论文部分内容阅读
对逆系统建模时,原系统的输出作为逆系统参数辨识时的输入.由于原系统输出存在测量噪声,且噪声方差未知,采用普通最小二乘法辨识,无法得到逆系统参数的一致无偏估计.为此,本文研究了一种有输入扰动的的逆系统无偏参数辨识算法,该算法先通过小波变换估计输入信号噪声的方差,再由估计得到的方差,通过偏差消除的递推最小二乘法,对逆系统的参数进行无偏辨识.该算法降低了对输入辨识信号为白噪声的要求,具有较强的实用性.由于采用递推运算,该算法也可以用于逆系统参数的在线辨识.最后,通过实验验证了该算法的有效性.
When modeling the inverse system, the output of the original system is used as the input of the inverse system parameter identification. Since the output noise of the original system exists and the noise variance is unknown, the unbiased estimate of the inverse system parameter can not be obtained by ordinary least square method. To solve this problem, this paper studies an unbiased parameter identification algorithm for inverse system with input perturbation, which first estimates the variance of the input signal noise by wavelet transform, and then estimates the variance by the variance of the recursive least square Multiplication and unbiased identification of inverse system parameters.The algorithm reduces the requirement of white noise for the input identification signal and has strong practicability.Because of the recursion algorithm, the algorithm can also be used to inverse the online parameters Finally, the effectiveness of this algorithm is verified through experiments.