自适应广义预测控制系统的鲁棒稳定性分析

研究当被控对象具有未建模动态、参数时变、非线性以及有界干扰时,自适应广义预测控制系统的鲁棒稳定性。提出了分析和设计鲁棒自适应广义预测控制系统的一般理论和方法。给出了一类满足弱化鲁棒稳定性条件的鲁棒参数自适应算法,并根据确定性等价原理将它和鲁棒广义预测控制律结合,建立了鲁棒间接自适应广义预测控制算法。采用统一的理论分析了已建立的自适应广义预测控制算法的全局稳定性和鲁棒性,得出了自适应控制系统具有鲁棒稳定性的新条件：自适应验后误差和对象参数估值变化的均值渐近小,从而弱化了现有自适应控制系统的鲁棒稳定性条件。

BOBUST STABILITY ANALYSIS OF ADAPTIVE GENERALIZED PREDICTIVECONTROL SYSTEMS

This research is devoted to the robust stability analysis of adaptive generalized predictive control(GPC) systems in the presence of unmodeled dynamics,time-varying parameters,nonlinearities and bounded disturbances. A general approach for designing and the theory for analysing robust adaptive GPC systems are presented. The design approach involves the development of a robust adaptive law which has less restrictive properties(the robust stability properties A5 A7)and the use of certainty equivalence principle to combine it with robust GPC structures,The global stability properties and robustness of the developed adaptive control schemes are established by using a general theory, which requires the properties of signals to be asymptotically small only in the mean sense over intervals of time and provides weak sufficient conditions that guarantee the robust stability of the adaptive control systems.