Decentralized Fault-tolerant Control Of Rigid-flexible Hybrid Space Robot

With the rapid development of China's aerospace industry,space robot will play an important role in the future.Space robot can be used for on-orbit service and space station construction.Due to the very bad environment and aging parts of space robot,actuator or sensor of space robot may go wrong in the process of space operation,so the research on fault diagnosis and fault-tolerant control is of great practical significance.At the same time,in order to reduce the consumption of precious fuel,the base's position is uncontrolled.Taking into account the complexity of the task and the stability of operation,the research and design of dual-arm space robot are also necessary.Furthermore,as the result of lightweight design,the flexible issues of space robot need to be taken into consideration.Based on the above background,this paper makes related research on decentralized fault-tolerant control of rigid-flexible hybrid space robot.The dynamic models of rigid single-arm,dual-arm space robot,flexible-arm space robot,and flexible-joint space robot are successfully established by using the second Lagrange equation,reasonable deformation description of flexible body,and linear spring model,and considering momentum and moment of momentum conservation law.The dynamic models lay the foundation for subsequent control scheme design.Based on decentralized back-stepping terminal sliding mode control,a method of decentralized fault-tolerant control for rigid single-arm space robot under partial loss of actuator is proposed.The system is decentralized into several subsystems.With the help of radial basis function neural networks,uncertainty and interconnection of subsystems are estimated.Fault-tolerant controllers are designed by back-stepping to achieve individual control of subsystems.For actuator's partial loss of dual-arm space robot with external disturbance,a fault-tolerant control method which combines decentralized back-stepping neural networks with delay control(TDC)is proposed.The external disturbance and actuator failure are regarded as unknown dynamics.When fault occurs,with the TDC,the unknown actuator faults are estimated by using one-step previous state information.Most of all,it doesn't need fault detection and isolation.In order to solve actuator's partial loss and flexible vibration suppression offlexible-arm space robot,based on singular perturbation theory,a composite control scheme of decentralized terminal sliding mode and PD feedback control is designed to realize fault tolerant control,precise trajectory tracking and flexible vibration suppression.As to actuator's partial failure and flexible vibration of flexible-joint space robot,with the help of the joint flexibility compensation concept and singular perturbation theory,a method which is composed of decentralized back-stepping neural network fault control and speed difference feedback control is proposed so that to achieve exact trajectory tracking and flexible vibration suppression.Finally,the proposed control methods are successfully applied to the numerical simulation of space robot,and the simulation results show the effectiveness of the proposed control methods.