If you’ve ever been to a hockey game, you’re familiar with the large Zambonis that roll onto the rink between periods to clean up the ice. Motorized ice surface cleaners first began making their debut on ice rinks in the 1950s, and now, researchers are looking to make those machines run autonomously.
A team of students at Carnegie Mellon University (CMU), in collaboration with Duquesne Light Company, the Pittsburgh Penguins, Zamboni and Locomation, a Pittsburgh-based automation company that creates autonomous systems for semi trucks, has developed an autonomous Zamboni for the Pittsburgh Penguins.
The collaboration started with Locomation, whose founders, Çetin Meriçli, now the CEO, and Tekin Meriçli, now the CTO, were attending a Penguins game when two saw an opportunity for the technology they’ve been developing to be used in a new way.
Locomation creates autonomous systems for semi-trucks, and the company takes a unique approach to autonomous driving. Locomation wanted to keep a human in the loop, so its system involves an autonomous semi-truck, with a resting human driver behind the wheel, following another autonomous truck being operated by a human driver. These trucks are electronically tethered to move together and can even swap places so that the driver can rest while another takes over.
While someday the company aims to develop fully autonomous semi-trucks that don’t need to follow human-driven ones, it wanted to prioritize getting its technology out into the world. This allows its autonomous driver to learn from real-world scenarios while on the road.
Typically, two Zambonis clean the ice in between hockey periods. Locomation determined that it could have a similar system for the Zambonis that it uses in semi-trucks. The company reached out to John Dolan, director of the Masters of Robotic Systems Development program at CMU, to see if a group of students would be interested in working on the project with them. Locomation spun out from CMU in 2018, so it was a natural collaboration fit for both organizations.
The project spanned three semesters, starting with a brainstorming phase where the students determined how they would develop the system and test it, going from a small RC Car, which is a small remote-controlled platform, to a Zamboni.
After testing its proof-of-concept on the RC Car, the team of students worked on an HE vehicle, equipped with all of the sensors needed for autonomous operation. This stage of the project allowed the team to work out any final kinks in the automation technology before putting it onto a Zamboni.
Zamboni provided an all-electric ice cleaner for the team to work with and retrofit. The team has completed its first ice test, with its autonomous Zamboni following another human-driven Zamboni but with an offset, so the two vehicles aren’t cleaning the same path of ice. Final demos of the system are coming up in about less than a week.