Study On Path Planning And Trajecotory Tracking Control Of Free-Floating Space Robot

Space robot will play an important role in the future human space activity. The research on space robot has become a research hotspot of aerospace technology in countries of the world. Free-floating space robot is one kind of space robot, whose attitude control and position control of the base do not work, that is, closed. Under such working mode, the fuel and electricity are saved prolonging the on-orbit life span of space robot. Therefore, free-floating space robot has outstanding advantage and wide application prospect compared with other categories. This thesis focused and studied on the problems of dynamics, trajectory planning and trajectory tracking control of free-floating space robot.Firstly, the concept and category of space robot were introduced. Then the current technology for space robot and its ground simulation system in main countries of the world were surveyed. And the theory development of space robotics in home and broad were summarized in depth, including the problems of system modeling, trajectory planning, tracking control.Using dynamics of multi-rigid body theory and"augmented body vector"method, this thesis derived kinematic model and dynamic model of free-floating space robot and an example model of planar two-link free-floating space robot was also given. It is noted that the external disturbance moment, caused by the end-effecter impacting with the object, is imported into free-floating space robot system. Based on this fact, which was ignored by most present controller for space robot, the nature and characteristic of such external disturbance moment was analyzed as well as other system uncertainties.Angular momentum of the free-floating space robot is conserved and there is dynamic couple between the manipulator and the base. Thus, system motion is constrained by nonholonomic constrain because of the non-integrability of angle momentum conservation equation. To solve the constrained motion problem of free-floating space robot using nonholonomic property, a dynamic equation considering nonholonomic constraints was developed and a new path planning method based on coordination transform and optimization method was proposed. Using this method, the attitude of base and position of end-effecter can arrive at desired value at the same time.Since solution obtained by trajectory planning is in kinematic level, it is important to realize the trajectory planned by nonholonomic planning method in dynamic level. The existed methods focused on model uncertainty caused by load change in value, ignoring the effect of the external disturbance which may deteriorate the control performance. In this thesis, two control methods were proposed in joint space. One was robust fuzzy neural network (RFNN) control and the other was the nonlinear robust control The FNN was designed to approximate an ideal controller, and the effect of approximation error was estimated and attenuated by a robust controller. A novel learning method established using Lyapunov function assured the stability of the system. The robust controller, which considered the model uncertainties and external disturbances at the same time, was proposed based on Backstepping method avoiding to solving HJI inequality Simulation results of a two-link planar free-floating space robot verified the validity of the proposed controller.During actual space operation, the desired trajectories are not always given one-by one in joint space, but given to end-effecter in task space (or inertial space).Therefore, it is important to research the problem of tracking control in task space. Since the Jacobian matrix of the system is relative with the dynamic parameters, it is difficult to transform the trajectory in task space to the one in joint space once there is uncertainty in Jacobian matrix. To track the trajectory in task space, considering the model uncertainties and external disturbances, two robust tracking controllers were proposed, i.e. adaptive robust controller and robust neural network (RNN) controller. Using Lyapunov direct method and dissipative theory, the adaptive robust controller can stabilize the system and guarantee the L2 gain from disturbance to tracking error is lower than the given index. For RNN controller, a RBF neural network was used to approximate the highly nonlinear model of free-floating space robot and a robust controller was used to compress the approximation error and external disturbance. New on-line adaptation laws of parameters and weights of RBF neural network was proposed using Lyapunov direct method. The robust controller was proposed based on dissipative theory. Above these assures the stability of the whole system, and L2 gain also can be less than the given index.