过采样滤波器组作为纠错码的阈值新算法

摘要

考虑量化噪声和信道噪声，搭建基于过采样滤波器组 OFBS（Oversampled Filter Banks）的联合信源信道编码系统，研究系统的传输检错纠错性能。提出新的阈值选取方法，该方法是基于贝叶斯假设检验为理论依据的，能根据实际信号和噪声统计特性做实时分析，得出适当的阈值。通过 Matlab 对系统进行仿真，结果表明信道在不同的转移概率条件，OFBS 作为纠错码的阈值新算法的计算复杂度低，并且有很好的检错纠错性能。通过对一维信号的纠错，信号的峰值信噪比平均提高1到2 db。%Considering the quantisation noise and channel noise,we construct OFBS-based joint source-channel coding (JSCC)system, and study the performance of transmission error detection and correction of the system.We also propose a new threshold selection method, which is based on Bayesian hypothesis testing and takes it for the theoretical basis,it can carry out real-time analysis according to practical signal and noise statistical characteristics to obtain the proper threshold.The system is simulated with Matlab,results show that when channel is under different transition probability conditions,the new threshold algorithm using OFBS as error-correcting codes has low computational complexity,and has good error correction performance as well.Through error correction on one-dimensional signal,the peak signal-noise ratio of the signal improves 1 db to 2 db in average.