Multi-Arm Path Generation Method for Humanoid Robots

The robot´s arm used to be limited to manipulation by the industrial manipulator, and most studies on the humanoid robot focus on robots´ walking methodologies. Since the multi-DOF (degree of freedom) system, like the humanoid robot´s arm, has a high- dimension configuration space that cannot appear geometrically and that has many kinematics and dynamics constraints, the path planning of the system requires much computation and, as such, is very complicated. In this paper, the path planning algorithm was employed for the humanoid robot´s (Maru-1) arms. The RRT (rapidly exploring random tree) algorithm, which has been proven to be an efficient approach to multi-arm motion plans, was selected for use in this study. The employed algorithm, with inverse-kinematics analysis, was evaluated in this study using 3D graphic simulation.