A parallel collision-avoidance algorithm for robot manipulators

Throughout the world, researchers are improving robots to enable them to do more complex tasks with less human intervention. One aspect of this research is the complex path-planning problem. Path planning derives collisionfree paths to enable robot manipulators to move from one location to another. Collision detection is straightforward using geometry, and moving a manipulator in an environment without obstacles is also straightforward. But, moving a manipulator safely around obstacles is difficult, the level of difficulty depending on the manipulator and the environment. You can approach this problem in various ways. If the planning will occur offline, a complete description of the environment must be known in advance (including the motions of mobile and shape-changing obstacles). This is a global approach—it finds a solution before implementing it. In online planning, the manipulator makes decisions as it moves toward its goal. This approach considers only the area in the manipulator’s immediate vicinity—a local method. In this case, the path found might not be the most optimal or efficient. Most previous methods to coordinate robot manipulators along collision-free trajectories use configuration space (a transformation from Cartesian space1). In this approach, the manipulator maps to a point (and everything else in the environment goes through the same mapping process). This point is guided through the environment, usually by potential fields that exert an attractive force toward the goal and a repulsive force away from obstacles.1–3