cluster, τ k, l is the delay of the k th multipath component relative to l th cluster arrival time and φk,l denotes the uniformly distributed phases. For a band pass system, the phase angle is taken as uniformly random distributed in a range from 0 to 2π [30]. An important component of the model which is the number of clusters is represent by L, which is supposed as Poisson-distributed. where, l is arrival rate of the cluster. [ ( T T ], l 0 p( TL TL 1 ) = l exp l L L 1) > (8) The arrival times of the ray are modeled with a mixture of two Poisson processes as follows, [ λ ( τ τ )] + ( β 1) λ exp[ λ ( τ τ )], k 0 P( τ k, lτ ( k 1), l ) = βλ1 exp 1 k, l ( k 1), l 2 2 k, l ( k 1), l > (9) where, β is the mixture probability, and λ 1 and λ 2 are the ray arrival rates [30]. The mean power of different component s exponential within each cluster is given as [30] 2 1 E α k, l =Ω1 exp( τ k, l / γ l) γl [(1 β ) λ + βλ + 1] 1 2 (10) Amplitude Clusters Γ Delay 1 / λ 1 / Figure 3.4: Principle of Saleh Valenzuela model

P a g e 35 where, Ω 1 is the integrated power of the l th cluster and γ l is the intra-cluster decay time constant. The mean (over small-scale fading) mean (over cluster-shadowing) energy (normalized to γ l ) of the l th cluster adopts a general exponential decay which can be expressed as [30] 10log( Ω 1 ) = 10log(exp( T 1 / Γ)) + M cluster (11) where, M cluster is a normal distribution random variables with cluster standard deviation around it [30]. The scenarios for the non line of sight (NLOS) define the shapes of the power delay profiles differently [29]. 2 { } γ1+ γ rise Ω1 E α k, l = (1 χ.exp( τ ( k, l) / γ rise)).exp( τ ( k, l) / γ1). (12) γ γ + γ (1 x) The parameter χ represents the attenuation of the first component, γ rise determines how fast the power delay profile γ increases to its maximum and γ 1 determines the decay at the last time. 1 1 rise 3.4.5 Delay dispersion Delay dispersion can be said to be occurring when the channel impulse response remains for a finite quantity of time or the channel happens to be frequency-selective [32]. The effect of delay dispersion can be expressed as the product of the delay spread and the bandwidth of the system. If this product is below unity, then its delay dispersion effect will be low on the system design. And, if the product is higher than unity, then it is said to have a strong delay dispersion effect in the system performance [18]. In multipath channel, delay dispersion is featured by two parameters: root mean square (rms) delays spread and mean excess delay [32]. The mean excess delay is the first moment of the power delay profile (PDP) according to [30]. τ m = PDP( τ ) τdτ PDP( τ ) dτ (13)

P a g e 36 The rms delay spread is the second moment of the power delay profile (PDP) according to [30]. 2 ( ) ( ) PDPτ τ dτ PDPτ τdτ τ rms = (14) PDP( τ ) dτ ( ) PDPτ dτ Delay spread depends upon the distance; nevertheless, for channel simplicity it is neglected. 2 3.4.6 Small scale fading Small scale fading refers to the changes in amplitude, multipath delays or phase of the received signals over a short period of time [29]. Small scale fading takes place because of destructive and constructive interference of multipath components that reaches the receiver at fairly different times [29]. The distribution of small scale amplitudes is Nakagami in this model [30]. 2 m pdf ( x) = Γ( m) Ω m x 2m 1 m exp x Ω 2, (15) where m 1/2 is the Nakagami m factor, gamma function is Γ (m), and mean square value of amplitude is Ω. The parameter m modeled as a log generally random distributed variable. Both values of logarithmic mean µ m and standard deviation m [30]. µ m o m ( τ ) = m k τ (16) σ = mo k mτ (17) Nakagami factor is deterministic and delay independent [30]. m ^ ^

P a g e 37 CHAPTER FOUR IR-UWB Receivers

P a g e 38 Chapter 4: IR-UWB Receivers 4.1 Introduction The immense bandwidth of UWB systems can make the design of the receiver quite difficult in conventional UWB systems that use antipodal or pulse-position modulation with very short pulses [33] [34]. The analog to digital transformation of the whole signaling bandwidth in simple low-power UWB receivers is very hard to implement [34]. Many digital UWB receivers have certain number of analog correlators to accumulate signal energy in a front-end RAKE receiver type architecture [34]. The efficient accumulation of energy in that kind of architecture can be expensive because of a large number of resolvable paths in the standard UWB fading environment. It can create problems in channel estimation even if allowable in the perspective of circuit complexity [34]. Because of these problems encountered in traditional impulsive UWB or DS-UWB, the approach regarding multiband UWB for short-range high data rate applications has been favored [34]. In the following sections, three reference-based noncoherent IR-UWB systems, i.e., Transmit-Reference (TR), Frequency-Shifted Reference (FSR) and Code-Shifted Reference (CSR) systems are discussed. 4.2 Transmitted Reference (TR) UWB System As UWB systems have pulses of short duration and they are characterized by limited power, these characters cause extreme dependence of these systems on timing requirements [44]. These difficult timing requirements make receiver design complicated. In such a situation, Transmitted-Reference (TR) UWB systems can offer a simple and cheap receiver that collects the energy from various multipath components for the correct detection of UWB data [35]. In TR-UWB systems, some amount of the transmitted energy is used for measuring channel [36]. Each frame of the transmitted signals contains two different pulses which are reference and data [37]. The reference pulse has a fixed polarity. The polarity of the data pulse indicates the data bit [38]. These two segments of the signal are separated in time domain. The transmitted TR-UWB signal can be mathematically expressed as the following [39],

