Path Planning Of Space Robot In Capturing Moving Target Autonomously

With the development of the space on-orbit service technology, the content andscope of on-orbit service of space robot has been expanded broad gradually. The spacemissions like repairing failure spacecraft on-orbit, recycling space lapse satellite andcleaning up space junk is more and more. The autonomous capture technology of thetarget spacecraft plays a key role in these missions. So, the autonomous capture offailure spacecraft is studied as the main line in this paper. The motion prediction methodof spacecraft is launched, and based on spacecraft motion predicted information,autonomous capture planning is put forward. At last, some mathematical simulation andphysics experiments is operated to verify the method.For the target in autonomous capture, structural features of general spacecrafts areanalyzed firstly. Then basing on space rigid body dynamics equations, motivating law offailure spacecraft is researched, the simplified model of spacecraft is constructed, andalso the system state equations and measurement equations for motion prediction.Thinking of the nonlinear characteristics of the state equation, Extended Kalman filterand Unscented Kalman filter is introduced for predicting the motion of spacecraft.Although the extended Kalman filter can be used in nonlinear systems, the predictionresults may converge for some strong nonlinear system. In using Unscented Kalmanfilter, the higher-order terms in model signals will not be neglected, so higher accuracyof predicted results for strong nonlinear systems can be achieved, also shortcomings ofdivergence predictive results and lower accuracy for extended Kalman filter isovercame.In the process of autonomous capture of spacecraft, path planning of space robotsand capture planning is included. In this paper, the continuous acceleration pathplanning method in joint space is proposed. By constraining the size and variation ofacceleration in the path points, force mutation is eliminated in controlling robot, alsosmooth path is achieved. In the process of a visual servo control robot autonomoustarget acquisition, time-delay must be faced directly. By adding predicted information toautonomous capture of spacecraft, the shortcoming of time-delay can be reducedeffectively. By using the predicted motion information, a autonomous capture planningmethod based on predicted pose of spacecraft is proposed in this paper. By this method,autonomous capture of spacecraft can be managed, also a higher capture accuracy andreduced time-delay.In order to verify if the motion predict method and autonomous capture planningmethod based on predicted pose of spacecraft is correct and feasible, somemathematical simulations and physics experiments is be practiced by mapping the spacemotion to the ground at last. Ground test system is composed by digital simulation platform and physical platform. By ground test system, the movement of the robot andtarget spacecraft can not only be seen in three-dimensional environment and but also inreal robot environment.The projects researched in this paper keep up closely with the step of developingstrongly in space on-orbit service technology. The research results takes a certain theoryand application significance for the development of on-orbit service technology in thefuture.