摘要
利用随机过程理论研究了时变多变量系统多新息投影辨识算法的均方收敛性，给出了参数估计误差上界的计算公式，并阐述了获得最小均方参数估计误差上界时，数据窗长度的选择方法．本文的研究成果对提高辨识算法的实际应用效果有重要意义．
Mean Square convergence of the multi-innovation projection identification algorithm for time- varying multivariable systems is proved by using stochastic process theory,and the formulae of computing the upper bounds of the parameter estimation error given by the niulti-innovation projection identification algorithm are derived. The way of choosing the data window length is stated for the minimum upper bound of the mean square parameter estimation error to be obtained.
基金
国家自然科学基金项目(60074029)
国家自然科学基金重点项目(69934010)
