基于Chebyshev数值逼近的实时信号预测及其应用

提出了基于Ｃｈｅｂｙｓｈｅｖ数值逼近的时间序列直接多步预测算法。该算法具有模型简单、所需的观测样本容量小、易于在线计算及预测精度较高的特点，特别适合于有高实时性要求的场合进行实时预测。解决了基于传统ＡＲＭＡ模型建模繁琐，模型阶次对预测精度影响大，以及神经网络模型收敛速度慢，难于满足实时性要求的问题。仿真及实验结果表明了该算法的可行性和有效性。

REAL-TIME SIGNAL PREDICTION BASED ON CHEBYSHEV NUMERICAL APPROXIMATION AND IT'S APPLICATION

A predicting algorithm based on Chebyshev numerical approximation to predict time series is puts forward.The algorithm possesses the features of model simplicity,small observed sample size needed,easy real-time calculation and high predicting accuracy.It overcomes the shortcomings of ARMA model and neural network model,as the traditional ARMA method needs a complicated model building,the order of the model has a great influence on predicring accuracy and the method based on neural network meets real-time demand difficultly due to it's slow rate of convergence.The simulation and experiment show that the algorithm proposed is feasible and effective.