Collision avoidance for a redundant manipulator in an unstructured environment with a simplified path planning and multiple self-organizing maps

This paper proposed a collision avoidance approach for a visuo-motor system, which composed of a redundant manipulator, cameras and self-organizing maps, in an unstructured and cluttered environment. The achievement of collision avoidance is basing on a simplified path planning and a motion control by the self-organizing maps. The reason why we say it is a simplified path planning is, in our research, that we only planed a path for the end-effector of the redundant manipulator in the collision avoidance. As the path of the end-effector is collision free from obstacles, the learned self-organizing maps promise to decide the joint angles of the manipulator with obstacle free poses. In order to settle visibility constraints, 3 CCD cameras are employed. Two self-organizing maps are directly connected to the cameras. According to the visibility of targets, more appropriate map is adopted to control manipulator. In our visuo-motor system, neither any priori knowledge about the manipulator nor the camera parameters is required. Simulation results are presented to demonstrate the effectiveness of the proposed method