Navigation in Multi Robot system using cooperative learning: A survey

Abstract

Multi-Robot system became an important research area in the Robotics and Artificial Intelligence. In the presence of obstacles, uncertainty and incomplete information Multi-Robot system is used to accomplish the task in more efficient way as compared to single robot system. Navigation of robots in its surrounding is essential to avoid unacceptable situation such as avoiding obstacles, trajectory planning. Cooperative behavior of multi-robots provides efficient way to avoid collision. It is a matter of concern that how a group of mobile robots should be controlled so that they can move in a cohesive way toward a single direction. For the problem of cooperation among robots, flocking strategy is a good solution. To learn cooperation among robots, various machine learning strategies have been developed. One of the machine learning techniques is called reinforcement learning. It is a very challenging issue in the area of robotics and artificial intelligence. Combining cooperative and learning strategy collision avoidance can be improved. Survey of multi-agent reinforcement learning and flocking control is presented in this paper.