Optimal Control Base For Satellite Attitude Based On Receding Horizon

In the control of satellite attitude maneuver, we always want the system meets optimal requirement. As the satellite usually has model uncertainty, nonlinearity and disturbance, it is difficult to use the classical optimal control on it. On the other hand, receding horizon control, which is a method of predictive control, can achieve a approximate optimal control. But the control law lacks stability proof and it only achieves a local optimal control. This dissertation combines the classical optimal control with receding horizon control, creates a kind of optimal control which can be used on the control of satellite attitude maneuver. The control law guarantees the global stability as well as achieves the global optimal control under certain condition. A simulation of this method is made. Based on the result of the simulation and the analyse of the algorithm, research on the computation amount is made. Several techniques are proposed to relieve the computational burden.At first we discuss the theory of optimal control, the contact between the value function in optimal control and control Lyapunov function is analyzed. Basis on the theory of inverse optimal, the condition for calculating optimal control from control Lyapunov function is summarized. The algorithm under certain condition is proposed and its applicable range is analyzed. Also we analyze the method of receding horizon control. A control law which combines control Lyapunov function and receding horizon control is created and its optimality is proved.Considering the features of attitude maneuver, rule of selecting control Lyapunov function is proposed based on the model of satellite in practice. A numerical algorithm of receding horizon control is proposed for the nonlinear system. A simulation is made with this control Lyapunov function and receding horizon control combined function.As the receding horizon control brings a heavy burden of computation. We analyze the calculation process in control. Some techniques are proposed to simplified the algorithm. A simulation is made to check the effect of these techniques.