基于实数BCH码的联合编码阈值新算法

为了解决实数BCH(Bose-Chaudhuri-Hocquenghem)编码的联合信源信道编码系统中阈值选取复杂问题。考虑量化噪声和信道噪声,提出了基于实数BCH解码阈值新算法。比较原有阈值算法与新阈值算法的计算复杂度以及阈值判决性能。该方法以贝叶斯假设检验为理论依据,分别画出只有信道噪声和只有量化噪声时校验子范数平方的直方图,通过直方图选出适当的信道噪声阈值。仿真结果表明信道在不同的转移概率条件下,实数BCH码作为纠错码的阈值新算法的计算复杂度低,阈值判决性能良好。在转移概率是10-4时,采用新阈值算法的信号的峰值信噪比约提高了1 d B。

A New Threshold Algorithm of Joint Encoding Based on Real BCH

In order to solve complexity of selecting the threshold of joint source-channel coding system based on real BCH. Considering the quantization noise and channel noise, a new threshold algorithm of real BCH decoding is raised. Compare original threshold algorithm and the new threshold algorithm from calculation complexity and judgment performance. The method is based on Bayesian hypothesis testing as the theoretical basis, and drawing histogram of the norm of the syndrome when there is only quantization noise and channel noise separately, then selecting the appropriate threshold of channel noise from histogram. Simulation results show that at different channel transition probability conditions, the new threshold algorithm of real BCH error correction code has a lower calculation complexity and good judgment performance. When the transition probability is 10-4 , the peak signal to noise ratio improves about 1dB using the new threshold algorithm.