SURVEY OF LINEAR BLOCK ESTIMATION ALGORITHMS FOR THE DETECTION OF SPREAD-SPECTRUM SIGNALS TRANSMITTED OVER FREQUENCY-SELECTIVE CHANNELS

The well known optimum approach to detect spread spectrum signals tran smitted in bursts over frequency selective radio channels is matched f iltering, which performs despreading, and subsequent Viterbi equalizat ion (VE) to cope with intersymbol interference (ISI). With respect to complexity, VE is feasible only if data modulation schemes with a few symbol levels as e.g. 2PSK are used and if the delay spread of the cha nnel is not too large. The paper gives a survey of suboptimum data det ectors based on linear block estimation. Such data detectors are less expensive than VE especially in the case of multilevel data modulation schemes as 4PSK or 16QAM. Special emphasis is laid on data detectors based on Gauss-Markoff estimation because these detectors combine the advantages of unbiasedness and minimum variance of the estimate. In co mputer simulations, the Gauss-Markoff estimation algorithm is applied to spread spectrum burst transmission over radio channels specified by COST 207. It is shown that the SNR degradation which is a measure of the suboptimality of the detector does not exceed a few dB, and that e ven moderate spectrum spreading considerably reduces the detrimental e ffect of channel frequency selectivity.