Design And Implementation Of Service Robot Cooperation And Simulation System Based On Multi-agent

Nowadays the application of robotics has changed from traditional industrial fields into new areas, such as agriculture, medical treatment, education, entertainment, hotel management and office. Meanwhile, a kind of robotics has developed rapidly, known as “service robotics”. Moreover, with the development of Internet, Internet of Things and Cloud Computing, networked collaborative robotics will become the trend in robotics. In view of the situation, the following work has been done.1. Studied the current research methods of Cooperation Problem and the current research status of Cooperation and Simulation System. The study found that the current research methods could be divided into two categories: the methods based on consultation and reaction and the methods based on Multi-Agent.2. Studied the theory and method of Multi-Agent and analyzed the key technical problem of Multi-robot Cooperation. We studied the properties and structure of Agent and three kinds of architecture of MAS(Multi-Agent System). Then, we discussed several key technologies: collaborative task description, collaborative content, collaborative form, resource statistics, task allocation, task scheduling, communication and interaction.3. Aiming at the problem of communication, we proposed a new kind of communication mechanism based on improved KQML, which could be applied on service robot cooperation. About the strong coupling between communication and knowledge sharing in KQML, we put forward a kind of communication mechanism based on status variables. In addition, we gave a pair of “message performatives” because KQML had not taken the communication across network levels into consideration.4. Concerning task allocation and scheduling problem, we raised a kind of self-adapting approach for server and robot, which included two sections: self-adapting allocation and self-adapting scheduling. The former was based on robot and network communication status, the latter was based on more fine-grained task immigration.5. Designed a self-adapting heterogeneous robot cooperation system, which had three characteristics: dynamics, self-adaption and universality. The dynamics meant the amount of service Agent could be changed dynamically, the self-adaption meant the task allocation module could allocate task based on the amount of resource and the university meant the system could be useful or helpful for many collaborative situations.6. Archived a collaborative SLAM(Simultaneous Localization and Mapping) mechanism based on Occupancy grid map, dynamic partition and following-wall in the system, which we experimented.