A novel global re-localization method for underground mining vehicles in haulage roadways: A case study of solid-state LiDAR-equipped load-haul-dump vehicles

Abstract

This study introduces a global re-localization method for mobile mining vehicles equipped with solid-state LiDAR in underground haulage roadways. The method includes feature extraction from LiDAR data, an improved fast Euclidean clustering algorithm for point cloud classification, and a descriptor based on Delaunay and Extended Delaunay Triangles centered on roadway features. A global re-localization process is established using a historical keyframe search strategy, enabling swift re-localization of mining equipment. 21 experiments were conducted with a load-haul-dump vehicle fitted with solid-state LiDAR across three underground mine haulage roadways. The proposed method achieves rapid global re-localization with an accuracy of within 0.2 m in under 40 ms, demonstrating the significant advantages and practicality of the proposed global re-localization method.