Influence of trellis termination methods on turbo code performances

Increasing the performances of turbo codes used in communication systems is an important issue at the moment. The method used for trellis termination influences the code performances. This paper describes and compares (in terms of bit error rate - BER and frame error rate - FER) four types of these methods: without trellis termination, termination only of the first trellis, post-interleaver flushing and dual termination. Turbo code performances are evaluated for two lengths of S-random interleaver, using binary phase-shift keying (BPSK) modulation, over Additive White Gaussian Noise (AWGN) channel. The simulations were performed for two generating matrices of the component encoders and the decoding algorithm we have used is a version of Max-Log-MAP.