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 ,所以该系统是值得推荐的系统 The possibility of decoding the convolutional code using the method of multi-layer perceptron Neural Network (NN) with Error Back Propagation (EBP) is analyzed.The simulation results show that the performance of the codeword is best when it is located in the middle of the decoding window , And it does not improve even considering the large window performance.When the neural network is large in scale, the neural network shows adaptive and redundant fault tolerance.The system with minimum window width of 3 is the most recommended system