神经网络在移动信道卷积码译码中的应用  

Application of Neural Decoder for Convolutional Codes on the Mobile Channel

作　　者：刘华章[1] 蔡新举[1] 任建存[2] 

机构地区：[1]海军航空工程学院电子工程系,山东烟台264001 [2]海军航空工程学院自动控制工程系,山东烟台264001

出　　处：《烟台大学学报（自然科学与工程版）》2002年第3期216-219,228,共5页Journal of Yantai University(Natural Science and Engineering Edition)

摘　　要：分析运用误差反向传播 (EBP)多层感知器神经网络 (NN)方法来进行卷积码译码的可能性 .仿真结果表明 :当涉及的码字位于译码窗口中间时 ,其性能最佳 ,而且即使考虑大的窗口性能也不会提高 .当神经网络规模较大时 ,神经网络体现出自适应、冗余容错性 .窗口宽度等于 3的系统最小 。The possibility of decoding convolutional codes by means of the multilayer perception neural entwork (NN) trained with the error back propagation (EBP) algorithm is considered. The result of the simulation shows:the best performance of the NN decoder is obtained when the relerence codeword position is at the center of the working window and the performance dose not improve considering larger window sizes.The better performance of adaptivity and redundancy fault tolerant of the NN is showed on the larger NN scale.The system is valuely recommended,because its scale is minimum and its decoding bit error probability Pb is optimal for the window width equals 3.

关 键 词：神经网络 移动信道 卷积码译码 衰落信道 误差反向传播算法 码字 译码窗口 

分 类 号：TN922.22[电子电信—通信与信息系统] TP183[电子电信—信息与通信工程]