作者: T.M. Aulin
DOI: 10.1109/26.752126
关键词:
摘要: The problem of performing breadth-first maximum likelihood sequence detection (MLSD) under given structural and complexity constraints is solved results in a family optimal detectors. Given trellis with S states, these are partitioned into C classes where B paths each class selected recursively symbol interval. derived result to retain only those which closest the received signal Euclidean (Hamming) distance sense. Each member SA(B, C) detectors (SA denotes search algorithm) performs constrained MLSD for additive white Gaussian noise (AWGN) (BSC) channel. unconstrained solution Viterbi algorithm (VA). Analysis tools developed asymptotic (SNR) probability losing correct path associated new measure AWGN case, vector (VED). traditional scalar special case this, termed (SED). generality this VED pointed out. Some general reductions exemplify VA approach.