Adaptive control for satellite formation flying with thruster uncertainties

Abstract

This thesis is mainly concerned with the control problem of relative position tracking for satellite formation flying missions in the presence of uncertainties of unknown parameters and external disturbance. Adaptive controllers are designed by using backstepping technique and composite adaptive scheme under uncertainties including satellite mass and thrust misalignment. The practical implementation of satellite formation flying missions depends on accurate control of the follower position relative to the leader, and requires low thrust and long lifetime for the propulsion system. For an ion thruster capable of high control accuracy and long lifetime, its thrust error such as magnitude error and misalignment is not negligible in precise relative control problems. In this study, thrust misalignment is modeled to have property of parameter linearization required in adaptive technique of nonlinear systems. The method to compute control input is addressed because it cannot be calculated explicitly from adaptive control law when thrust misalignment is regarded as unknown parameter and estimated from adaptation law. Adaptive backstepping controllers are designed for two cases considering satellite mass and thrust misalignment as uncertainty, respectively. The proposed controllers are proven to guarantee system stability by using Lyapunov-based analysis framework. Compared to a general controller considering thrust misalignment as disturbance, the simulation results show that the proposed adaptive backstepping controller is more efficient in terms of fuel consumption during formation maneuver under the uncertainty of thrust misalignment. Composite adaptive controller is proposed by using information of tracking error and prediction error in the presence of uncertainties of both satellite mass and thrust misalignment. It is also proven by using Lyapunov stability analysis framework that the proposed controller guarantees global asymptotic convergence for the p...