• Learning Algorithms
• Game Theory
• Detection
• Defence
• Coverage
• AUV
• Asymmetric threats
• Naval platform
• Sonar

An important application of intruder detection system concerns the protection of a main ship in asymmetric threat scenario. A team of robots is deployed to compensate for degradation of available surveillance sensors on naval platform. The multi-robot coordination must deal with critical environmental conditions such as limited communication and climatic conditions.
The team must guarantee the coverage of the area surrounding the ship while tracking a possible intruder. This problem is modelled as a potential game for which learning algorithms are introduced. The algorithms take into account the model of AUVs by imposing constraints on feasible actions selected by each robot.
Moreover an estimation of detection performance is introduced to model on-board sensors as real-like diver detection sonar. The algorithms performances have been evaluated in different dynamic features of environment (i.e. threat characteristics).