Trend AnalysisOther Engineering

Autonomous Vehicles Underground: Navigating Mines Without GPS

Underground mines are among the most challenging environments for autonomous vehicles: no GPS, limited visibility, narrow tunnels, and hazardous conditions. Recent advances in LiDAR-based localization, behavior-based control, and post-blast UAV inspection are making autonomous mining operations increasingly practical.

By Sean K.S. Shin
This blog summarizes research trends based on published paper abstracts. Specific numbers or findings may contain inaccuracies. For scholarly rigor, always consult the original papers cited in each post.

Underground mining is one of the most dangerous occupations globallyโ€”hazards include rock falls, toxic gases, flooding, equipment collisions, and blasting accidents. Autonomous vehicles could significantly reduce human exposure to these risks by performing haulage, inspection, and rescue operations without human operators in the most dangerous areas. But underground environments present technical challenges that surface autonomous vehicles do not face: no GPS signals, limited or absent lighting, dust and particulates that degrade sensors, narrow and irregular tunnel geometries, and the need to operate safely alongside human workers.

The Research Landscape

Post-Blast UAV Inspection

Nordstrรถm and Dahlquist (2025), with 5 citations, report what they describe as the first fully autonomous UAV mission to perform gas measurements after a real blast in an underground mine. The mission was deployed approximately 40 minutes after blastingโ€”a period when the area is hazardous due to residual gases (CO, NOโ‚‚, SOโ‚‚) and unstable rock.

The UAV navigated autonomously through the mine using LiDAR-based SLAM (Simultaneous Localization and Mapping), measured gas concentrations at multiple points, and returned safely. The practical value is clear: currently, human workers must wait hours before entering post-blast areas for gas clearance. Autonomous UAV inspection can perform this assessment much soonerโ€”potentially allowing production to resume faster while keeping workers safe.

Behavior-Based Navigation for Narrow Tunnels

Badr and Almaghout (2024), with 1 citation, address the specific challenge of autonomous navigation in narrow mine tunnels where conventional path planning algorithms struggle. Their behavior-based control approach draws on robotics techniques that decompose complex navigation into simple reactive behaviors (wall-following, obstacle avoidance, goal-seeking) that combine to produce robust tunnel navigation.

The approach is notable for its robustness: it handles irregular tunnel geometries, dynamic obstacles (fallen rock, other vehicles), and sensor degradation (dust, water on LiDAR lenses) better than optimization-based planners that assume accurate environmental models.

Articulated LHD Automation

Wu and Lu (2025) present an autonomous driving system for articulated Load-Haul-Dump (LHD) machinesโ€”large vehicles that scoop ore and transport it through tunnels. LHDs present additional challenges: they are articulated (bending in the middle), non-holonomic (cannot move sideways), and operate in feature-poor environments where LiDAR scans show little variation between one section of tunnel and the next.

Their system was validated in a real underground mine, demonstrating autonomous loading, hauling, and dumping cyclesโ€”the core operational workflow that LHDs perform continuously during production.

Cyber-Physical Safety System

Behera, Agarwal, and Badr & Almaghout (2024), with 2 citations, take a systems approach: building a cyber-physical system (CPS) that integrates an unmanned ground vehicle with environmental mapping, gas sensing, and computational intelligence for mine safety and rescue support. The UGV maps unknown mine sections, identifies hazards, and communicates findings to surface operators through a real-time data link.

Critical Analysis: Claims and Evidence

<
ClaimEvidenceVerdict
Autonomous UAVs can safely inspect post-blast mine areasNordstrรถm et al.'s real-mine demonstrationโœ… Supported โ€” first real-world deployment
Behavior-based control handles narrow, irregular mine tunnelsBadr & Almaghout's simulation and limited field testsโœ… Supported โ€” robustness demonstrated
Autonomous LHD haulage is feasible in real underground minesWu & Lu's mine validationโœ… Supported โ€” complete load-haul-dump cycles
CPS-based UGVs can provide mine safety and rescue supportBehera et al.'s prototype systemโš ๏ธ Uncertain โ€” prototype demonstrated; operational deployment not yet

What This Means for Your Research

For mining engineers, autonomous vehicles are moving from research prototypes to operational systems. The post-blast UAV inspection use case (Nordstrรถm et al.) is likely the nearest to commercial deployment. For robotics researchers, underground mines provide a uniquely challenging testbed for autonomous navigation.

Explore related work through ORAA ResearchBrain.

References (5)

[1] Nordstrรถm, S., Stathoulopoulos, N., & Dahlquist, N. (2025). Safety Inspections and Gas Monitoring in Hazardous Mining Areas Shortly After Blasting Using Autonomous UAVs. Journal of Field Robotics.
[2] Badr, A.A. & Almaghout, K. (2024). Navigating Narrow Margins: Behavior-Based Control for Autonomous Mining Vehicles. Regular and Chaotic Dynamics.
[3] Wu, K. & Lu, Z. (2025). An Autonomous Driving System for Articulated Underground LHD Machines. Proc. RICAI 2025, IEEE.
[4] Behera, L., Agarwal, S., & Sandhan, T. (2025). A cyber-physical system based UGV for safety inspection and rescue support in underground mine. International Journal of Intelligent Unmanned Systems.
Behera, L., Agarwal, S., Sandhan, T., Sharma, P., Kumar, A., Ranjan, A., et al. (2025). A cyber-physical system based unmanned ground vehicles for safety inspection and rescue support inย anย underground mine. International Journal of Intelligent Unmanned Systems, 13(1), 92-128.

Explore this topic deeper

Search 290M+ papers, detect research gaps, and find what hasn't been studied yet.

Click to remove unwanted keywords

Search 8 keywords โ†’