A distributed multi-AUV control framework for cooperative missions

Intelligent multi-AUV systems are becoming well-established due to the advantages they offer in various underwater monitoring applications. One current challenge is designing and implementing distributed control frameworks to enhance cooperation between agents while accounting for the limitations imposed by the underwater environment. The only viable option for communication between agents underwater is acoustic communication, which is known to be unreliable due to low bandwidth, high latency, and frequent packet loss. In this scenario, we propose a sequential multi-agent decision framework to tackle monitoring missions that require explicit cooperation between agents. The cooperative mission is formulated as an optimization problem using a Partially Observable Markov Decision Process, in which each agent must meet specific operational objectives to improve the overall mission outcome. This work presents the software architecture developed to handle the guidance and control of the multi-AUV system and validates the proposed approach in a simulation environment that reproduces a cooperative monitoring mission while considering the major limitations imposed by the acoustic communication channel.