Abstract
With the improvement of intelligence mine, underground operation robots are gradually being applied in underground operation areas to replace manual operations in complex coal mines. Due to the interference of uneven lighting, water mist, dust, and other environmental factors, it is difficult for the robots to identify the boundaries of underground roadways and achieve autonomous movement toward the target working area effectively. To solve the problem of identifying roadway boundaries in complex underground scenarios and provide necessary scene information for underground mobile equipment, a reliable detection method for underground roadway boundaries based on the 4D Millimeter-wave radar is proposed in this paper. While fully utilizing the insensitivity of 4D millimeter-wave radar to environmental factors such as lighting conditions, water mist, and dust, a point-cloud dense reconstruction method based on spatiotemporal composite is proposed for the 4D millimeter-wave radar, which solves the problem of sparse 4D millimeter-wave point clouds. Further, a method for extracting underground tunnel boundaries is designed to provide reliable scene information for the autonomous movement of underground mobile equipment. Finally, the simulation roadway experimental platform and underground experimental platform are established to conduct experimental research. The results show that the maximum unilateral error for extracting boundaries is less than 0.1m, and the completeness and correctness rate are greater than 95%.
Get full access to this article
View all access options for this article.