基于纠错算术码的信源信道自适应联合编解码系统

对信源信道自适应联合编码方法进行了研究,提出了一种新的基于纠错算术码的联合信源信道编解码系统。该系统在编码端利用算术码内嵌禁用符号实现信源信道一体式编码,即利用马尔科夫信源模型和根据信道状态信息自适应地调整禁用符号概率大小从而调整编码码率来实现信道自适应;在解码端,推导出了基于MAP的解码测度数学公式并基于此测度公式提出了一种改进的堆栈序列估计算法。与传统的信道自适应编码算法不同,该自适应编码算法只需调整一个参数:禁用符号,且理论上可获得连续可变的编码码率。实验结果表明,与经典的Grangetto联合编码系统以及分离编码系统相比,所提出的编码系统具有明显改善的性能增益。

Adaptive joint source-channel coding and MAP decoding of error correction arithmetic codes

The adaptive joint source/channel coding methodologies was investigated,and a novel joint source/channel coding/decoding system was proposed based on error correction arithmetic codes.At encoding,this system can achieve an adaptive source/channel integrative coding by utilizing the arithmetic codes with a forbidden symbol.In particular,the Markov source model was used,and the probability of the forbidden symbol was adjusted adaptively according to the channel state information(CSI)so as to provide an adaptive source coding;and then at decoing,an improved sequential search algorithm based on new MAP decoding metric was proposed.Unlike the conventional channel adaptive coding algorithm,only one parameter(ie.a forbidden symbol)must be need to adjust,and a continuous coding rate could be achieved theoretically.Some experimental results show that the proposed system can achieve a higher performance gain,comparing with systems proposed by Grangetto and the separate coding system.