Study On Autonomous Cooperation Control And Reinforcement Learning For Multi-robot Systems

In the late 1980s, with the development of Artificial Intelligence (AI) theory, some researchers applied Multi-agent theory to the study of Multi-robot systems (MRS), and later started to study a new direction in robot field—Multi-robot cooperation technologies.In the past two decades, with the deeper and deeper research of DAI (Distributed Artificial Intelligence) and complex systems, MRS has improved greatly in theory and application. As an important embranchment of robotics, it shows a wide prospect in the fields of aviation and aerospace, geologic exploration, medical treatment service, traffic management, scientific expedition, military and civil affairs est. It can be predicted that in following years Multi-robot technologies will be widely used in every aspect of our common lives. And especially the major development directions on MRS will focus on enlarging the number of robots, improving the cooperation ability, environments of MRS will be changed from known structured to unknown unstructured and dynamic, improving the robustness and rebuilding ability of system, improving the adaptability of single robot and the cooperation adaptability of swarm robots system, designing standard test platform and translating the theoretical results into real applications. It is under the above new requests to study MRS in unknown dynamic environment. Therefore, it has significant theory and application value.This paper studies the cooperation problems of the MRS in complex dynamic environment based on transit and forging tasks. The main research work about this...