An Efficient Channel Impulse Response Estimation Method for DS-CDMA Communications

Abstract

A method is proposed for determining the minimum mean square error (MMSE) channel impulse response estimations for a direct-sequence code-division multiple-access (DS-CDMA) communication system with multipath reception. Typical systems employ a RAKE receiver with tap weights estimated using a simple correlation followed by energy detection. While computationally simple to implement, such a method results in a sub-optimal solution. The standard MMSE estimator, however, requires calculation of the sample autocorrelation matrix and its inverse in order to determine the optimal weights. While providing an optimal solution to the AWGN channel impulse response estimation problem, it is often too computationally intensive to implement in real-time systems. The proposed method, based on a generalized side lobe canceller (GSC), performs a direct computation of the optimal weights using a systolic array, bypassing the need to determine the autocorrelation matrix and its inverse. Development of the channel impulse response estimation algorithm is described and estimations for simulated single and multiple access scenarios are presented for evaluation.