Switching LMS linear turbo equalization

摘要: Turbo equalization using linear filters for data detection has been shown to perform nearly as well those based on the original maximum a posteriori probability (MAP) approach. Such methods have taken many forms in literature, from simple least-mean-square (LMS)-based adaptive filtering approaches, minimum mean square error (MMSE)-based that are recursively computed each output symbol iteration. In this paper, we consider class of turbo algorithms which complexity requirements dictate fixed set filter coefficients must be used all symbols and iterations. By computing one such via LMS algorithm assuming unreliable soft information, another highly reliable show switching strategy can employed, achieving performance recomputing at