The founders of Argo AI, the self-propelled startup supported by Ford, are returning to their alma mater for a $ 15 million investment in Carnegie Mellon University to fund the creation of a new research center.
The Carnegie Mellon University Argo AI Center for Autonomous Vehicle Research (phew) will use these funds to continue "advanced research projects to overcome obstacles that allow self-driving vehicles to work in a variety of real-world conditions, such as winter weather. or construction zones ", the company and the university announced Monday. Argo was founded in 2016 by a team of CMU alumni.
Argo – which is testing its vehicles together with Ford in Miami, Washington, DC, Palo Alto and, most recently, Detroit – will support research into advanced perception and decision-making algorithms for autonomous vehicles. In other words: the software and hardware that enhance the ability of a self-driving car to "see" and "think".
But this is not an Argo navel-gazing project, nor is it a benevolent gift from a handful of CMU alumni who have made it big. (In 2017, Ford said it would spend $ 5 billion on Argo for five years.) This research project focuses on enabling "large-scale, global deployment" of self-driving cars. This is money to get self-driving cars on the road faster and on a larger scale.
Autonomous vehicles are tested in small batch implementations in cities around the world, but they are still a long way from & # 39; global deployment & # 39 ;. To get there, the cars must be proven to be safe to work in all types of road and weather conditions. People have to trust the technology – which they don't currently do – and they have to be cheaper and more efficient than taxis, popular apps like Uber and Lyft, and vehicles for personal use.
Argo gave a taste of the types of projects it hopes to sponsor through this new partnership with CMU:
For example, how can autonomous vehicles "see" their surroundings and operate safely in bad weather, such as very heavy rainfall, falling snow and fog? How can we reduce or eliminate confidence in high definition maps without compromising on safety and performance? How can autonomous vehicles reason in very unstructured situations with bad traffic that are often found in some major international cities, where roadside actors completely ignore road rules? How can we reduce our need for labor-intensive High Definition map data when moving to new cities? When vehicle fleets are deployed, how can we efficiently use the experiences of an autonomous fleet to ultimately achieve exponential improvements that go beyond the original launch possibilities?
Argo released it last week Argoverse, its HD card dataset. By making datasets such as these available to the research community free of charge "the performance of different (machine learning – deep net) approaches can be compared to solve the" same "problem," said Raj Rajumar, a professor of electrical and computer technology to Carnegie Mellon University, which is not affiliated with Argo. "In other words, they offer a kind of standard benchmark."
Argo is not the only company that focuses on supporting the research community. Last year, Intel launched the Institute for Automated Mobility in the driverless testing hotbed of Phoenix, Arizona. The institute combines the three state universities with the state departments of Transportation, Public Safety and Commerce and the companies working on automated cars, trucks and drones. Intel has not announced its financial involvement with the institute.