Skip to Main Content
A kind of system architecture based on multi-agent is proposed in this paper for autonomous control system of an autonomous underwater vehicle (AUV). The autonomous control system is divided into several agents according to its software function and hardware structure. Each agent is designed as a deliberative/act hybrid mode. The agent not only recognizes science events and makes decision unattached, but also cooperates with the other Agents to accomplish the complex task. Based on Multi-Agent System (MAS), dynamic collaboration mechanism for AUV task planning is researched. The DEDS (Discrete Event Dynamic System) model of the task planning is created by improved Petri Net (PN). The architecture and task planning method have been integrated and experimented in lake.