煤矿狭窄非结构化巷道中履带式机器人动态避障方法
打开文本图片集
中图分类号：TD67 文献标志码：A
Dynamic obstacle avoidance method for tracked robot in narrow unstructured coal mine roadways
ZHOU Weimingl，XUNa¹，LIU Zhigang （.SchoolofAutomation Engineering，HenanPolytechnic Institute，Nanyang473ooo，China; 2.School ofResources and Environment， Nanchang University，Nanchang 33oooo， China）
Abstract：Existing robot obstacle avoidance methods mostly rely on a single sensor，which leads to large obstaclecalibration errors and insuficient safety margins in complex unstructured roadway environments where dynamic obstacles appear randomly.To addressthese problems，adynamic obstacleavoidancemethod fortracked robots based on multi-sensor perception and deep reinforcement learning Was proposed for narow unstructured coal mineroadways.The high-resolution imaging capability in the visible spectrumand the sensitivity to thermal radiationin the infrared band were used to perceiveroadway environments with low illumination and high reflectivity.The Mean Shift algorithm was introduced to perform kermel density estimationon theoccurrence probabilityofobstacles intheroadway，andthethree-dimensional spatialcoordinatesofobstacles werecalibrated to overcome the limited field of view and occlusion caused by the narrow roadways.The spherical envelope method wasused to constructthe safety potential field boundary corresponding to the thre-dimensional spatial coordinates of obstacles as the constraint condition for theobstacle avoidancereward in deep reinforcement learning，and therobot obstacle avoidance behavior Was optimized according to thereward to achieve dynamic obstacle avoidance.Experimental results showed that under conditions of high dust concentration， strong light， and weak light，the mean of cross-modal structural similarity between the visible light images and the infrared images of the perceived results was higher than 55% ， enabling accurate perception of the roadway environment. The maximum error between the calibrated obstacle position and the actual position was only 0.4m During movement， the minimum distance between the robot using the proposed method and obstacles was greater than the safety threshold，and no collsion occurred，indicating a sufficient safety margin for obstacle avoidance.
Key words： robot dynamic obstacle avoidance; tracked robot;；unstructured roadway； multi-sensor perception; deep reinforcement learning; Mean Shift algorithm; obstacle calibration
0引言
煤矿狭窄非结构化巷道是缺乏固定、明确空间结构的地下通道，存在黑暗、高粉尘、空间受限等极端条件，使机器人巡检面临高风险与低效率的严峻挑战[1]。（剩余9911字）