Nested Distributed Turbo Code for Relay Channels

Abstract

Distributed turbo coding (DTC) has been shown to be able to approach the information-theoretic capacity of wireless relay networks. However, decoding errors at the relay can lead to severe error propagation and hinder the achievement of such capacity. In order to increase the decoding success rate at the relay, thereby enhancing the DTC system performance, we propose a novel DTC scheme, referred to as nested DTC, for a two-hop triangular relay network. In this scheme, we use a turbo code instead of a single convolutional code as does in the conventional DTC. At the source node, the four output streams of the two component convolutional encoders are converted into two streams by adding in modulo-2 addition. An extended 2-D log MAP decoding is used at both relay and destination to retrieve the information symbols. Numerical results show that the error performance of the proposed nested DTC scheme significantly outperforms the existing DTC schemes and performs within only 1 dB of the DTC technique with automatic repeat request (ARQ). The performance superiority of the nested DTC scheme is attributed to the retention of the full information due to the nested encoding structure.