采用偏最小二乘法，是学习PCA特征提取的很好的学习资料，多目标跟踪的粒子滤波器，分析了该信号的时域、频域、倒谱，循环谱等,给出接收信号眼图及系统仿真误码率，用于建立主成分分析模型，模式识别中的bayes判别分析算法。- Partial least squares method, Is a good learning materials to learn PCA feature extraction, Multi-target tracking particle filter, Analysis of the signal time domain, frequency domain, cepstrum, cyclic spectrum, etc. The received signal is given eye and BER simulation systems, Principal component analysis model for establishing, Pattern Recognition bayes discriminant analysis algorithm.