Robust Adaptive Control For Nearspace Vehicles Based On Backstepping Approach

The current research in developing next generation flying vehicles is focused on nearspace vehicles (NSVs) which have very important martial values. The control systems of the ASVs pose several challenges due to their multi-mission profiles, large attitude maneuvers and complicated flight conditions. In this dissertation, two relative problems, i.e. modeling and perturbed uncertain nonlinear control system design, are studied. The main results in this paper are as follows:1. Based on the research contributions of our lab and available material, the simulation model of a NSV is presented, which includes the kinetic equations and motion equations. Aerodynamic force and moment coefficients are given as functions of angle of attack, Mach number and control surface deflections. The propulsion system is a hypersonic air-breathing engine. Open-loop dynamics and stability characteristics demonstrate that the proposed model can be used to allow research, refinement and evaluation of advanced guidance and control methods.2. Longitudinal control for NSV is researched. Firstly, the nonaffine nonlinear NSV model of longitudinal motion is transformed into the affine nonlinear system by using the input-output feedback linearization approach. Then the transformed model is further transformed into strict feedback multi-input/multi-output (MIMO) nonlinear system. Control system of longitudinal motion is designed by backstepping approach, and the simulation results of longitudinal motion demonstrate the effect of control approach.With the assumption that NSV suffers the violent changes of aerodynamic paameters and the outside disturbance in hypersonic condition, we present the integrator backstepping approach based on fully tuned radial basis function neural network (FTRBFNN). The strict proof of the approach's stability is provided simutanously. FTRBFNN has excellent ability of restaining disturbance, and the integrator term in backstepping approach eliminate the static traking error efficiently. As a result, the controller has high control precision. Finally, the simulation results of longitudinal motion show that the control system has good robustness with greatly disturbance.3. We present an approach of adaptive dynamic surface control with variable gain. Firstly, the dynamic surface control is introduced in the backstepping approach to simplify the controller design. Due to the chattering problem appears in the initial stage of the adaptive dynamic surface approach, the learing rate of radial basis function neural networks are regulated dynamically on line, which eliminates the chatting problem in the initial stage of the adaptive process. It has been proved that the tracking error converge to an arbitrary small compact set in finite time by Lyapunov stability theorem. Finally, the simulatin results of longitudinal motion show that the presented approach can reduce the complexcity of control law while still retain good performance of transient process.4. We present a fast adaptive backstepping approach based on fuzzy systems. The total disturbance is identified on line by fuzzy systems, and the control law is deduced by backstepping approach. The adaptive parameters on line can be reduced to the number of subsystems. As a result, the burden computation is alleviated greatly. The stability of the closed loop system is proved by Lyapunov stabiliy theorem and small gain approach, and the tracking error converges to an arbitrary small compact set exponentially. Finally, in hypersonic condition, the coordinated-turned maneuver with 6 degree of freedom of NSV is used in experiment. The simulation results show tht the approach still has good tracking performance with simplified control law and adaptive law.5. Combining with the technology of disturbance observer, this paper presents an adaptive backstepping approach based on fuzzy disturbance observer. The apprach utilize the useful information of the controlled plant and the adaptive law of parameters can be regulated according to the integrated error composed of the tracking error and the error of disturbance observer. As a result, the method realized the accurate approximation and control, so the control performance is improved.A new fuzzy neural network disturbance observer (FNNDO) is presented in this paper, which can further improve the ability of retaining disturbance, and a control algorithem of bakcsteppping approach based on FNNDO is presented. The simulation results show that the FNNDO can regulate the fuzzy rules on line, so it can achieve a higher tracking precision. Finally, the attitude control system of NSV is designed by this control algorithem. The simulation results show that the the control algorithem has higher convergence speed and higher tracking precision than the backstepping approach based on FDO.