Research And Simulation Of The Arithmetic Of Multi-Robot Cooperative Location

As one of the greatest human inventions in the20th century, robots represent the most advanced intelligent control technology in the current society. Multi-robot system has many incomparable advantages over single robot system, mainly contained:better fault tolerance, higher efficiency and better scalability, and easy to research and development. In numerous studies problem of multi-robot systems, robot localization is one of the most basic research problems, which is the fundamental guarantee of multi-robot system to complete the task. Multi-robot systems in practical application mainly work in complex and dynamic environment, so the research on the multi-robot cooperative localization has important theoretical significance and practical significance.Firstly, this paper outlines the superiority of multi-robot systems, then expounds the basic problem, technical features and developing trend of multi-robot cooperative localization, explains the research background, status-quo of multi-robot cooperative localization technology and research significance.Then different algorithms which combine robot motion information and relative observation information to cope with multi-robot system is studied. The study is focused on enhance robot location accuracy. Through establish robot point target movement model, Kalman filter equation which combine motion information and relative information is studied. And location features based on relative orientation is analyzed. Simulation results verified the accuracy and effectiveness of EKF location algorithmAs the EKF algorithm is high to the initial conditions, This paper researched the use of information fusion particle filter, Simulation results show that particle filter, have the same level on precision with EKF. The particle filter algorithm has good positioning effect but computationally expensive.Then, PF-EKF location algorithm is studied. The algorithm is used to cope with multi-robot cooperation location in unknown environment. In initial stage, the algorithm uses PF to location multi-robot system,when particle constrict to approach Gaussian distribution, EKF algorithm is used instead of PF Simulation result shows that:this method can solve realize multi-robot cooperative localization with unknown initial conditionsAs the increasing number of robot cases a decrease in efficiency and precision of the algorithm, this paper proposes reference target selection method based on maximum information entropy and combines with EKF algorithm to select reference target. The simulation experimental results show that when satisfying cooperative localization accuracy, the method can significantly reduce amount of calculation.The above research obtains satisfactory results by a series of simulation experiments, which provides a basis for further research of multi-robot cooperative localization and Multi-humanoid robot cooperative localization.