Great potentials of robotic networks have been found in numerous applications such as environmental monitoring, battlefield surveillance, target search and rescue, oil and gas exploration, etc. A networked multi-robot system allows cooperative actions among robots and can achieve much beyond the summed capabilities of each individual robot. However, it also poses new research and technical challenges including novel methods for multi-agent data fusion, topology control and cooperative path planning, etc. In this paper, we review recent developments in cooperative control of robotic networks with focus on search and exploration. We shall first present a general formulation of the search and exploration problem, and then divide the overall search strategy into different modules based on their functions. Methods and algorithms are illustrated and compared following the classification of the modules. Moreover, a 3D simulator developed in our laboratory is introduced and its application is demonstrated by experiments. Finally, challenges and future research in this area are provided.
An Enhanced Model-Free Reinforcement Learning Algorithm to Solve Nash Equilibrium for Multi-Agent Cooperative Game Systems