研究离散时间参数不确定的线性随机系统的加权多模型自适应控制(Weighted multiple model adaptive control,WMMAC)问题,采用一种改进的加权算法,在模型输出误差可分的情况下,可以保证其收敛性;然后在加权收敛的前提下,借助虚拟等价系统的概念和方法证明了此类加权多模型自适应控制系统的稳定性和收敛性.本文所采用的分析方法和结论不依赖于局部控制策略和加权算法的具体形式,而只取决于它们的某些属性.最后,基于Matlab对相应的加权多模型自适应控制系统进行了仿真,仿真结果验证了加权算法的收敛性和闭环控制系统的稳定性、收敛性。
This paper is concerned with the weighted multiple model adaptive control(WMMAC) of discrete-time linear stochastic system with uncertain parameters. Firstly, an improved weighting algorithm is adopted with convergence guaranteed under smooth conditions. Then based on virtual equivalent system methodology, the stability of resulting WMMAC systems for both linear time-invariant(LTI) and parameter jump plants is presented. The analysis method is independent of specific local control strategy and specific weighting algorithm. Finally, the theoretical results have been verified by simulation results using Matlab.