Major step toward fully autonomous underground mining

Researchers from Luleå University of Technology have successfully carried out large-scale field trials within the ReNAM project at the Epiroc Test Mine in Kvarntorp, outside Örebro. The experiments mark an important step toward fully autonomous and reactive mining vehicles, enabling safer, more efficient, and more intelligent underground operations. The project is a collaboration between LTU, Epiroc Rock Drills and Algoryx Simulation.

During the trials, the research team tested and demonstrated new algorithms for autonomy and perception in realistic mining environments. The experiments included scenarios such as mixed traffic, blocked tunnels, and rockfall simulations.

A key achievement was the integration of the research team’s autonomy system and an Epiroc MT42SG mining truck, equipped with custom-designed sensing, computation, and autonomy stacks developed within the project. This setup enabled two-way communication between the system and the truck, allowing researchers to both receive and send real-time data.

In one of the most significant demonstrations, the team successfully tested a Body-and-Dynamics Aware Path Planner, designed to enable mining vehicles to perform obstacle avoidance maneuvers in tight spaces. Meanwhile, operators at Epiroc’s Control Room could follow the experiments in real time through a visualization interface showing sensor data and algorithm output.

“This collaboration shows what’s possible when research and industry work side by side,” says Björn Lindqvist, Associate Senior Lecturer. “It’s a key milestone toward integrating our autonomy systems on full-scale mining vehicles.”

“These trials clearly demonstrate how advanced perception and reactive decision-making can be brought together in real mining conditions,” says Anton Koval, Associate Senior Lecturer at the Robotics and AI group. “Seeing our algorithms handle dynamic, unstructured environments in real time is an exciting step toward truly autonomous underground operations.”

From research to real mining operations

The ReNAM project aims to push the limits of existing mining autonomy by introducing a layer of robotics-inspired reactive navigation to mining machines. This approach allows vehicles to handle complex, dynamic situations beyond predefined routes — such as meeting other vehicles, avoiding unexpected obstacles, or adapting to environmental changes in real time.

ReNAM also extends the use of simulation-driven development, using high-fidelity physics-based simulations to efficiently evaluate autonomy concepts before real-world deployment. This work was enabled by the state-of-the-art virtual prototyping and testing solution developed by Algoryx Simulation AB, extended for this project with new capabilities. The continuous simulation integration was a major factor for speed and efficiency of the development iteration pipeline.

The successful evaluation of ReNAM’s ramp driving and mixed-traffic scenarios demonstrates how autonomous mine trucks can operate safely in shared environments reducing the need for isolated vehicle zones. On a larger scale, advances in onboard perception and integration with a central Mining Management System bring the mining industry closer to driverless trucks and full digitalization. Real-time data from sensors, vehicle positions, and detected objects can now be streamed, shared, and updated continuously across systems.