摘要
本文提出一种非线性码神经网络译码方案,在纠错能力范围内对满足码距特性的一般非线性码以零错误概率进行纠错译码,并在检错能力范围内检错。文中具体描述了神经网络模型构造、学习算法及其理论依据。最后通过非线性等重码的译码实例表明此方案的有效性及理论和应用价值。
A decoding strategy for nonlinear codes using a neural network is presented. Within the capacity of error correction, it can correct and detect errors of general nonlinear codes which have some specific code distance with zero error probability. This paper describes structure of the neural network, learning algorithm and theory analysis. Finally, one decoding example: nonlinear constant weight code is demonstrated to prove the availability and values of theory and application.
基金
国家自然科学基金(批准号 69772035
69896240)
关键词
非线性码
译码
神经网络
Nonlinear codes, Decoding, Neural networks
