Robust Output Feedback Predictive Control For Markov Jump Systems

Markov jump system is a suitable mathematical model to represent a class of dynamicsystems subject to random abrupt variations in their structures, and this system exists invarious fields. Markov jump system can be regarded as a special kind of stochastic systemswith system matrices changed randomly at discrete time instances governed by a Markovprocess, and remain linear time invariant between the random jumps. Model predictivecontrol is the most efficient control strategy in handling constraints and optimization problem.So model predictive control has received much attention as a powerful tool for the control ofprocess systems including Markov jump systems.Due to the reason that the feedback control method can delivery the real information ofcontrolled object to the system timely, the controller designed in this article is studied by thefeedback control method. Considering the state cannot be measured directly, therefore, thestudy of control problem based on output feedback is more practical valuable. The researchcontents mainly relate to robust predictive control via output feedback controller design. Thecontroller design approaches are based on feedback control theory and all results can bereduced to a feasible problem of linear matrix inequalities (LMIs). The research works of thisdissertation are divided into five chapters, and the main contents are as follows:An output feedback predictive controller design method is proposed for a class ofparameters uncertain Markov jump systems. Based on the idea of variable transformation andLMI methods, the infinite time domain “min-max” optimization problems are converted intolinear programming problems, output feedback control law is obtained. The controllerguarantees the closed-loop system robust mean square stable.For quasi-linear parameter varying Markov jump system, when the state can not bemeasured completely, a manner of output feedback robust model predictive control isproposed with polytopic uncertainty and bounded disturbance. Through solving the min-maxoptimization problem with infinite horizon performance cost, the output feedback law of closed-loop system is obtained. With the circumstance of input and output, the stochasticstability of the closed-loop system is guaranteed by utilizing the quadratic boundedness.