Second-Order LMS Based Per-Survivor Processing Decoder for Space-Time Trellis Code

An adaptive space-time trellis code (STTC) decoder with thesecond-order LMS (SOLMS) based channel tracking is studied in thispaper. We first show that per-survivor processing (PSP) can beadopted as the approximation of the adaptive maximum likelihoodsequence detection (MLSD) of STTC when there are no periodicallyinserted orthogonal pilot sequences (PIOPS). With thispresupposition, we propose a new decoder for STTC, called SOLMSbased PSP decoder, in which we use min-max optimum SOLMSalgorithm for channel tracking when some ``boundary'' statisticalknowledge of the fading channel is available. Intensive computersimulations have been done to compare the performance of theproposed decoder with other known decoding methods as applied tolong burst transmissions.