基于快速扩展随机树的7R机械臂避障达点运动规划

Motion Planning of Reaching Point Movements for 7R Robotic Manipulators in Obstacle Environment Based on Rapidly-exploring Random Tree Algorithm

基于单树随机树搜索算法(Single directional rapidly-exploring random tree,single-RRT)和双树随机树搜索算法(Bi-directional rapidly-exploring random tree,bi-RRT),对7R机械臂的避障达点运动规划展开系统研究。基于single-RRT算法进行避障达点运动的数值仿真和实物样机试验。提出一种新的bi-RRT算法,结合末端姿态调整和关节自运动来生成目标树。并利用7R机械臂的解析逆解,生成目标位形。传统的bi-RRT算法只给定了一个目标位形,而新算法中目标点树根是由一群目标位形组成。在给定的障碍物环境中,机器人自动选择某一合适的位形作为目标节点来引导搜索树最有效地生长。通过数值仿真,验证该方法的优越性。利用Matlab和C++的混合编程和OpenGL开发了7R机械臂避障达点运动规划仿真软件。利用该软件,对基于bi-RRT的7R机械臂避障达点运动规划进行虚拟样机的试验研究。

Based on single directional rapidly-exploring random tree（single-RRT） and bi-directional rapidly-exploring random tree（bi-RRT）,the motion planning of reaching point movements for 7R robotic manipulators in obstacle environment is systematically studied.Numerical simulations and physical experiments are conducted by using the single-RRT algorithm.A new version of bi-RRT algorithm is proposed by incorporating the pose adjustments and the self-motion in joint space into the expansion of target trees.A general analytical inverse kinematic algorithm for 7R robotic manipulators is used to get desired target configurations.In this way,the given target configuration in the standard bi-RRT is replaced by a group of configurations in the new algorithm.In a given obstacle environment,the robot can automatically choose one of the appropriate configurations as the target node to direct the expansion of searching tree in the most efficient manner.Through numerical simulations,the superiority of this algorithm is further verified.By using Matlab,VC＋＋ and OpenGL,a simulation software is developed to further conduct experimental researches on the motion planning of reaching point movements for 7R robotic manipulators.