A study on turbo decoding iterative algorithms [articol]

Abstract

In the paper, a study of some turbo decoding iterative algorithms: MAP, MaxLogMAP, LogMAP, is presented. For the correction of the approximation used in the MAxLogMAP algorithm two methods are proposed obtaining two LogMAP algorithm variants. All algorithms variants have been simulated to make possible a comparison from the bit error rate (BER) point of view, in order to provide an optimization for each algorithm. The simulations were made for AWGN channel. Two component codes with generator matrix: G₁ =[1, 5/7] and two interleaver types: pseudo-random [1] and S-interleaver (S=29) are used. The interleaving length is N=1784. The number of emitted blocks in one simulation depends of signal to noise ration, SNR, to obtain a good precision of resulted curve.