Optimising multistatic sensor locations using path planning and game theory

A novel approach towards the optimisation of a network of fixed multistatic sonar sensors is described. The scenario of interest is the deployment of a fixed network of multistatic sonar sensors at sea, to detect a transiting submarine threat. The study considers that any optimisation of sensor positions must account for an assumed level of threat intelligence. A worst-case threat with complete knowledge of the network is modelled using the A* path planning algorithm. This provides a measure of the performance of a given network configuration against a worst-case, most intelligent, threat. However, the implied assumption of complete knowledge of sensor positions is unrealistic. An important advantage of a multistatic network is the inability of the threat to counter-detect, and thereby determine the position of, the receiving nodes. In order to address this problem, and to account for a more realistic threat with incomplete knowledge of sensor positions, a game theoretic approach is described.