A soft computing approach to cooperative robotics

The problem of coordination of multiple autonomous mobile robots, in the execution of tasks requiring or benefiting from coordination, is examined. Robots with limited computational resources (restricted memory, low-power microprocessors), simple sensors (distance ranging, such as sonar and infra-red, versus video cameras) and limited power are considered. A proxy control architecture, with coordinator and scout robots, based on a multi-agent fuzzy logic behavior hierarchy, and ad-hoc communication range-determined group formation, is proposed as a solution, augmented by a compact environment representation, in the form of a fuzzy occupancy grid map. This map reduces the memory required for its storage and the bandwidth required for its transmission and sharing among robots, therefore reducing the energy used for communication, while maintaining a level of detail equivalent to that of other mapping methods. The proposed architecture is tested with a simulated autonomous distributed mapping problem, where the environment is initially unknown, with exception of its overall dimensions. The robots successfully navigated the environment and generated a fuzzy occupancy map. The simulation results are discussed and directions for future research are outlined.