Generalized projection algorithm for blind interference suppression in DS/CDMA communications

Abstract

New blind adaptive interference suppression algorithms for DS/CDMA communication systems are presented. A generalized projection algorithm is presented to update a multiple access interference (MAI) suppression filter in a static communication environment. The algorithm adaptively adjusts the filter coefficients by iteratively projecting them onto constraint sets. The constraint sets consist of data sets and property sets, which are obtained from the received signal and prior knowledge of the desired signal, respectively. A space alternating generalized projection algorithm with approximate EM mapping is also presented to suppress the MAI in a dynamic communication environment. This algorithm jointly estimates the filter and desired signal amplitude using a stochastic approximation of the EM algorithm. The performance characteristics of the proposed algorithms are shown by output signal-to-interference ratio and bit-error-rate simulation results. The results show that the proposed algorithms significantly outperform the existing adaptive algorithms in steady state.