Research on intelligent visual obstacle avoidance of underground mobile robot

In view of problems that existing obstacle avoidance methods of underground mobile robot cannot accurately detect obstacle position information when facing complex obstacles and is inability to perform accurate obstacle avoidance control for the underground nonlinear obstacles, an intelligent visual obstacle avoidance method of underground mobile robot based on fuzzy control was proposed. First, binocular stereo vision module is used as obstacle detection sensor to perceive underground environment information, detect distribution of obstacles in real time, and construct occupation grid map. Then， the octree structure model is used to construct three-dimensional point cloud, and the tree structure is used to describe point cloud data structurally, which is mapped to the occupation grid map to obtain regional distribution of obstacles. Finally, a fuzzy control strategy is used to process distribution of obstacles detected in real time in the occupation grid map, and distribution of obstacles in the occupation grid map at the current moment and the running speed of the mobile robot are used as input variables of the fuzzy controller. The fuzzy control algorithm is used to calculate steering angle and acceleration of the mobile robot at the next moment, so as to realize intelligent obstacle avoidance control of the underground mobile robot. According to actual space occupied by the mobile robot, an external bounding box is designed to further stabilize control algorithm, and the obstacle avoidance strategy is combined to perform intelligent obstacle avoidance to avoid collision between mobile robot and obstacle. The experimental results show that the method can accurately describe distribution of underground obstacles, and enable the mobile robot to avoid obstacles accurately and autonomously according to the designed fuzzy control rules, so as to realize adaptive movement.