Low-Complexity Iterative Equalisation and Decoding for Wireless Optical Communications

Abstract

A low-complexity scheme of iterative equalisation and decoding by combining a recursive systematic convolutional code and a pulse-position modulation is proposed here. A graph-based equalisation for intersymbol interference (ISI) known at both transmitter and receiver is considered. By representing the memory channel with ISI as the factor graph and applying sum-product (SP) algorithm to this graph, a posteriori probability (APP) of the desired symbol necessary to implement iterative equalisation and decoding is derived. A partial response precoding is used to reduce the span of ISI from a possible infinite number of two baud periods. This precoding scheme makes the factor graph of memory channel cycle-free, and SP algorithm for combating ISI converges to an optimum detection. Numerical results show that the proposed low-complexity strategy has almost the same performance as the optimum turbo equalisation.