Intelligent exploration and surveillance algorithms for multi-agents robotics systems

In this work are considered two problems in a multi-agent robotics surveillance scenario, exploration and surveillance, with which intruder entities must be detected and tracked. The goals are minimize the overall exploration time and increase the time in which intruders are tracked by the agents in a static environment. An intelligent approach is presented for the coordination of multiple robots using a market-based model considering three levels to find the suitable task for an agent: knowledge acquired, decision making algorithm and social interactions. For the two problems the decision making algorithms have been implemented using computational intelligent techniques. The proposed approach has been implemented in an indoor real-world environment using cell maps and its performance is tested in simulation runs. Experimental results and conclusions emphasize the advantages of using the coordination mechanisms developed in this work by contrasting the effectiveness in coverage area and tracking tasks to similar non-intelligent approaches.