Delay processing vs. per survivor techniques for equalization with fading channels

This paper analyses the performance of equalizers for unknown fading channels. We compare two versions of a delay processing equalizer, with a per-survivor equalizer, and with a Viterbi algorithm with known channel information. New adaptive algorithms are developed which adjust the decision delay based on the trellis path metric ratios and the path merging lag. The analysis focuses on the trade-off between computational complexity and achievable bit error rate. The crucial factors include fading rate, noise, equalization trellis size, and accuracy of channel estimates. Simulations are used to determine the effect of each of these factors on algorithm performance.