一类非齐次时变线性模型及自适应预测控制策略

A Class of Inhomogeneous Time-varying Linear Model and the Adaptive Predictive Control Strategy

通过分析系统闭环预测存在的缺陷,为了改善系统的预测性能、便于系统辨识和实现自适应控制,提出了一类适合于一般非线性系统的非齐次时变线性模型,给出并证明了其存在性定理。而后给出了这类非齐次时变线性模型的渐消记忆递推最小二乘参数估计,推出了基于该类模型的自适应预测控制算法。仿真结果表明基于该类模型的自适应预测控制策略具有优良的控制品质。

By analyzing the defect of closed loop prediction of systems, and in order to improve the prediction behavior, to be convenient for system identification and to realize adaptive control, a class of inhomogeneous time-varying linear model is proposed, which is suitable for general nonlinear systems. The existence theorem of the model is given and demonstrated. Subsequently the recursive least squares with variable forgetting factor algorithm to estimate the model parameters is given, and the adaptive predictive control algorithm based on the model is deduced. Simulation result shows that the adaptive predictive control strategy based on the inhomogeneous time-varying linear model has excellent control quality.