 Shopping Bag ( 0 items )

All (12) from $63.99

New (8) from $74.27

Used (4) from $63.99
More About This Textbook
Overview
Editorial Reviews
From the Publisher
"...presents the latest findings from militarysponsored academic research on wireless communications from a hardware design perspective..." (SciTech Book News, Vol. 25, No. 4, December 2001)Product Details
Related Subjects
Meet the Author
George Haddad, PhD, is Professor of Electrical Engineering and Computer Science at the University of Michigan. He has authored numerous publications in the areas of solidstate microwave devices and circuits.
James Harvey is a Research Program Manager at the Electronics Division, Army Research Office, with primary responsibility for the fields of electromagnetics, antennas and antenna structures, innovative microwave and millimeterwave circuit integration, lowpower/minimumpower system design, and landmine detection.
Read an Excerpt
RF Technologies for LowPower Wireless Communications
John Wiley & Sons
Copyright © 2001 John Wiley & Sons, Inc.All right reserved.
ISBN: 0471382671
Chapter One
WIRELESS COMMUNICATIONS SYSTEM ARCHITECTURE AND PERFORMANCEWayne Stark
Department of Electrical Engineering and Computer Science The University of Michigan, Ann Arbor
Larry Milstein
Department of Electrical and Computer Engineering University of CaliforniaSan Diego
1.1 INTRODUCTION
Low power consumption has recently become an important consideration in the design of commercial and military communications systems. In a commercial cellular system, low power consumption means long talk time or standby time. In a military communications system, low power is necessary to maximize a mission time or equivalently reduce the weight due to batteries that a soldier must carry. This book focuses attention on critical devices and system design for low power communications systems. Most of the remaining parts of this book consider particular devices for achieving low power design of a wireless communications system. This includes mixers, oscillators, filters, and other circuitry. In this chapter, however, we focus on some of the higher level system architecture issues for low power design of a wireless communications system. To begin we discuss some of the goals in a wireless communications system along with some of the challenges posed by a wireless medium used for communications.
1.2 PERFORMANCE GOALS AND WIRELESS MEDIUM CHALLENGES
A system level (functional) block diagram of a wireless communications system is shown in Figure 1.1. In this Figure the source of information could be a voice signal, a video signal, situation awareness information (e.g., position information of a soldier), an image, a data file, or command and control data. The source encoder processes the information and formats the information into a sequence of information bits [member of] {[+ or ]1}. The goal of the source encoder is to remove the unstructured redundancy from the source so that the rate of information bits at the output of the source encoder is as small as possible within a constraint on complexity. The channel encoder adds structured redundancy to the information bits for the purpose of protecting the data from distortion and noise in the channel. The modulator maps the sequence of coded bits into waveforms that are suitable for transmission over the channel. In some systems the modulated waveform is also spread over a bandwidth much larger than the data rate. These systems, called spreadspectrum systems, achieve a certain robustness to fading and interference not possible with narrowband systems. The channel distorts the signal in several ways. First, the signal amplitude decreases due to the distance between the transmitter and receiver. This is generally referred to as propagation loss. Second, due to obstacles the signal amplitude is attenuated. This is called shadowing. Finally, because of multiple propagation paths between the transmitter antenna and the receiver antenna, the signal waveform is distorted. Multipath fading can be either constructive, if the phases of different paths are the same, or destructive, if the phases of the different paths cause cancellation. The destructive or constructive nature of the fading depends on the carrier frequency of the signal and is thus called frequency selective fading. For a narrowband signal (signal bandwidth small relative to the inverse delay spread of the channel), multipath fading acts like a random attenuation of the signal. When the fading is constructive the bit error probability can be very small. When the fading is destructive the bit error probability becomes quite large. The average overall received amplitude value causes a significant loss in performance (on the order of 3040 dB loss). However, with proper error control coding or diversity this loss in performance can essentially be eliminated.
In addition to propagation effects, typically there is noise at the receiver that is uncorrelated with the transmitted signal. Thermal (shot) noise due to motion of the electrons in the receiver is one form of this noise. Other users occupying the same frequency band or in adjacent bands with interfering sidelobes is another source of this noise. In commercial as well as military communications systems interference from other users using the same frequency band (perhaps geographically separated) can be a dominant source of noise. In a military communications system hostile jamming is also a possibility that must be considered. Hostile jamming can easily thwart conventional communications system design and must be considered in a military communications scenario.
