The launch and recovery (L&R) of an autonomous underwater vehicle (AUV) by a manned submersible (MS) is required for joint AUV/MS missions. The authors consider L&R as a distributed artificial intelligence (DAI) task. The approach has been taken to provide assistance in definition of subtasks, allocation of task responsibility, and specification of required data acquisition, processing, and communication. An architecture for the task requires special-purpose communications and multi-agent planning. The authors define the problem, the agents. the goals and constraints, review DAI, and discuss some real-world experiences. Three candidate DAI architectures are examined, and a hybrid, consisting of elements of all three, is suggested for further investigation.
展开▼
