虚拟参考反馈校正控制器参数的预测误差辨识  

Prediction Error Identification of Controller Parameters in Virtual Reference Feedback Tuning

作　　者：唐小军[1] 王道波[1] 唐得志[1] 李猛[1] 

机构地区：[1]南京航空航天大学自动化学院,江苏南京210016

出　　处：《华南理工大学学报（自然科学版）》2013年第5期15-21,共7页Journal of South China University of Technology(Natural Science Edition)

基　　金：航空科学基金资助项目(20101352015)

摘　　要：针对未知系统中双控制器的设计问题,采用虚拟参考反馈校正方法,通过最小化由一组观测数据组成的L2范数所构成的代价函数,达到直接设计控制器目的.联合前馈控制器分离出的固定部分和反馈控制器作为滤波器,对原输入/输出关系式进行重参数化,得到标准的预测误差辨识形式,在所得式中,对控制器未知参数矢量采用可分离迭代的非线性最小二乘法进行估算.文中还对算法的收敛性进行了理论分析,并通过仿真验证了所提方法的有效性.In this paper, a method that minimizes the 2-norm-type cost function consisted of a set of observed data, namely the virtual reference feedback tuning, is employed to directly design a dual-controller for an unknown system. In this method, the fixed part decomposed from the feedforward controller is combined with the feedback controller to form a filter, a normal formula of the prediction error identification is derived by re-parameterizing the original input/output relationships, and a separable iterative nonlinear least square method is used to estimate the unknown parameter vectors of the controller. Moreover, the convergence of the proposed algorithm is theoretically analyzed, and the effectiveness of the algorithm is finally verified by simulations.

关 键 词：虚拟参考反馈校正 预测误差辨识 滤波器 重参数化 非线性最小二乘法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]