Autonomy algorithms developed by researchers at the University of Toronto Institute for Aerospace Studies (UTIAS) could one day make cargo transport on the moon safer and more efficient for astronauts.
As part of a team led by MDA Space, Professor Tim Barfoot and PhD student Alec Krawciw are developing technology to help Canada's proposed lunar utility vehicle navigate between cargo drop-off points during future lunar missions, addressing a key transportation challenge once astronauts land on the moon.
"Lunar exploration involves a landing site and a habitat site about five kilometres apart," says Barfoot, who also serves as director of the U of T Robotics Institute , an institutional strategic initiative .
"The landing site is flat for safe shuttle arrival, while the habitat needs to be shielded from radiation - typically behind rocky terrain. This creates a transportation challenge: astronauts must be able to move all cargo from the shuttle to the habitat."
Unlike previous planetary missions where rovers explore terrain in multiple directions to collect data, the lunar utility vehicle will make regular round trips between fixed locations to deliver goods and equipment to astronauts. This marks the first time a space rover will be required to repeat the same path, making Barfoot's visual teach-and-repeat navigation framework well-suited for the mission.
"Teach-and-repeat algorithms allow us to pilot the rover along a predetermined path by manually or physically driving it, [but] once it learns the path, it can automatically repeat the route as many times as you like," Barfoot says. "By automating this part of the mission, it saves astronauts time and energy returning to the landing site to pick up cargo, limits astronaut exposure to lunar elements and increases mission productivity."
As part of his PhD research, Krawciw is adapting the self-driving technology for integration with the Canadian Space Agency's test vehicle, the Lunar Exploration Light Rover (LELR).
In December 2024, Krawciw and Barfoot joined teams from MDA Space and the Centre de Technologies Avancées BRP at the Université de Sherbrooke to trial the autonomous system at the space agency's analogue terrain facility in Montreal, which replicates the surface of Mars. The field test provided an opportunity for the teams to identify and address any hardware and software constraints when operating in lunar-like conditions.
"Adapting our code to the LELR came with some unexpected challenges," says Krawciw. "Simulating lunar conditions introduced a five-second delay in command and feedback, so we couldn't rely on joystick control like we normally would. That pushed us to develop a new semi-autonomous teaching method using short path segments - something we hadn't done before."
"Despite the technical challenges, it's always exciting to see something I worked on in the lab come to life in a real space-focused mission."
After a successful field trial, the team was selected by the space agency in July 2025 to conduct an early-phase study for Canada's proposed lunar utility vehicle as part of the agency's lunar surface exploration initiative. This will be Canada's next contribution to NASA's Artemis program, which aims to establish a sustainable human presence on the moon.
As the team prepares the vehicle to be mission-ready, Krawciw is focused on improving the system's performance in real-world conditions and ensuring it is ready for long-duration deployments.
"We learned a lot from running the system continuously in the field," says Krawciw.
"It wasn't just about getting the autonomy to work - it was about making it reliable and user-friendly for operators who might be using it all day, in tough conditions. That perspective is shaping how I approach the next phase of development."
