Abstract: At present,the inspection work of the drainage equipment of the coal mine pump room still needs to be completed manually. The inspection is not in place,and the digitalization of inspection results is inconvenient. Combining odometer positioning technology,wheeled mobile robot navigation and obstacle avoidance technology,and based on the ROS robot development platform,this paper studies a global path planning and tracking algorithm for an inspection robot suitable for an underground pump room environment. The inspection robot moves autonomously according to the set path without human intervention and without the assistance of magnetic tracks,cables or GPS to complete the inspection of the pump room equipment. This paper firstly establishes a two-dimensional grid map according to the floor plan of the pump house,and plans an obstacle-free path through all detection points and travel points in the built two-dimensional grid map through the PRM path planning algorithm,and then uses the Pure Pursuit path tracking algorithm. Calculate the real-time linear velocity and angular velocity required for the inspection robot to walk along the path,and send it to the inspection robot motion controller,and then the controller monitors the obstacles within 10cm of the robot in real time according to the radar data,and decides whether the inspection robot adopts the current The linear velocity and angular velocity of the robot walk along the planned path,and finally,after the robot reaches the detection point,the drainage equipment is photographed and the operation data is collected. Through the analysis of the simulation experiment results of the inspection robot,it is concluded that the inspection robot can avoid obstacles on the path and reach the inspection point accurately. The average error of the actual position of the robot is 4 cm,and the average error of the attitude angle is 0.095 rad,which meets the requirements of the robot inspection work.