Model Predictive Control For Uncertain Singular Systems With External Disturbances

As a system description form,it is widely known that singular systems has more extensive applicability than the state space system which can also be easily show the process of many practical systems,such as aerospace system,robotic system and economical system,etc.Uncertain singular large scale systems which are coupled by a number of singular subsystems and also have extensive practical background.With the increasing requirements of control accuracy in modern production system,finding the factors that caused system instability or bad performances have attracted wide attention by many scholars,uncertainties and time delays are often have a significant impact on the performance of the system,and even cause the system can not run normally.so the study of singular system with uncertainty is also particularly important.Model predictive control(Model predictive control,MPC)can effectively handle system constraints and parameter uncertainty factors,such as,can make the system have better performance,has become a more extensive application of modern control strategy.In this paper the research on robust model predictive control is constrained for a class of uncertain singular systems and uncertain singular large scale systems with external disturbances and input constraints.The main contents are summarized as follows:For a class of uncertain singular system(USS)with bounded disturbance and input constrain,the problem of how to design the model predictive controller is discussed.By taking into account of singular system with convex polytopic uncertainty,the design of dual model predictive control algorithm is given based on system model,besides,the corresponding feedback controller is designed.Numerical simulation shows that the proposed method can guarantee the system satisfying input to state stability(ISS).The decentralized robust H-infinity and model predictive control problem for singular large-scale systems with polytopic uncertainty in system matrices(USLS)is considered.A decentralized control strategy is proposed,a sufficient condition for the existence of decentralized robust H-infinity and model predictive controller for uncertain singular large-scale systems is presented in terms of the solvability to a certain system of linear matrix inequalities and the constructed Lyapunov function.The infinite horizon optimization problem is transformed into a convex optimizationproblem.Simulation results illustrate the effectiveness of the proposed method.For a class of uncertain singular systems with time-delay,we obtain a simplified bounded real lemma(BRL)by matrix transformation.Compared with the existing conclusions,the new BRL is more simpler.The model predictive control design algorithm is obtained by solve the corresponding LMI constraints problem at each sample time.By the method of MPC rolling optimization,a MPC controller is given to ensure the resulting closed-loop singular system is not only satisfies the input constrain,but also satisfies a prescribed performance index.