P a g e 39 f ( p[ t ( j f + i) Tf ] + b j p[ t ( j f + i) Tf D] ) x( t) = (1) j= 1 i= 0 where p (t) is a UWB pulse with duration T p, frames, f >>1, b j Tf is the frame length, {1,-1} is the information bit transmitted during th j f is the number of time duration. D is the delay between the reference and data pulse. There is a UWB pulse per each frame interval [39]. At the standard TR-UWB receiver side shown in figure 4.1, the received signal is filtered and correlated with a delayed version of itself. The correlated signal is integrated from ( to constitute a decision variable [48]. The value of T M ranges ( j f + i) T f to j f + i) Tf + TM from T p to T f. T f f r (t) Matched Filter Delay D ( j f + i) T f + TM ( j f + i) T f 1 f i= 0 rˆ j sign r ˆ ) ( j bˆ j Figure 4.1: Block diagram of the TR-UWB receiver [34] 4.3 Frequency-Shifted Reference (FSR) UWB System The TR-UWB architecture is able to provide a simple receiver design for a UWB system. However, the implementation of TR-UWB receiver can be quite a challenge. In a low-power integrated circuit environment desired by the TR-UWB receiver, it is hard to construct the delay that handles a wideband signal [34]. To solve this complexity problem of the delay element, a new system called Frequency-Shifted Reference Ultra-Wideband (FSR-UWB) is introduced [34]. The main idea behind FSR-UWB is to propose a TR-UWB system in the frequency domain which excludes the delay element of the standard TR-UWB receiver. Frequency translation of a wideband signal is much easier than implementation of its delay element. In FSR-UWB, the reference signal is translated in frequency (instead of time) to be orthogonal to the data signal [40].

P a g e 40 The key principle proposed in [34] is to enforce the frequency shift of the data signal relative to the reference signal over a symbol period rather than over a frame period. This permits a significant overlapping of the frequency bands occupied by the data-bearing and reference signal. The transmitted FSR-UWB signal can be mathematically expressed as the following: ( f 1 p[ t ( j f + i) Tf ] + bj 2 cos(2 f0 t) p[ t ( j f + i) Tf ]) x( t) = π (2) j= i= 0 where p (t) is a UWB pulse with duration T p, Tf is the frame length, f is the number of frames, f >>1, f = 1/T is the frequency shift of the data signal relative to the reference signal, 0 s b { 1, 1} is the information bit transmitted during j th j time duration [39]. Figure 4.2 T f f shows the FSR-UWB receiver structure. The detailed explanation about FSR-UWB receiver is given in [34]. r (t) Matched Filter ( j f + i) T f + TM ( j f + i) T f 1 f i= 0 rˆ j sign r ˆ ) ( j bˆ j 2 cos(2 π f 0 t ) Figure 4.2: Block diagram of the FSR-UWB receiver [34] A modified form of the traditional FSR-UWB is proposed in [41] which is called Multi- Differential (MD) FSR-UWB, where multiple data carriers use a single reference carrier. Every data signal is a slightly frequency-shifted version of the reference signal [41]. The data carrier frequencies are cautiously selected such that all data signals and the reference signal are orthogonal to each other over the symbol period [41]. This adjustment expands the amount of freedom accessible for signaling in the system [41]. The transmitted MD FSR-UWB signal for M carriers can be mathematically expressed as following [41], 1 f M x( t) = p[ t ( j f + i) Tf ] + bjk 2 cos(2π fkt) p[ t ( j f + i) Tf ] (3) j= i= 0 k= 1

P a g e 41 where b jk {1,-1} is the th k information bit transmitted over th j time duration. The T f f carrier frequency of the th k data signal is expressed as f = ( 2k+ 1) / T. Figure 4.3 shows the MD FSR-UWB receiver structure. The detailed explanation about MD FSR-UWB receiver is given in [41]. For moderate data rate applications, the FSR-UWB scheme performs better than the TR-UWB scheme. However, it is not preferable for high data rate systems, because of the presence of intersymbol interference [41]. Apart from this, when there are many users in the system, the frequency oscillator requires a lot of power [34]. k s r(t) Matched Filter rˆ ( t) (. ) 2 cos( 2πf 1t) cos( 2π t) f k f 1 i= 0 rˆj1 b ˆ j 1 sign ( r ˆj 1) ( j f + i) T f + TM ( j f + i) T f r ij f 1 i= 0 rˆ jk sign r ˆ ) ( jk bˆ jk cos( 2πf M t) f 1 i= 0 rˆ jm sign r ˆ ) ( jm bˆ jm Figure 4.3: Block diagram of MD FSR-UWB receiver [41] 4.4 Code-Shifted Reference (CSR) UWB System Recently a new scheme called Code-Shifted Reference (CSR) has been proposed for IR-UWB systems. In this scheme, the reference and data pulse sequences are separated by a set of shifting and detection codes instead of getting separated by time (TR) or frequency (FSR) [33]. In CSR- UWB, a reference pulse sequence and a single or multiple data pulse sequences are instantaneously transmitted [42]. Every pulse sequence is coded by a particular shifting code. A set of detection codes are used to detect the information bits from the data pulse sequences at the CSR receiver side. The CSR scheme has been able to remove the ultra wideband delay element required in the TR-UWB transceiver because the separation of the reference pulse sequence and the data pulse sequences is performed by the employment of code shifting rather