Abstract: Aiming at the problem of low efficiency and unreal effect of motion planning for virtual human assembly simulation in narrow space,an algorithm by combining kinematics with dynamics was proposed to solve and optimize the moving path of virtual human arm.Seven Degree of Freedom (DOF) arm motion planning was divided into two steps,which were the collision-free virtual hand path planning solved by bi-directional Rapidly-exploring Random Tree (bi-RRT) and the entire virtual arm motion planning solved by Inverse Kinematics (IK).To acquire natural motion path,a multi rigid dynamics model was established on the arm chain,with which the arm motion path was optimized according to a maximal comprehensive ergonomics comfort index.The proposed methods were implemented on a radar assembly simulation task,and the results showed that the generating efficiency of arm operation in narrow space was improved,and the ergonomics comfort was enhanced.

Key words: narrow space, virtual human arm, motion planning, rapidly-exploring random tree, dynamics optimization

摘要： 针对虚拟人在狭小空间内进行装配拆卸操作仿真时手臂运动路径规划求解难、姿态不自然的问题,提出一种运动学与动力学相结合的分层求解优化方法。将七个自由度的手臂无碰撞运动路径规划问题分成虚拟手和手臂两个层级,即虚拟手无碰撞运动路径规划采用bi-RRT(双向快速扩展随机树)算法;手臂运动路径规划采用基于肘圆的逆向运动学的分段规划方法。基于多刚体系统动力学模型,以最大关节舒适度为指标对手臂运动路径进行寻优。雷达天线转台虚拟人装配操作仿真实例表明,上述方法可以提高狭小空间的手臂操作路径生成效率,并可提高手臂关节的舒适度。

关键词: 狭小空间, 虚拟人手臂, 运动规划, 快速扩展随机树, 动力学优化