Controlling a team of unmanned underwater vehicles performing water region search tasks

Abstract

The article discusses a problem of performing a mission and controlling a team of unmanned underwater vehicles which perform a task of water region search. The team of underwater vehicles has been modelled as a multi-agent system, which made it possible to use the mechanism of multi-agent systems for controlling the team of cooperating underwater vehicles, with each vehicle being treated as an individual agent. Solutions to cooperation problems in multi-agent systems were proposed, including the presentation of the system structure and the data exchange method. Issues relating to cooperation actions in the team of underwater vehicles were described and solved using the theory of games and the vector quality index. The process of negotiations between the vehicles was modelled as the multiplayer cooperative game, which secured successful realisation of the task performed by the team of vehicles with respect to time, which represents the economic aspect of the task. The optimal solution was selected based on the vector quality index, which provided opportunities for minimisation of costs having the form of energy used by the entire team, and maximisation of profits in the form of time needed for completing the water region search task. The proposed method for dividing the action space in the water region search tasks performed by the team of underwater vehicles consists in assigning to each of cooperating underwater vehicles a task the realisation of which contributes to the realisation of the globally defined task. Finally, the results of water region search tasks performed by a team of cooperating underwater vehicles are presented and discussed taking into account possible vehicle failures happening during task realisation.