In this study, we propose an off-line and on-line planning algorithm for cooperative tasks of multiple mobile robots. Sweeping means a motion in which a robot covers a 2-dimensional area with its effectors (Fig..1). This task has variety of applications such as floor sweeping, painting, and object exploring. The former studies on this topic applied reactive planning based only on sensor information. These approaches have disadvantage of decreasing task efficiency immediately as the shape of environment becomes complicated. The planner proposed in this study has two phases: the off-line planner generate the sweeping trajectory based on environmental model, and the on-line planner navigates robot’s motion that covers the map errors and robot motion errors. As the off-line planner, we introduce both edges of the configuration space and Voronoi diagram so as to compute paths in the whole area. Appropriate paths are assigned to each robot by considering path length (Fig.2）．As the on-line planner, we propose (a) an algorithm in which a robot partially recognizes unknown obstacles and re-constructs its path locally by utilizing the stacker-crane problems in graph theory(Fig.3), and (b) an algorithm that measures a part of contour on the map while running along it; and generate the motion with correcting the map and adjusts the path(Fig.4). Experimental results with holonomic robots show the effectiveness of the proposed algorithm from the viewpoint of efficiency and adaptability.

1. Daisuke Kurabayashi et al.: Cooperative Sweeping by Multiple Mobile Robots, In Proc. 1996 IEEE Int. Conf. Robotics and Automat.,2, 1744/1749 (1996).
2. Daisuke Kurabayashi et al.: Local Path Re-Planning for Unforeseen Obstacle Avoidance by an Autonomous Sweeping Robot, In Proc. 1998 IEEE Int. Conf. Robotics and Automat., to be appeared(1998).
3. Daisuke Kurabayashi et al.: Real-time Path Adaptation for Sweeping by Autonomous Mobile Robots, In Distributed Autonomous Robotic Systems 3, Springer, to be appeared(1998).