Golay码的神经网络译码算法

本文提出了Golay码的一个神经网络译码器N(23,12),证明了Golay码的最大似然译码等价于N(23,12)收敛于能量函数的全局稳定状态。N(23,12)是一个异步的组合逻辑电路,能用12个大数逻辑门和77个异或门电路来实现。根据Golay码的循环结构,本文还提出了一个神经网络译码算法。计算机模拟表明,该算法达到了Golay码的纠错能力,并且,译码速度优于Kasami的修正捕错译码算法。新算法作为一种完全译码算法,不仅适用于Golay码,而且适用于任何循环码,特别是平方剩余码的译码。

A Neural Network Decoding Algorithm for the Golay Code

A neural network decoder N (23,12) for the Golay code is presented in this paper. The maximum likelihood decoding of the Golay code is proved to the equivalent to that N (23,12) converges the globe stable state of a energy function. N (23,12) is an asynchronous combinational logic circuit which can be implemented with 12 majestic-logic gates and 77 exclusive-OR gates. Based on the cyclic structure of the Golay code,a neural network decoding algorithm is also put forward. Computer simulation indicates that the algorithm reaches the error correcting ability of the Golay code. Moreover,its decoding speed is faster than that of the Kasami's improved error-trapping decoding algorithm. The new algorithm, as a complete decoding algorithm,is suitable to decoding not only the Golay code, but also all the cyclic codes,especially,the quadratic residual codes.