前馈反馈线性时变多变量控制系统性能评价

本文研究一类具有可确定“时滞”的线性时变多变量(linear time-varying multiinput-multioutput, LTV MIMO)控制系统的性能评价方法。定义了该类系统中相当于单回路“过程滞后”的矩阵因子——关联矩阵，并利用该关联矩阵和正常乘法提出一套建立线性时变多变量控制回路前馈反馈最小方差(minimum variance, MV)基准的系统化算法和过程．需要考虑到LTV系统特有的计算特性，利用实际最小方差输出而不是关联滤波最小方差输出获得MV基准。理论和仿真研究表明，所提方法能准确和有效地评价线性时变多变量过程．

Performance Assessment of Feedforward and Feedback Linear Time-varying Multivariable Control System

This paper presents a performance assessment method for a kind of linear time-varying multiinput-multioutput(LTV MIMO) control systems of which the time delays can be decided.A matrix factor named interactor matrix,equivalent to process delay in univariate loops,is defined.Using the interactor matrix and normal multiplication,a set of systematic algorithm and process to set up feedforword plus feedback minimum variance(MV) benchmarking is proposed for LTV MIMO control loops.In order to acquire MV benchmarking,the computing characteristics of LTV systems should be taken into consideration,and the actual MV output instead of the interactor filter MV output should be used.Theoretical research and simulation show that the presented method can assess LTV MIMO process accurately and effectively.