BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

Toyota Invests $50 Million In Artificial Intelligence Research For Vehicle Robotics

This article is more than 8 years old.

Toyota just revealed the 2016 Prius and, based on an announcement last week, we know a bit more about what a fully self-driving Toyota will look like. But it probably won't appear or act anything like the new Prius, and also won't be available for quite some time.

While many automakers are competing to provide partially autonomous cars and Google is going for the moon-shot of fully autonomous vehicles without a steering wheel or gas and brake pedals, Toyota hasn't jumped into the self-driving car race. But that's doesn't mean the Japanese automaker is taking a backseat to Google and others, and in fact made a sizeable bet on automation last week.

Toyota announced that it's investing approximately $50 million over the next five years to establish a joint research initiative with Stanford and MIT to study artificial intelligence (AI) and its "application to intelligent vehicles and robotics." Toyota also announced the hiring of Dr. Gill Pratt, former program manager at the Defense Advanced Research Projects Agency (DARPA) and leader of its renowned Robotics Challenge events, to direct the AI research for the automaker.

I sat down with Pratt following a press conference in Palo Alto, California where Toyota announced the new Stanford/MIT research initiative and he explained that while automakers, suppliers and Google have made significant strides in self-driving technology, major challenges like initiating the "hand-off" between human and machine driving will require much more time and research.

"In the last five years, we’ve come a long way," Pratt said of self-driving technology. "But the important part is to understand that for fully autonomous driving – and even for intervention where you leave the person in control most of the time – getting to the last bit is the hardest part."

Pratt will coordinate research led by Daniela Rus, director of the MIT's Computer Science and Artificial Intelligence Laboratory, and by Fei-Fei Li, director of the Stanford Artificial Intelligence Laboratory. The university research teams will work with Toyota "to develop advanced intelligent systems to recognize, understand and act in complex traffic environments," the automaker said in a statement.

According to Pratt, one of the biggest hurdles is helping humans cope with the transition to robot cars. He noted that having the car drive on its own about "99.9 percent of the time is easy. The question is what do you do for that last [0.1] percent, where suddenly something completely unexpected happens? With a fully autonomous car, you have to handle this situation 100 percent of the time because you can’t expect a person to suddenly have a tenth of a second to figure out what’s going on and take over."

Pratt pointed out that "if there were no people involved and all the cars could be run by software, everything would be easy. But what we have is a whole bunch of cars and pedestrians and bicyclists and other things that are very unpredictable. So we need to build predictable [self-driving] cars."

Pratt said that the goal of the Toyota research is to "push the frontier on hard cases" such as how to deal with driver-machine hand-off. "That’s the challenge, and it involves as much understanding of human beings as it does computer science. I think it’s a very exciting problem and it’s one that’s very right for AI right now because AI systems can perceive the world in ways that just a few years ago they could not.

"The purpose of this whole effort with the universities is just the beginning," Pratt added. "In a few months, I’m expecting you’ll see a whole lot more from us."

You'll also likely be seeing more new Priuses on the road when the car becomes available in early 2016. But probably not fully self-driving Toyotas for a long time.