Improvement On The Decoding Algorithm Of Turbo Codes And Its Application On Distributed Video Coding

As the most efficient channel code, Turbo codes have drawn much attention of various research groups. Under the support of the National 863 Project: Digital Video and Audio Coding, Transmission and Application System (No.: 2002AA119010), this dissertation has done a lot of researches on Turbo codes. Firstly an overview of coding technology is given in this thesis. And then some key points of Turbo codes and RCPT codes are presented including component codes, interleaver, decoding algorithm, etc. After the comparative studying of MAP decoding algorithms, a novel decoding algorithm of turbo codes is proposed. Based on the MAX-Log-MAP algorithm, the novel algorithm not only uses the segment linearity function but also takes the statistic of the decoding data into account. Experimental results demonstrate that it performs better than the MAX-Log-MAP algorithm using traditional segment linearity function. This thesis also introduces the application of Turbo codes in the distributed video coding. It enables low-complexity encoding by shifting the bulk of the computation to the decoder. Such schemes are promising for mobile application environment, both for transmission and storage applications. After explicating the theory of the distributing video coding, an application model based on H.264 is proposed, which uses turbo codes as distributing coder. Experiment results show that the distributed video coding using Turbo codes outperforms conventional intraframe coding but is not as good as the performance of a convolutional interframe video coder.