Sandvik Mining and Rock Solutions has introduced the "Auto-Tram+" autonomous haulage package for its LH518i 18-ton underground loaders (LHDs). Underground mines present a harsher automation environment than open pits: tunnels are narrow, GPS signals are absent, visibility is zero due to dust, and human miners in light vehicles share the same decline roads. Previous semi-autonomous systems required a human operator at a remote surface station for every machine, limiting productivity gains.
Auto-Tram+ solves this with a "Fleet Digital Twin" architecture. Each LHD is equipped with "Front and Rear 3D LiDAR Scanners," an inertial measurement unit, and wheel odometry sensors. During a single manually-driven "learning run" down each tunnel, the LiDAR builds a centimeter-resolution point cloud map of the driveage. All machines in the fleet share these maps over the mine's underground Wi-Fi 6 backbone, meaning a tunnel only needs to be mapped once by one machine for the entire fleet to navigate it.
The engineering breakthrough is the "Light-Vehicle Detection and Negotiation Logic." Because miners' pickup trucks and utility vehicles also travel the declines, the system cannot simply reserve roads exclusively for autonomous machines. The forward LiDAR continuously classifies detected objects. When a light vehicle is recognized approaching from the opposite direction, the LHD calculates the distance to the nearest designated "refuge niche" (a widened passing bay carved into the tunnel at regular intervals). The autonomous loader automatically pulls into the niche, halts, and flashes its blue autonomous-operation beacon, allowing the human-driven vehicle to pass. If no niche is available within range, the LHD reverses to the last one it passed. Mine trials in a Canadian nickel operation demonstrated 8% higher tonnage per shift simply by eliminating operator shift changes and travel time to and from the face, while fuel-free electric cable reels cut ventilation costs by a further margin.