The receiver's goal is to reproduce at the output of the source decoder the informationbearing signal, be it a voice signal or a data file, as accurately as possible with minimal delay and minimal power consumed by the transmitter and receiver. The structure of the receiver is that of a demodulator, channel decoder, and source decoder. The demodulator maps a received waveform into a sequence of decision variables for the coded data. The channel decoder attempts to determine the information bits using the knowledge of the codebook (set of possible encoded sequences) of the encoder. The source decoder then attempts to reproduce the information.
In this chapter we limit discussion to an information source that is random data with equal probability of being 0 or 1 with no memory; that is, the bit sequence is a sequence of independent, identically distributed binary random variables. For this source there is no redundancy in the source, so no redundancy can be removed by a source encoder.
There are important parameters when designing a communications system. These include data rate [R.sub.b] (bits/s, or bps), at the input to the channel encoder, the bandwidth W (Hz), received signal power P (watts), noise power density [N.sub.0]/2 (W/Hz), and bit error rate [P.sub.e,b]. There are fundamental tradeoffs between the amount of power or equivalently the signaltonoise ratio used and the data rate possible for a given bit error probability, [P.sub.e,b]. For ideal additive white Gaussian noise channels with no multipath fading and infinite delay and complexity, the relation between data rate, received power, noise power, and bandwidth for [P.sub.e,b] approaching zero was determined by Shannon as
(1:1) [R.sub.b] < W [log.sub.2](1 + P/[N.sub.0]W):
If we let [E.sub.b] = P/[R.sub.b] represent the energy used per data bit (joules per bit), then an equivalent condition for reliable communication is
[E.sub.b]/[N.sub.0] > [2.sup.[R.sub.b]/W]  1/[R.sub.b]/W
This relation determines the minimum received signal energy for reliable communications as a function of the spectral efficiency [R.sub.b]/W (bps/Hz). The interpretation of this condition is that for lower spectral efficiency, lower signal energy is required for reliable communications. The tradeoff between bandwidth efficiency and energy efficiency is illustrated in Figure 1.2. Besides the tradeoff for an optimal modulation scheme, the tradeoff is also shown for three modulation techniques: binary phase shift keying (BPSK), quaternary phase shift keying (QPSK), and 8ary phase shift keying (8PSK).
In this figure the only channel impairment is additive white Gaussian noise. Other factors in a realistic environment are multipath fading, interference from other users, and adjacent channel interference. In addition, the energy is the received signal energy and does not take into account the energy consumed by the processing circuitry. For example, power consumption of signal processing algorithms (demodulation, decoding) are not included. Inefficiencies of power amplifiers and low noise amplifiers are not included. These will be discussed in subsequent sections and chapters. These fundamental tradeoffs between energy consumed for transmission and data rate were discovered more than 50 years ago by Shannon (see Cover and Thomas). It has been the goal of communications engineers to come close to achieving the upper bound on data rate (called the channel capacity) or equivalently the lower bound on the signaltonoise ratio.
To come close to achieving the goals of minimum energy consumption, channel coding and modulation techniques as well as demodulation and decoding techniques must be carefully designed. These techniques are discussed in the next two sections.
1.3 MODULATION TECHNIQUES
In this section we describe several different modulation schemes. We begin with narrowband techniques whereby the signal bandwidth and the data rate are roughly equal. In wideband techniques, or spreadspectrum techniques, the signal bandwidth is much larger than the data rate. These techniques are able to exploit the frequencyselective fading of the channel. For more details see Proakis.
1.3.1 Narrowband Techniques
A very simple narrowband modulation scheme is binary phase shift keying (BPSK). The transmitter and receiver for BPSK are shown in Figures 1.3 and 1.4, respectively. A sequence of data bits [b.sub.l] [member of] [+ or ] 1 is mapped into a data stream and filtered. The filtered data stream is modulated onto a carrier and is amplified before being radiated by the antenna. The purpose of the filter is to confine the spectrum of the signal to the bandwidth mask for the allocated frequency. The signal is converted from baseband by the mixer to the desired center or carrier frequency (upconversion). The signal is then amplified before transmission. With ideal devices (mixers, filters, amplifiers) this is all that is needed for transmission. However, the mixers and amplifiers typically introduce some additional problems. The amplifier, for example, may not be completely linear. The nonlinearity can cause the bandwidth of the signal to increase (spectral regrowth), as will be discussed later.
