Robust Adaptive Coordinative Control For Near Space Vehicle Based On Multiple Models Switching

Near space vehicle (NSV) has great strategic value both in military and civilian area, whichcombines the advantages of traditional aircrafts and spacecrafts. It has become a focus of concern andresearch for many military powers. NSV has special characteristics, such as special flightenvironments, large flight envelop, multiple work modes and multiple flight states. These make NSVpossess some very complicated features. Therefore, the NSV control system design is a challengingand meaningful research project. The dissertation carries out a series of research work in NSVmodeling and robust adaptive flight control for nonlinear uncertain systems.First of all, on the basis of the published literatures and the contributions of our research group,the flight mechanics and dynamics of the NSV with variable sweep angle are systematically analyzed,and a set of equations are established to describe the flight motion in the earth coordinate system. Theopen-loop control dynamics with disturbances and uncertainties are studied. They demonstrate thatthe developed model can embody the characteristics of NSV, such as serious nonlinearity, intensestate-coupling, fast time variation and uncertainties and can meet the requirements of flight controlsystem design.Secondly, a robust soft-switching control method based on multiple fuzzy models is presentedfor the nonlinear flight motion equations. The work space of NSV is divided into multiple fuzzyregions according to the sweep angle. Then local robust controllers are designed by the descriptorsystem method, which can reduce the calculation and complexity. The soft switching is realizedthrough combining local controllers with fuzzy weights. Simulations of NSV with variable sweepwings demonstrate the effectiveness of the proposed method.Thirdly, to decrease the scale of the model set, a robust adaptive control strategy based on theswitched multiple nonlinear systems is considered in the presence of unknown aerodynamicparameters and external disturbances. Fuzzy systems are employed to approximate the unknownfunctions, and robust controllers are designed to compensate for the approximation errors. Then thestability of the closed-loop system is proved according to the common Lyapunov function theory.Furthermore, the variable universe method is introduced to get better approach results. The fuzzy rulescan be increased without more complexity. The simulation results show that the two control methodscan realize robust adaptive control on NSV.In the following, a multi-model switching controller based on multi-model fuzzy observer isinvestigated in consideration of the disturbances in the environment and the unmodeled uncertainties in the NSV flight. The multi-model fuzzy observers are constructed by using the known systeminformation to approximate the unknown functions with dynamic lumped disturbances. Then robustcontrollers are also designed. The adaptive laws of the fuzzy weights and robust gains are derived bythe common Lyapunov function theory and the closed-loop system can be ultimately uniformlybounded. Further more, a fully tuned fuzzy observer is introduced to avoid overly relying on expertiseand improve the fuzzy system performance. The centers, widths and fuzzy weights of the fuzzysystem can be adjusted on line. And a multi-model switching control method based on fully tunedfuzzy observer is proposed. At last, the two schemes are used to track the attitudes for the NSV andreach ideal control effects.Finally, considering the actual flight requirements, a multi-model switching control law withconstrained-input and constrained-output is presented. The constrained-output part is constructedaccording to the given error range. Then the command filter is applied to the designed controller to getactual input. The switched fuzzy system is introduced to approximate the difference between thedesigned and actual controller as well as lumped disturbances. The estimates can be used tocompensate for the controller. The simulation results show that the designed method can realizehigh-precision tracking control with appropriate input and output.