【作者】 薛驭风； 石雨豪； 李强； 徐世杰； 霍明英； 齐乃明；

【Author】 XUE Yufeng;SHI Yuhao;LI Qiang;XU Shijie;HUO Mingying;QI Naiming;School of Astronautics,Harbin Institute of Technology;Beijing Aerospace Long March Vehicle Research Institute;

【机构】 哈尔滨工业大学航天学院； 北京航天长征飞行器研究所；

【摘要】 提出了一种利用机器人完成以大型空间太阳能电站为代表的大型空间结构在轨装配的设计方案,并针对机器人的运行轨迹及任务规划问题,提出了一种基于Bezier形函数和蒙特卡洛树搜索算法的在轨装配自主规划方法。首先,设计了空间在轨装配机器人和大型结构系统;之后,引入基于包围盒和网格的碰撞检测算法进行装配障碍检测,生成机器人运动轨迹可行域,在该可行域中,使用Bezier形函数法进行机器人无碰撞运动规划;最后,基于机器人无碰撞运动规划求得的机器人能量损耗为优化条件,引入蒙特卡洛树搜索算法进行在轨装配任务规划,形成装配方案。仿真结果表明,该方案能够快速得到机器人的无碰平滑关节轨迹,提高自主任务规划效率,降低空间在轨装配中的能量损耗。更多还原

【Abstract】 In this paper,a design scheme for using robots to complete the on-orbit assembly of large-scale space structures represented by large-scale space solar power stations is proposed,and an autonomous planning method for on-orbit assembly based on Bezier shape function and Monte Carlo tree search algorithm is proposed for the running trajectory and task planning problems of robots.Firstly,a space-in-orbit assembly robot and a large-scale structural system were designed.Then,a collision detection algorithm based on bounding box and mesh was introduced to detect assembly obstacles,and a feasible domain of robot motion trajectory was generated,in which the Bezier shape function method was used to plan the collision-free motion of the robot.Finally,based on the energy loss of the robot obtained from the collision-free motion planning of the robot,the Monte Carlo tree search algorithm was introduced to plan the on-orbit assembly task and form the assembly scheme.The simulation results show that the proposed scheme can quickly obtain the smooth joint trajectory of the robot without collision,improve the efficiency of autonomous task planning,and reduce the energy loss in space onorbit assembly.更多还原