A University of Washington professor and expert in machine learning and robotics is leading a new company focused on off-road autonomous vehicle technology.
“Overland AI believes that autonomy is not just for roadways and warehouses,” Boots told GeekWire.
Boots is the principal investigator of the UW RACER team, one of three research groups involved with the U.S. Department of Defense Advanced Research Projects Agency’s (DARPA) Robotic Autonomy in Complex Environments with Resiliency (RACER) competition.
Overland AI was registered as a corporation on Jan. 12, according to Washington state corporations filings. The first public sign of the startup came when Boots posted about it on LinkedIn last week.
But key details remain under wraps. The UW spinout has raised funding, but Boots declined to name investors or provide more information about connections between Overland and DARPA.
Overland’s other two co-founders are part of the UW RACER team:
- Greg Okopal, a principal engineer at the UW Applied Physics Lab
- Stephanie Bonk, director of strategic initiatives at the UW Robot Learning Laboratory
The company’s CTO, Jonathan Fink, worked as a researcher for the U.S. Army Research Laboratory for more than a decade. He earned a PhD in electrical systems engineering from the University of Pennsylvania.
Boots earned a PhD in machine learning from Carnegie Mellon University, a leading robotics research institution. He then spent two years at the UW as a postdoc scholar and later became an assistant professor at Georgia Tech.
In 2019 Boots returned to the UW, where he is currently the Amazon Professor of Machine Learning at the Paul G. Allen School of Computer Science and Engineering.
Boots is also director of the UW’s Robot Learning Laboratory and co-chairs the IEEE Robotics and Automation Society Technical Committee on Robot Learning. He previously was a principal research scientist at NVIDIA Research’s Seattle Robotics Lab.
Overland is launching as other autonomous vehicle companies lay off workers or shut down and are struggling to raise more capital. Investor confidence has been undermined as projects in the field are exceeding anticipated timeframes and cost requirements.
Many self-driving startups have been focused on commercial driving, which presents various complexity.
But off-road navigation also has its own challenges, such as diverse vegetation, uneven surfaces, changing terrain, and other obstacles.
The goal of DARPA’s RACER competition is to create off-road autonomous driving tech that can navigate complex terrain at speeds that match or exceed human capabilities.
UW’s RACER team equips Polaris off-road vehicles with onboard sensors, computers and batteries to navigate rugged terrain without relying on GPS systems.
In September, the UW RACER team completed its second-round of testing in a hilly terrain in central California.
UW competes with Carnegie Mellon’s National Robotics Engineering Center and NASA’s Jet Propulsion Laboratory as part of the DARPA program.
The 2004 DARPA Grand Challenge, a competition where 15 teams constructed and raced their own fully-autonomous vehicles, is regarded as a catalyst for the self-driving vehicle industry. Its aim was to advance the development of autonomous vehicles for transporting cargo and military supplies to combat zones.
Since then, autonomous vehicles have found a number of off-road use cases in industries such as agriculture, construction and mining. Seattle startup Carbon Robotics, for example, has raised $36 million for its self-driving weed-zapping machines.