Research On Design And Control Method Of Truss-shaped Manipulators For Grasping Of Space Targets

With the development of space science and technology,the research of outer space in the world is also deepening.There are many invalid spacecrafts and some valuable asteroids and meteorites in the present space.For these non-cooperative targets,they need to be operated on orbit,such as capture recovery or dragging into the grave orbits.Therefore,how to capture the noncooperative targets in space is an important challenge in the field of space robotics.In the process of capturing the noncooperative target of the manipulator,it is necessary to ensure that the manipulator can grasp the target and avoid the damage of the manipulator.It is necessary to consider the truss type structure with high stiffness and high reliability,and the control algorithm of the manipulator is also a very important research content.In this paper,a kind of truss manipulator is designed,several control algorithms for space capture tasks are designed,and some intelligent control algorithms are studied in detail.As follows:A truss-type manipulator was designed with reference to the planar 2R mechanism and parallelogram mechanism for capturing non-cooperative targets.Considering the complexity of its dynamic model,and in order to realize its safety control in the capture process,considering the sliding mode control in modern control theory and PID control which is widely used in traditional industry,an approximation of sliding mode control algorithm is proposed.This control algorithm can ensure that the manipulator has accurate and rapid tracking ability during free operation and can produce over damped recovery motion after the actuator is saturated.On the basis of this control,the impedance control algorithm of position type is further studied,acceleration feedforward compensation and high-dimensional sliding mode surface were added to the position controller to improve the control accuracy.Finally,the simulation analysis is carried out to verify the effectiveness of the algorithm.In order to estimate and compensate the dynamic modeling error of the truss manipulator,the radial basis function(RBF)neural network control and disturbance observer are deeply studied,and the complexity of the dynamic modeling of the truss type manipulator is considered.The approximation ability of the nonlinear function is used to solve the problem of the dynamic modeling error of the truss type manipulator by the RBF neural network,and the uncertainty in the parameters of the dynamic model is estimated.The disturbance observer is used to deal with the external disturbance of the control system and the estimation error of the RBF neural network.Combining these two methods,a position controller is proposed,and the stability of the controller is proved by constructing Lyapunov function.In addition,considering the control problem of the contact force between the manipulator and the capture target in the process of space capture,a kind of impedance controller with high power control ability is designed with the advantages of RBF neural network and disturbance observer,and a simulation system is constructed in MATLAB,and the simulation analysis of the proposed method is carried out.According to the truss manipulator designed for space non-cooperative target arrest,and for the experimental verification problem of the control algorithm,the manufacture of the engineering experiment prototype and the construction of its control system were completed.The proposed algorithm was implemented on the experimental platform.