For now, assume that the filter, mixer, and amplifier are ideal devices. In this case the transmitted (radiated) signal can be written as
(1.2) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
where P is the transmitted power, T is the duration of a data bit or the inverse of the data rate [R.sub.b], [f.sub.c] is the carrier frequency, and h(t) is the impulse response of the pulseshaping filter. There are various choices for the pulseshaping filter. A filter with impulse response being a rectangular pulse of duration T seconds results in a constant envelope signal (peaktomean envelope ratio of 1) but has large spectral splatter, whereas a Nyquisttype pulse has high envelope variation and no spectral splatter. The disadvantage of high envelope variation is that it will be distorted by an amplifier operating in a power efficient mode because of the amplifier's nonlinear characteristics. Thus there is a tradeoff between power efficiency and bandwidth efficiency in the design of the modulation.
The simplest channel model is called the additive white Gaussian noise (AWGN) channel. In this model the received signal is the transmitted signal (appropriately attenuated) plus additive white Gaussian noise:
(1.3) r(t) = [alpha]s (t) + n(t). The noise is assumed to be white with twosided power spectral density [N.sub.0]/2 W/Hz.
The receiver for BPSK is shown in Figure 1.4. The front end low noise amplifier sets the internal noise figure for the receiver. The mixer converts the radio frequency (RF) signal to baseband. The filter rejects outofband noise while passing the desired signal. The optimal filter in the presence of additive white Gaussian noise alone is the matched filter (a filter matched to the transmitter filter). This very simplified diagram ignores many problems associated with nonideal devices. For the case of ideal amplifiers and a transmit filter and receiver filter satisfying the Nyquist Criteria for no intersymbol interference, the receiver filter output can be expressed as
[X.sub.l] = [square root of ([bar.E][b.sub.l1] + [[eta].sub.1]),
where bar.E is the received energy (bar.E = [[alpha].sup.2]PT) and [[eta].sub.l] is a Gaussian distributed random variable with mean zero and variance [N.sub.0]/2. The decision rule is to decide [b.sub.l1] = +1 if [X.sub.l] > 0 and to decide [b.sub.l1] = 1 otherwise. For the simple case of an additive white Gaussian noise channel, the error probability is
[P.sub.e,b] = Q ([square of root (2[bar.E])/[N.sub.0]
where [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. This is shown in Figure 1.5.
From Figure 1.5 it can be seen that in order to provide error probabilities around [10.sup.5] it is necessary for the received signaltonoise ratio to [bar.E]/[N.sub.0] = 9:6 dB. The capacity curve for BPSK in Figure 1.2, however, indicates that if we are willing to lower the rate of transmission we can significantly save on energy. For example, it is possible to have a nearly 0 dB signaltonoise ratio if we are willing to reduce the rate of transmission by 50%. Thus about 9.6 dB decrease in signal power is possible with a 50% reduction in transmission rate.
The above analysis is for the case of additive white Gaussian noise channels. Unfortunately, wireless channels are not accurately modeled by just additive white Gaussian noise. A reasonable model for a wireless channel with relatively small bandwidth is that of a flat Rayleigh fading channel. While there are more complex models, the Rayleigh fading channel model is a model that provides the essential effect. In the Rayleigh fading model the received signal is still given by Eq. (1.3). However, [alpha] is a Rayleigh distributed random variable that is sometimes large (constructive addition of multiple propagation paths) and sometimes small (destructive addition of multiple propagation paths). However, the small values of a cause the signaltonoise ratio to drop and thus the error probability to increase significantly. The large values of a corresponding to constructive addition of the multiple propagation paths result in the error probability being very small. However, when the average error probability is determined there is significant loss in performance. The average error probability with Rayleigh fading and BPSK is
(1.4) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII],
where f(r) is the Rayleigh density and [bar.E] is the average received energy. The average error probability as a function of the average received energy is shown in Figure 1.6. Included in this figure is the performance with just white Gaussian noise. As can be seen from the figure, there is a significant loss in performance with Rayleigh fading. At a bit error rate of [10.sup.5] the loss in performance is about 35 dB.
Continues...
Table of Contents