Dynamic assignment of probability distribution of extrinsic information for turbo decoding over AWGN and Rayleigh fading channels

This paper proposes a strategy using a generalized Gaussian distribution (GGD) to model the statistical properties of the extrinsic information generated from the component decoders applied to the maximum a posteriori (MAP) algorithm in order to improve the performance of an iterative turbo decoder for finite block lengths. A matching technique based on the measured moments and distance criterion is introduced to dynamically select the appropriate parameter values of the GGD conditional probability density functions of the extrinsic information in each iteration. Simulation is performed to evaluate the performance of the proposed strategy for various block lengths and to examine the statistical behaviour of the intrinsic information in different iterations for both Additive White Gaussian Noise (AWGN) and Rayleigh fading channels. The results indicate that the GGD model with dynamic parameter assignment is appropriate to track the varying characteristics of the extrinsic information and helpful for the proposed strategy to offer performance gain in case of medium block lengths.