A FAST MOTION PLANNING ALGORITHM FOR A MOBILE ROBOT USING A DISTANCE TRANSFORMATION IMAGE

This paper proposes a new high-speed motion planning algorithm for the mobile robot functioning on the distance-transformation image. The method utilizes the property that the value of the distance transformation image represents the distance from the obstacle. By the proposed method, the robot can move by translation to the farthest possible point only by examining a small number of points on the straight-line of the progress. This makes it possible to realize the fast motion planning. When a deadlock is produced, a new posture of the robot is determined at the position so that the collision with the obstacle does not occur. Then, the part of the motion down to the initial point is regenerated. Finally, the speed of the proposed algorithm is compared to that of the conventional motion planning algorithm, and the fast operation is demonstrated experimentally.