Abstract: To solve the pose planning problem of virtual human maintenance operation in a virtual maintenance simulation environment,a virtual human upper limb operation method based on an improved Rapidly-exploring Random Trees(RRT)-Connect algorithm was proposed.In this solution,the foraging behavior rule of Artificial Fish  Swarm Algorithm(AFSA) was introduced into the traditional RT-Connect algorithm as a guiding strategy,so that the expansion of the initial point and the target point in the path planning task was oriented.The path point was used as the motion track point of the virtual human upper limb terminal joint.To obtain the virtual human maintenance attitude,inverse kinematics were used to derive the attitude set in accordance with the degree of freedom of the human joint in each path point.In addition,to select the optimal path of pose concentration,the fast upper limb evaluation method was used to obtain the optimal pose of virtual human upper limb maintenance.The feasibility of this method in virtual maintenance applications was verified by taking a virtual maintenance operation simulation of the unmanned boat as an example.

Key words: virtual human upper limb, motion path planning, rapidly-exploring random trees-connect algorithm , human-machine ergonomics, artificial fish swam algorithm

摘要： 为了完成虚拟维修仿真环境下的虚拟人维修作业姿态规划问题,提出一种基于改进RRT-Connect算法的虚拟人上肢作业姿态求解方式。在该求解方式中,将人工鱼群算法(AFSA)的觅食行为法则作为导向性策略引入传统RRT-Connect算法,从而使得路径规划任务中初始点与目标点的扩展具有导向性;将路径点作为虚拟人上肢末端关节运动轨迹点。为获得虚拟人维修姿态,利用逆向运动学进行推导,以在每一个路径点中得到符合人体关节自由度的姿态集。另外,为了选择出姿态集中的最优路径,利用快速上肢评估法求取最优的虚拟人上肢维修作业姿态。最后,以无人艇虚拟维修操作仿真为例,验证了该方法在虚拟维修应用中的可行性。

关键词: 虚拟人上肢, 运动路径规划, RRT-Connect算法, 人机工效学, 人工鱼群算法