Adaptive Group Formation in Multirobot Systems

[摘要] Multirobot systems (MRSs) are capable of solvingtask complexity, increasing performance in terms of maximizingspatial/temporal/radio coverage or minimizing missioncompletion time. They are also more reliable than single-robotsystems as robustness is increased through redundancy. Manyapplications such as rescue, reconnaissance, and surveillance andcommunication relaying require the MRS to be able to self-organizethe team members in a decentralized way. Groupformation is one of the benchmark problems in MRS to studyself-organization in these systems. This paper presents a hybridapproach to group formation problem in multi-robot systems. This approach combines the efficiency of the cellular automata asfinite state machine, the interconnectivity of the virtual grid andits bonding technique, and last but not least the decentralizationof the adaptive dynamic leadership.

[学科分类] 人工智能