Genetic Algorithm (GA)-Based Detection for Coded Partial-Response Channels

Abstract

The Bahl-Cocke-Jelinek-Raviv (BCJR) detector for turbo equalization over coded partial-response channels has a complexity growing exponentially with channel memory length. In this paper, we consider the soft-in/soft-out (SISO) channel detection from a combinatorial optimization viewpoint and propose a low-complexity detector based on an efficient implementation of the genetic algorithm (GA). Simulation results show that the proposed detector can approach the bit-error-rate (BER) performance of the optimal BCJR algorithm and outperform other suboptimal schemes.