Research On Robust Model Predictive Control Based On LPV Model For Hypersonic Vehicles

Due to its astonishing speed and rarely explored flight altitude, hypersonic vehicle have attracted so much attention. Stable control algorithm design plays a vital role in the whole hypersonic flying project. To track the reference signals of wide ranges, suppress the effects of disturbances and varying parameters, the method of robust model predictive control based on LPV model is taken advantage of to design the hypersonic vehicle control system.To achieve stable control and restrain the disturbance, the polytopic LPV model of the hypersonic vehicle is established, and robust MPC controller is designed. Tensor- product modeling method is used to get the suitable polytopic LPV model. The tensor which is gotten by sampling the Jacobian linearization matrix of the hypersonic vehicle is dealt with high order singular value decomposition. With the help of the norm-bounding technique, robust MPC forces the predicted states of the hypersonic vehicle stay in an invariant ellipsoid. The invariant ellipsoid is transformed into an equivalent LMI constraint. The LMI solver is used to solve the optimization problem offline.The optimality of offline calculated robust model predictive control method is not that good. In order to improve the optimality, dynamic programming and a kind of various control horizon robust MPC are explored. Imitating the method of dynamic programming, a series of nested invariant ellipsoids are established in the neighborhoods of the hypersonic vehicle system stable point from the inside out. The multiple ellipsoids make the control horizon of MPC various. And the various control horizon is equivalent to increasing the freedom degrees of the optimizing calculation. The proposed method not only enlarges the convergence domain of the hypersonic vehicle control system, but also improves the optimality of the hypersonic vehicle control system.In order to solve the problem of tracking tremendous large flight envelope and time varying parameters, feedback linearization and optimization function are combined to calculate the large envelope LPV model. And ISS predictive control law is constructed. The varying parameters in the original nonlinear model of the hypersonic vehicle are transformed into the multiple vertices and bounded disturbances. MPC utilizes the LMI to find an inequation about the local quadratic form ISS lyapunov function and the disturbances. This inequation ensures the input-to-state stability of the large envelope LPV system. The states of the large envelope LPV model are bounded because of the boundness of the disturbances. Ultimately, the hypersonic vehicle predictive control system could track the tremendous large flight envelope under the condition of varying parameters.