Information fusion as input source for improving multi-agent system autonomous decision-making in maritime surveillance scenarios

Decision making problems are usually referred as a cognitive process resulting in the selection of an action among several alternatives. Autonomous entities taking these actions like multi-agent systems, needs to process and understand its environment state to frequently update its beliefs, and then, select an optimal action. As an environment can be composed by several sources of information, it is useful for a multi-agent system, a way to process integrated information of multiple data which represents the same real-world object. This information can improve the agents knowledge and let select better actions than processing simple raw data. Most information fusion research has had a technical and algorithmic focus, and takes little attention to high level decision making, although some studies relate fusion to human decision making. However, in this paper is proposed the use of fused information as an input source for supporting and improving the decision making capabilities of autonomous agents in maritime surveillance scenarios.