虚拟人手臂避障抓取运动规划  

Planning Collision-Free Grasping Actions of Virtual Human

作　　者：杨宇盟 聂斌 方红根 黄正东[1] 

机构地区：[1]华中科技大学国家CAD支撑软件工程技术研究中心,武汉430074 [2]上海航天设备制造总厂,上海200245

出　　处：《计算机辅助设计与图形学学报》2014年第8期1362-1373,共12页Journal of Computer-Aided Design & Computer Graphics

基　　金：上海航天科技创新基金(SAST201208);国家自然科学基金(61173115;51075162)

摘　　要：针对强干涉环境下虚拟人手臂无碰撞抓取物体动作规划问题,提出一种逆向运动学与运动规划相结合的求解方法.首先在现有FABRIK逆向运动学算法的基础上提出避障FABRIK算法,用于实现手臂无碰撞抓取姿态设计,作为最终抓取姿态;然后采用Bi-RRT算法,综合考虑7维角度域约束和对非光滑路径的简化问题对手臂进行无碰撞运动规划,产生从给定初始姿态到最终抓取状态之间的中间姿态.仿真实验结果表明,该方法速度快,产生的路径能有效地避开障碍;可用于强干涉环境下的虚拟人动作设计.Collision-free grasping actions of virtual human are difficult to plan especially when the surrounding scene is complex.Firstly,the forward and backward reaching inverse kinematics（FABRIK）algorithm is improved and an obstacle-avoiding FABRIK algorithm is put forward to design a collision-free grasping posture.Then,the bidirectional rapidly-exploring random tree（Bi-RRT）algorithm is adopted to plan the motion of human arm from a given initial posture to the grasping posture,which is composed of a series of intermediate postures.In addition,some strategies are developed to handle the issues of the bound constraints on the arms 7angular parameters and the simplification to the zigzag paths generated with the Bi-RRT algorithm.This paper provides some results of computation experiments from the implementation of the proposed method,which show that the method can be used for the arm motion design of virtual human in the scene with obstacles.

关 键 词：虚拟人 逆向运动学 运动规划 避障 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]