Information-driven cooperative distributed motion planning for long range search over marine areas

This work aims at proposing a method for building a map of a priori percentage density of expected findings over an area from historical qualitative information and at defining a cooperative distributed exploration algorithm guided by the minimization of the information gain over the a priori map. The exploration algorithm is implemented by partitioning the area with the Equitable Power Diagrams theory, through potential functions for motion planning and taking into account communication constraints. Simulations are carried out using the available archaeological data base of the Tuscan Archipelago, Northern Tyrrhenian Sea, and assuming that the exploration payload in each vehicle is a side scan sonar. The algorithm is compared with a standard systematic search strategy, based on regular transects over the area, and with the Rapidly Exploring Random Tree (RTT) planning approach which exploits the same a priori information as our algorithm.