This paper discusses the use of cooperation and distributed intelligence within a robot team to solve a maze problem in which a set of ordered control point tasks has to be performed before the robots exit the maze. At first, the maze structure and the control point locations are unknown. While the maze is explored, each robot dynamically maps the maze structure with information provided by its sensors and sent by the other robots. Within each robot, a dynamic scheduler uses the gathered information to update the robot goal.
C. R. F. Costa, Eliana P. L. Aude