Research On Aerodynamic Parameter Identification And LPV Mixing Control For Hypersonic Vehicle

As hypersonic vehicle is characterized by extremely fast and agile, it is seen as arevolution in aviation and aerospace field. However, as hypersonic vehicle has thefeatures of highly complex dynamic model, nonlinearity, heavy coupling anduncertainty, the aerodynamic parameters always change severely. Focusing on theaerodynamic parameter identification and tracking control problem of hypersonicvehicle under above control challenges, four aspects of work are discussed in thisdissertation.In order to solve the problem of building a reliable aerodynamic model under thecondition that the aerodynamic model is highly nonlinear, extremely complex andeasily influenced by stochastic disturbance, an offline identification method base onmaximum likelihood method and interior-point algorithm is developed. Firstly,boundaries of unknown parameters are introduced as prior knowledge to covert theestimation problem into a constrained optimization problem. Then the interior-pointalgorithm is applied to solve this problem, which can greatly simplified thecomputation of the algorithm based on lagrangian multiplier and relaxation factor.As the aerodynamic parameters change fiercely in hypersonic flight condition andonly limited measurement data is available for online estimate, it is very difficult toonline estimate the aerodynamic parameters precisely. Aiming at solving this problem,the offline identification method is extended to an online identification method byusing recursive maximum likelihood method. Also, prior knowledge of unknownparameters is introduced as constrains to enhance the performance of traditionalrecursive maximum likelihood method. By utilizing the barrier function directly, thecomplex computation of the online estimate algorithm using lagrangian multiplier andrelaxation factor is avoided. Therefore, fast online aerodynamic parameteridentification is realized.As for the problem of single controller may not perform well in the presence oftime-varying parametric, a robust LPV mixing controller is designed. Firstly, a LPVmodel is built to determine the variation range of important parameters. Then, themixing strategy is utilized to build a robust LPV mixing controller and the trackingcontrol of velocity and altitude is realized. In order to achieve attitude tracking control in reentry process in which theaerodynamic parameters change in wide range, a robust LPV mixing control systembased on hysteresis switching strategy is developed. Firstly, the whole flight envelopeis divided into several sub flight envelopes according to velocity and altitude.Secondly, robust LPV mixing controllers are designed to achieve tracking control ofattack angle in each sub flight envelope. Lastly, the robust LPV mixing controllers areswitched by the hysteresis switching strategy. Therefore the global LPV controlsystem is got and the attitude tracking control is realized in the whole flight envelop.