A motion planning method for a hyper multi-joint manipulator using genetic algorithm
T. Nishimura, Koh Sugawara, Ikuo Yoshihara, Kôji Abe · 2003
A motion planning method for a hyper multi-joint manipulator whose workspace includes several obstacles is proposed. Because hyper multi-joint manipulators have redundancy and can take various orientations, these manipulators are useful in complicated and narrow workspaces which consist of several obstacles. On the other hand, motion planning for hyper multi-joint manipulators is difficult because these manipulators have high degrees of freedom. Motion planning for a hyper multi-joint manipulator is divided into two parts, "path planning" for the tip of manipulator and "collision-free sequence generation" for the whole of the manipulator. We use "collision-free sequence generation" to optimization of the sequence of variations in joint angles. A motion planning method for a hyper multi-joint manipulator combined "path planning" which adopts the artificial potential field concept and "collision-free sequence generation" which uses GA coding of the sequence of variations in joint angles as a 2 dimensional variable-length gene is proposed. Then, a motion plan for a hyper multi-joint manipulator which can lead the tip of the manipulator to its goal without any collisions with obstacles is obtained. The effectiveness of the proposed method is evaluated through computer simulations of motion planning for a 12-joint manipulator.