MSL researchers refine autonomous mine equipment in Sweden

Posted on May 02, 2017

Engineering professor and Mining Systems Laboratory director Josh Marshall has spent most of the last year leading a team of Queen’s graduate students on a series of research and development projects in Örebro, Sweden.

Mine operators around the globe continue to build for a future of increasing autonomous mine operation, and Marshall and his students have already done extensive work to develop and refine control systems for autonomous trucks and loaders. They have, for the last five years, experimented with a small proof-of-concept test loader at Innovation Park in Kingston. The work in Sweden is chance to now test and refine those systems on full-size, operational underground loaders and trucks in a working mine. It’s also a chance to share in the process with other experts in the field.

Marshall’s team arrived at Atlas Copco’s automation R&D facilities and the Kvarntorp mine in August last year for the nearly year-long collaboration. As well as partnering with the mining equipment giant, Marshall’s team is also working with researchers at the Centre for Applied Autonomous Sensor Systems (AASS) at nearby Örebro University. The funding for the exchange comes in part from both the Natural Science and Engineering Research Council of Canada (NSERC) and the Swedish Knowledge Foundation.

“With me are Heshan Fernando, a PhD student in mechanical engineering; Lukas Dekker, a Master’s student in mechanical engineering; and Jordan Mitchell, a Master’s student from mining engineering,” says Marshall. “There are a few subprojects – one for each student – but the main gist is that we’re trying to make autonomous Atlas Copco-built trucks work much faster than they can now. We’re also continuing with a robotic loading problem that we started work on a few years ago with Atlas Copco: How can we make autonomous loaders learn how to load better? How can we design robotic loaders that get better at digging every time they do it?”

Fernando is working on developing and refining learning-based control algorithms associated with the autonomous loading problem. Dekker is exploring new, faster and more accurate ways autonomous vehicles can guide themselves under ground. Mitchell is working on developing a concept design for an auto-rotating cavity scanning device that could cheaply and automatically generate detailed maps of the insides of shafts or cavities.

“Fully equipped, these robotic Atlas Copco load-haul-dump machines are $1-million to $2-million pieces of equipment, so it’s kind of cool that they’re letting us write and test our code on their big, expensive, hardware,” says Marshall. “It has all worked out pretty well as a sabbatical year, and being able to bring the grad students was a bonus, too.”

Dr Johan Larsson, Lukas Dekker, and Dr Joshua Marshall
FUTURE OF MINING: Dr Johan Larsson (Atlas Copco, with his PhD from AASS), Lukas Dekker (Master’s student from Queen’s), and Dr Joshua Marshall, underground at Kvarntorp mine working on autonomous driving for mining vehicles.

PhD Student Heshan
THE SOFTWARE OF LEARNING: Heshan Fernando (PhD student from Queen’s) underground at Kvarntorp Mine, working on robotic loading for mining machines.

Master's Student Mitchell
SITUATIONAL AWARENESS: Jordan Mitchell (Master’s student from Queen’s) working at AASS on a new UAV-based scanning system for underground cavities.