Obstacle avoidance method based on the movement trend of dynamic obstacles

摘要

Obstacle avoidance is a well-known problem in the field of robotics. In this paper, movement trends of obstacles are proposed to be considered to improve the classical obstacle avoidance method. According to the movement laws of the dynamic obstacle, the movement trends of the obstacle are predicted to help the robot to plan the route and the mode of avoidance. Fuzzy logic technique is applied to predict the movement trend of the obstacle and dominate the linear velocity of the robot. In this paper, movement trend of obstacle are divided into approaching to the robot, moving away from the robot, translating with the robot. According to the three trends, the robot can choose to increase the velocity, decrease the velocity to pass by the obstacle, or stop to wait for the obstacle to pass. Some simulation results show that the proposed method can help the robot to accomplish obstacle avoidance without changing the initial moving path.