抄録
This paper deals with a method to design the control architecture for cooperative behavior of multiple autonomous mobile robots using an extended Petri net based approach. Centralized planning with hierarchical net models has been adopted to be easily implemented in Petri net formalism. A multi-robot task net model can be defined as the union of several single-robot net models enriched with synchronization constraints between actions of different robots. The synchronization constraints are used to add temporal constraints in the execution of the actions in a multi-robot task, such as sequential execution and simultaneous execution of two actions by two robots. Given a multi-robot task net model, the single-robot net models can be automatically derived by replacing synchronization constraints with communication actions and can be executed in an autonomous decentralized manner, where all the conflicts in the actions of different robots are resolved without the need of a central coordinator.
