This project deals with the coordination and control of multiple cooperating autonomous underwater vehicles with limited communication capabilities. The realistic condition considered is that communication below the surface is considerably more difficult and energy consuming than above the surface. This places special requirements on how the vehicle will share information and position themselves geographically and in relation to each other. In addition, it is assumed that when a vehicle surface, it can exchange information through a cloud service, which is able to process and store data. Through the cloud service vehicles can by an appropriate scheme surface to now and then exchange state information and update their plans. The cloud service could also minimize the risk of data loss if something happens to an individual vessel.
During this initial project period, the work will focus on decentralized coordination algorithms and control designs that respect communication limitations. In particular, we will consider that each vehicle will design its control and task planner based on the limited information it receives at dis-crete time instances from the cloud service. The team objectives will include both cooperative control tasks such as convergence to prescribed formation patterns and moving towards a common direction (a.k.a. flocking behaviour), but also higher-level tasks that evolve sequence of actions that may be defined in a team based centralized as well as an individual vehicle decentralized manner. These specifications may be defined as an automata leading to a distributed hybrid control approach. A novel aspect in the latter is the considerations of limited communication at the task planning and control layers simultaneously. The study will include modeling, analysis, and control de-sign. A simulation environment will be developed to enable efficient comparison between approaches and evaluate the influence of vehicle model and environmental uncertainties.