一类随机连续信号的时间序列分析与建模-控制与决策.PDF

第 15 卷第4 期 2000 年7 月 Vol. 15 No. 4 CONTROL A NDDECISION 　July 2000 * 赵 明旺 (430081) 讨论由随机微分方程描述的随机连续信号的辨识建模问题。提出并证明了非平 的连续 Wiener 过程通过 定的连续线性系统后为平 随机过程, 且均值和自相关函数阵为时间遍历的。基于 状态空间分析, 给出了连续随机信号建模的时间序列分析方法, 并证明了参数估计的一致收敛性。仿真 结果显示了所提出方法的有效性。 随机连续信号, 时间序列分析, 建模, 系统辨识,Wiener 过程 211.61 O Modeling and Time-series Analysis for a Class of Stochastic Continuous Signals Zhao Mingwang ( Wuhan U niversity of Science and echnology) Abstract he modeling problem for stochastic continuous signals, described by st ochastic differential , . - equations is discussed It is proved that t he stochastic process produced by a non stationary cont inuous - Wiener process and a st able linear filt er is stat ionary and its expectation and self correlation function are ergodic. Based on the state-space analysis, the time-series analysis method for identification of t he stochast ic continuous signals, proved as consistent convergence, is given. Simulation results show t he effectiveness of the proposed method. , - , , , Key words stochastic continuous signal time series analysis modeling syst em identification Wiener process [ 1, 2] 1 引　　言 1) ( CLS) , ( SA) ; [ 3, 4] CLS