Efficient trellis search algorithms for adaptive MLSE on fast Rayleigh fading channels

Analyzes trellis search algorithms for joint sequence estimation and channel tracking, assuming rapidly varying frequency-selective Rayleigh fading. The generalized Viterbi algorithm (GVA) and the M-algorithm (MA) are considered for approximately searching the maximum likelihood path in the trellis diagram and compared with the Viterbi algorithm (VA). All algorithms perform channel tracking utilizing per-survivor processing (PSP) techniques by associating a channel estimate to each hypothetical trellis path, according to the least mean square (LMS) or recursive least square (RLS) algorithm. To reproduce a typical mobile digital communication system, a time-division multiple access (TDMA) data frame is assumed, where each user transmits a block of information symbols with known preamble and tail. For Doppler bands up to one hundredth the symbol frequency and a three path delay profile, the MA is shown to exhibit a slightly better performance than the GVA for an equal number of survivors, and outperform the VA.