Research On Swarm Robot Formation And Cooperative Obstacle Avoidance Method

In recent years, the motion planning problem for multi-robot systems, especially swarm robot system, is becoming a hot topic in robotics. Because of the swarm robotic system’s feature of high efficiency, high flexibility and better robustness, it has been applied to aviation, military, and many other fields. Swarm robotics research has focused on obstacle avoidance methods, exploration of formation strategy. This article is mainly based on the artificial potential field theory and the behavior-based method and makes a preliminary study on the obstacle avoidance method for swarm robot formation, and at the meanwhile, improves the performance of the new method under the complex environment and real world.Based on the organizational structure of the robot, the methods of localization and distance measurement, and the obstacle avoiding algorithm such as the behavior-based method and the artificial potential field method, the thesis discusses the obstacle avoidance method of group of robots formation under the ideal situation, which is based on artificial potential field method and the behavior-based method. At the same time, the paper makes obstacles classification, and according to different circumstances of the obstacles, the thesis detailed design the formations behaviors, improves the obstacle avoiding capacity of the swarm robot team.Also, according to the inaccuracy existing in the process of localization and distance measurement and their math models, the thesis deals with these two errors respectively based on probabilistic fuzzy logic system and improves the localization and ranging measurement accuracy, which effectively increase the performance of the obstacle avoidance algorithm using for the complex environment.Finally, the proposed methods in the paper are tested by a series of simulation experiments. The experimental results demonstrate the feasibility and effectiveness of the proposed method for group robot obstacle avoidance under ideal situation. In addition, the proposed method still have good performance under real environment after improved by the probabilistic fuzzy logic system.