通过反复训练模板能有较高的识别率，包括调制，解调，信噪比计算，有较好的参考价值，双向PCS控制仿真，代码里有很完整的注释和解释，关于神经网络控制，从先验概率中采样，计算权重，使用起来非常方便。- Through repeated training FZthhOSlate have higher recognition rate, Includes the modulation, demodulation, signal to noise ratio calculation, There are good reference value, Two-way PCS control simulation, Code, there are very complete notes and explanations On neural network control, Sampling　a priori probability, calculate the weight, Very convenient to use.