Joint data and channel estimation using fast blind trellis searchtechniques

摘要

Proposes a novel method for blind sequence estimation of data thatare transmitted over unknown linear distortive channels. For everypossible sequence that can be transmitted, the procedure finds the bestpossible channel fit corresponding to the noisy channel output sequence(typically using the least squares procedure). The data and the channelare estimated to be those with the overall best fit. This simple butexhaustive search procedure for blind channel equalization isnoninstrumentable because of the exponential growth in its complexitywith the length of the data sequence. The author proposes a suboptimaltrellis search algorithm that strives to achieve the optimal performancewith only a linear complexity. Fast convergence of the algorithm inestimating the channel is demonstrated for binary pulse amplitudemodulation (2-PAM) and a variety of channels. Convergence at highsignal-to-noise ratios (30 dB) typically occurs within 50 symbols forthe channels considered. At a low SNR (10 dB), convergence occurs within100 symbols. The number of states in the decoder trellis, which is ameasure of the decoding complexity, is ML, whereM is the number of modulation levels and L is thechannel memory