DESIGN AND PERFORMANCE OF TIME-FREQUENCY-CODED SIGNALS FOR SPREAD-SPECTRUM SYSTEMS

KHIEM VAN CAI, Purdue University

Abstract

A fundamental problem in multiple-access communication systems employing spread-spectrum signals is that of designing a large set of signals that have uniformly small crosscorrelation properties. The uniformly small crosscorrelation property is necessary to permit the signals to be detected with matched filters and a threshold operation. The size of the signal set should be as large as possible in order that the greatest number of potential users, each having a uniquely assigned signal, can be accommodated by the system. The basic theoretical problem that emerges in this context is of determining the number of signals. We consider a set of equal-energy signals having a duration of T and occupying a bandwidth of B, but not necessarily having the same energy spectrum within that band. These signals are further constrained to have nearly uniformly small crosscorrelation functions for any pair of signals having any relative time shift. The theoretical maximum number of such signals that can exist, having a specified time-bandwidth product (TB) and desired crosscorrelation properties, has been determined. The new bound appears to be very tight and there are some methods of signals design that achieve this bound. The solution to the above basic theoretical problem suggests that there may exist a very large set of signals that have both small rms value and maximum value of crosscorrelation function. In the second part of this research, we devise some constructive methods for designing large sets of time-frequency-coded signals that have both specified time-bandwidth product and desired crosscorrelation properties. The coding sequences of these signals have the so-called "t('th) - order of difference" property that tends to keep the crosscorrelation functions uniformly small. A model of a receiver for the spread-spectrum multiple-access system employing the above frequency-hopping signals is investigated. The receiver consists of matched filters (or correlator equivalent), a threshold operator and operates in an asynchronous manner. The behavior of the designed signals, when this type of receiver is applied, is investigated in detail. The analysis indicates that when many active users access the system simultaneously, the signal-to-noise ratio at the output of the demodulator is quite high. This suggests that the probability of error is very small. The methods of signals design presented in this research provide an extremely large set of signals with specified crosscorrelation properties and can be used for the cellular mobile communication system proposed recently.

Degree

Ph.D.

Subject Area

Electrical engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS