Control For Stochastic Markovian Jumping Systems With Partly Unknown Transition Probabilities And Saturating Actuator

As a class of stochastic hybrid systems,Markovian jumping systems have been an attractive subject of research during the last decades,since they can model the behavior of many practical processes and systems,such as networked control systems.It is noticeable that most of the analysis and synthesis about Markovian jumping systems have been obtained under the assumption of a completely known of transition probabilities.However,owing to many kinds of complicated factors,the exact values of transition probabilities can be accessible partly.So,in this article,we study the Markovian system with partly known transition probabilities.On the other hand,because of the limitation of physical conditions,actuator saturation become a very common kind of nonlinear.This can severely degrade the performance of closed-loop systems and sometimes even make a stable closed-loop system unstable.The systems with partly known transition probabilities and actuator saturation are very common in many practical conditions.But the studies in this aspect are limited.And there is only a few results using anti-windup method.There is further room for investigation.This thesis take the Markovian jumping system in the presence of additive noise,standard Wiener noise and internal parameter uncertainties as model,finished following studies:On one hand,we use direct approach to study this kind of stochastic Markovian jumping system considering the problem of stochastic stability analysis,H? performance analysis and robust H? performance analysis.Through using free connection weight matrix,we decrease the conservation of the system.Finally a non-fragile state feedback is designed to ensure the closed-loop system with H? performance robustly stochastically stable and the domain of attraction of system is maximized.On the other hand,by constructing appropriate Lyapunov function using anti-windup approach,finish the stochastic stability analysis,H? performance analysis and robust H?performance analysis.Anti-windup compenstors are designed to ensure the the closed-loop system with H? performance robustly stochastically stable and the domain of attraction of system is maximized.