Maximum-likelihood sequence estimator with variable number of states

摘要

A maximum-likelihood sequence estimator (MLSE) with a variable number of states. A channel response estimator (101) calculates channel responses from a received signal having a predetermined burst length and a reference signal. A controller (102) checks the latest (L-th) response having a larger power than a predetermined threshold value of the channel responses and determines the L number of effective channel responses having the larger power. A Viterbi equilizer (103) with a variable number of states executes a maximum-likelihood sequence estimation on the basis of the trellis with M ^(L-1) states (M is a multi-value number of a modulation signal) using the L number of effective channel responses. The MLSE is operated with the minimum number of states every burst to reduce an average processing amount without degrading characteristics and to achieve a low consumption power of receivers.