Blind Estimation of Multi-Path and Multi-User Spread Spectrum Channels and Jammer Excision via the Evolutionary Spectral Theory

Despite the significant advantages of direct sequence spreadspectrum communications, whenever the number of users increases orthe received signal is corrupted by an intentional jammer signal,it is necessary to model and estimate the channel effects in orderto equalize the received signal, as well as to excise the jammingsignals from it. Due to multi-path and Doppler effects in thetransmission channels, they are modeled as random, time-varyingsystems. Considering a wide sense stationary channel during thetransmission of a number of bits, a linear time-varying modelcharacterized by a random number of paths, each beingcharacterized by a delay, an attenuation factor and a Dopplerfrequency shift, is shown to be an appropriate channel model. Itis shown that the estimation of the parameters of such models ispossible by means of the spreading function, related to thetime-varying frequency response of the system and the associatedevolutionary kernels. Applying the time-frequency orfrequency-frequency discrete evolutionary transforms, we show thata blind estimation procedure is possible by computing thespreading function from the discrete evolutionary transform ofthe received signal. The estimation also requires the synchronizedpseudo-noise sequence for either of the users we are interestedin. The estimation procedure requires to adaptively implementingthe discrete evolutionary transform to estimate the spreadingfunction and determine the channel parameters. Once the number ofpaths, delays, Doppler frequencies and attenuations characterizingthe channel are found, a decision parameter can be obtained todetermine the transmitted bit. We will show also that ourestimation approach supports multiuser communication applicationssuch as uplink and downlink in wireless communicationtransmissions. In the case of an intentional jamming, common inmilitary applications, we consider a receiver based onnon-stationary Wiener masking that excises such jammer as well asinterference from other users. Both the mask and the optimalestimator are obtained from the discrete evolutionarytransformation. The estimated parameters from the computedspreading function, corresponding to the closest to the line ofsight signal path, provide an efficient detection scheme. Ourprocedures are illustrated with simulations, that display thebit-error rate for different levels of channel noise and jammersignals.