Performance Simulation Of Turbo Codes And Its Application Study In Still Image Transmission Systems

Turbo codes were presented by C.Berrou in International Conference on Communications in 1993. They hang convolutional code and random interleaver together to realize the idea of random coding, and adopt soft input/soft output (SISO) iterated decoding to approach maximum likelihood decoding (MLD). Turbo codes come closer to approaching Shannon's limit. The excellent performances of Turbo codes were the results of the computer simulations, not from the theory investigation, so the foundational theory of Turbo codes are still unperfect up to the present. They still need to be studied. This thesis systematically studies the performances of Turbo codes over both AWGN channels and Rayleigh fading channels, and discusses the applications of Turbo codes in still image transmission systems. Some significative conclusions are obtained. The main contents are:1. Deriving formulas of MAP algorithm and its modified algorithms in the log domain (including Max-Log-MAP and Log-MAP), andsimulating the performances of Turbo codes over additive white Gaussian noise (AWGN).2. Modifying Log-MAP algorithm to be suitable for Rayleigh fading channels, and researching the performances of Turbo codes over Rayleigh fading channels based on above deriving. Simulation results show that Turbo codes exhibit superior error correction performance over fading channels.3. Studying still image transmission systems without channel coding or with Turbo codes over both AWGN and Rayleigh fading channels by Matlab simulation. Simulation results show that Turbo codes not only advance the dependability of communication systems, but also reduce channel signal-to-noise ratio(SNR).This project is supported by scientific research project of Guangxi Education Department (N0.[2000]392).