摘要

A novel fault tolerant attitude tracking control scheme is developed for flexible spacecraft with partial loss of actuator effectiveness fault. Neural networks are first introduced to account for system uncertainties, and an adaptive sliding mode controller is derived by using on-line updating law to estimate the bound of actuator fault such that any information of the fault is not required. To further address actuator saturation problem, a modified fault tolerant control law is then presented to ensure that the resulting control signal will never incur saturation. It is shown that the roll, pitch and yaw angle trajectories can globally asymptotically track the desired attitude in the face of faulty actuator, system uncertainties, external disturbances and even actuator saturation. A simulation example of a flexible spacecraft is given to illustrate the effectiveness of the proposed controller.