Teach-and-Repeat Driving Could Automate Lunar Cargo Delivery – Universe Today (www.universetoday.com)

Researchers at the University of Toronto’s Robotics Institute, led by Alec Krawciw and Prof. Tim Barfoot, are adapting a “teach-and-repeat” autonomous driving algorithm for Canada’s Lunar Exploration Light Rover (LELR) to automate repetitive lunar cargo runs between rocket landing pads and habitats. The approach has an astronaut drive the route once to collect training data, after which the rover can autonomously repeat that specific journey — a critical capability for routine 5 km transfers required by early lunar bases (landing pads must be placed away from habitats to avoid rocket shrapnel and radiation). The team ran a successful field trial at a Canadian analog terrain facility and has integrated their software into the LELR for further development toward Artemis logistics. Technically, the work involved adapting an algorithm not originally built for space, addressing operational delays from remote control by decomposing the long trip into shorter, chained segments, and ironing out real-world hiccups in analog conditions. Significance: automating mundane, repetitive cargo tasks reduces astronaut workload and exposure to lunar hazards, lowers operational risk, and helps scale sustained surface logistics. Beyond the Moon, lessons from this space-constrained teach-and-repeat deployment could improve robustness and practicality of terrestrial autonomous driving and route-following systems.