The drive to better understand robotics and artificial intelligence now includes a cool little robotic race car thanks to researchers in the Personal Robotics Laboratory at the University of Washington’s Paul G. Allen School of Computer Science & Engineering.
MuSHR, or Multi-agent System for non-Holonomic Racing, looks a bit like the kind of radio-controlled car you might want to launch off some sweet jumps in the backyard. But it’s actually an open-source, full-stack robotics platform, and the easy-to-assemble, low-cost nature of it all makes it a fit for advanced robotics research and education — or plain old hobbyists and hackers.
Allen School professor Siddhartha Srinivasa called MuSHR an ideal test platform for some of the tough challenges in robotics, including those involving autonomous vehicles.
“Beyond research, we also need to think about how we prepare the next generation of scientists and engineers for an AI-driven future,” Srinivasa said. “MuSHR can help us answer that challenge, as well, by lowering the barrier for exploration and innovation in the classroom as well as the lab.”
MuSHR, with a name derived from the UW’s Husky mascot and sled-dog racing terminology, is built using off-the-shelf and 3D-printed parts. Everything needed to assemble and operate one is detailed on the MuSHR website.
And the UW is particularly proud of the fact that it’s bringing the technology to the public for a much lower cost than competitors. According to Allen School News, a more complex system such as Georgia Tech’s AutoRally can cost upwards of $10,000. MIT RACECAR, the platform that inspired the UW team’s project, is more budget friendly at $2,600, with basic sensing capabilities. A MuSHR race car with no sensing can be assembled for as little as $610, while a high-end car equipped with a multitude of sensors can be built for around $930.
The MuSHR team believes its system rises above simple, educational platforms that have very little functionality while also being way less expensive and easier to use than complex, research-oriented platforms.
At an Amazon cloud technology conference last fall, the tech giant unveiled a first-of-its-kind global autonomous racing league called AWS DeepRacer using mini race cars designed to help developers learn more about machine learning.
Read more about MuSHR on the Allen School website.
UW researchers this week unveiled a separate toolkit that helps developers build privacy and security features into augmented reality apps.