On decoding of multidimensional single parity turbo product codes

Abstract

The paper deals with an iterative decoding algorithm for two or more dimensional turbo product codes, built using single parity block codes. Decoding is based on soft-input/soft-output (SISO) decoders for decoding component codes. The soft inputs of the decoder include two values, the soft channel value and an a priori value. The soft output of the component's decoder is an extrinsic value, which is an estimation of the log-likelihood ratio (LLR) of binary decisions given by the decoding algorithm. The extrinsic values are used in feedback as a priori information. Simulations show that simple parity block codes used as the component codes gives quite good results, although a very simple decoding algorithm is used.