IMPLEMENTATION OF TURBO DECODERS

Abstract

Turbo codes are error correction codes which approach Shannon's limit very closely. They involve parallel concatenation of recursive systematic convolutional codes and are decoded using an iterative decoding algorithm. The decoding strategy used to decode turbo codes is the basis of `turbo principle', which connotes the iterated exchange of soft information between different blocks of a communication receiver. The potential importance and applicability of this principle has been found to extend to a wide range of problems in communication. In this dissertation, simulation of turbo coding and decoding has been carried out in MATLAB environment. The performance of turbo codes has been evaluated using the MAP (maximum aposteriori probability) algorithm and its reduced state variant, the threshold-BOJR algorithm and finally, using soft output viterbi algorithm (SOYA).