Research On Entropy And LDPC Coding Based Joint Source Channel Decoding Technologies

With the development of wireless communication technologies, a variety of multimedia services emerge at a historic moment and develop rapidly so that multimedia data transmission has become the mainstream of communication services. However, in multimedia communication, the amount of data transferred is very large and the channel environment is complicated. Therefore multimedia applications put forward higher requirements on the validity and reliability of the communication system. Joint source channel decoding (JSCD)technologies, which provide technical support for the next generation communication system,are proposed to solve this problem. In the conditions of assuring validity, JSCD can provide higher reliability compared to separate source channel decoding (SSCD). Based on the national defense pre-research project, this paper analyzes the JSCD technologies based on entropy codes with low-density parity-check (LDPC) codes.First of all, the basic principle of LDPC codes and its main encoding and decoding algorithms are introduced. Then the demodulator hard decision decoding method is studied.Secondly, we research Huffman codes soft-in soft-out (SISO) decoding based on Bahl-Cocke-Jelinek-Raviv (BCJR) algorithm and a modified JSCD scheme which concatenates Huffman codes with LDPC codes is proposed with low implementation complexity. This scheme makes Huffman-SISO decoder applies only once BCJR forward-backward operation in the process of decoding. Simulation results show that the proposed JSCD method provides a gain of about 0.25 dB with respect to SSCD based on Huffman codes and LDPC codes. Thirdly, the theory and the realization of arithmetic codes (AC) are discussed. Then, we study the encoding principle and error detection mechanism of error resilient arithmetic code. For the security of AC, a new M-ary symbol security arithmetic coding (M-SAC) approach is presented. The M-SAC, which satisfies highly security without loss of compression efficiency, has strong resistant to selective plaintext and ciphertext attacks. Finally, we research BCJR and Chase-type algorithms for AC-SISO decoding. The theoretical analysis and simulation results show that the AC-SISO decoder based on Chase-type algorithm has lower computational complexity and does not effect on compression efficiency. Next, a novel low-complexity iterative JSCD scheme which serially concatenates AC with LDPC codes is proposed, where the reliabilities of information bits are passed between AC-SISO decoder and LDPC decoder to improve the performance of decoding. Results show that the iterative JSCD we proposed outperforms SSCD based on AC and LDPC codes and an additional gain of about 0.2 dB is achieved.