Toyota: We’ve Only Solved the “Easy Problems” When it Comes to Autonomous Driving
The 2016 International Consumer Electronics Show (CES) in Las Vegas, Nevada, is all about how software and technology are changing everyday life. But the conference has grown to focus increasingly on cars and automotive technology in recent years, and 2016’s show has already seen major announcements from startups like Faraday Future, and a new partnership between Lyft and GM.
The subtext is that the future of the automobile may not be in the hands of the traditional automakers, which are facing fierce competition from upstarts like Tesla Motors, and tech giants like Google and Apple. Against this backdrop, Toyota hosted a press briefing to discuss its $1 billion investment into the Toyota Research Institute, which is building new offices in Palo Alto, California, and Cambridge, Massachusetts, within walking distance of Stanford University and the Massachusetts Institute of Technology, respectively. Toyota is partnering with leading academics as part of a plan to build cutting edge artificial intelligence systems to underpin its future autonomous vehicle offerings.
Between the two institutions, Toyota has initiated 30 individual research projects. At Stanford, one of them is the “Uncertainty on Uncertainty” project, which is pursuing artificial vehicle intelligence capable of understanding and responding to obstacles that the car’s programming has never seen before. Accordingly, when the car encounters such obstacles and responds, it needs to be able to tell the manufacturer and driver why it responded the way it did. That’s the purpose of “The Car Can Explain,” another project based at MIT, which ensures that cars can produce a post-mortem to explain any unexpected behavior or responses.
Much of what has been collectively accomplished has actually been relatively easy. And the reason for that is that most driving is easy. Where we really need autonomy to help us the most is when driving is not easy: We need to solve driving when it’s hard.
CEO of the Toyota Research Institute, Dr. Gill Pratt, who was once head of DARPA’s Robotics Challenge, argued that these projects are critical because what has already been achieved with autonomous vehicles is only the low hanging fruit. “Industry has made incredible strides in recent years, but we are a long way away from a future of fully autonomous cars,” Pratt stated. “These systems can only handle certain speeds, weather conditions, and traffic. Much of what has been collectively accomplished has actually been relatively easy. And the reason for that is that most driving is easy. Where we really need autonomy to help us the most is when driving is not easy: We need to solve driving when it’s hard. That’s what TRI is working to address.”
Pratt made clear that Toyota, as the world’s largest automaker, is the company best positioned to solve this challenge because of the number of miles driven by Toyota vehicles every day. “To achieve full mobility we need reliability that is a million times better than what has already been achieved. We need trillion mile mobility. There are roughly 1 million Toyota cars and trucks in service at any time around the world. They travel 1 trillion miles a year.”
The reference to one million autonomous miles driven is presumably a jab at Google, which has logged 1.3 million miles in its autonomous vehicles.
The emotion of autonomous
Pratt emphasized that the emerging autonomous vehicle space will be one that is largely unforgiving of accidents, so the company is striving to build an autonomous vehicle that is “incapable” of being in an accident. “Society tolerates a lot of human error. Around the world it’s about 1.2 million deaths a year from automobile accidents. We expect machines to be much better–ever ready and nearly perfect. This has to work at the trillion mile scale.”
The company is striving to build an autonomous vehicle that is “incapable” of being in an accident.
Passenger safety is, thus, a critical concern. Bob Carter, Senior Vice President of Automotive Operations for Toyota, stated, “Now, cars can see things and react quicker than humans–cars are becoming intelligent. We now see ourselves at a point where the most important focus of all is the driver-vehicle interface. But maybe we should be calling it the driver-vehicle relationship. Cars have emotional effects on people.”
Carter emphasized that this increasingly emotional and codependent relationship between humans and machines must be harnessed to maximize safety. “We are now capable of creating a true interrelationship between a driver and an intelligent vehicle that will save more and more lives on the highway. Like teammates, the intelligent car and driver are learning from each other. They watch, listen, and remember. They adapt, communicate, and assist when needed. Over time, a foundation of trust is built.”
Autonomous and alternative
In the 100 year history of the automobile, 2015 was a high water mark: A new record in which 17.5 million new vehicles were sold worldwide.
Pratt quipped, “Toyota is the world’s best manufacturing company, that just happens to make cars.”
Advancing human mobility means incorporating software and artificial intelligence, but it also means designing and manufacturing a tremendous number of cars that are up to the challenge. In another likely comment on the company’s advantage against tech companies reaching into the automotive space, Pratt quipped, “Toyota is the world’s best manufacturing company, that just happens to make cars.”
Additionally, Carter stressed that the car of the future will be autonomous, but will also run on hydrogen. “Just last year we announced that we would openly share all of our 5,600 patents related to hydrogen fuel cell vehicles. Wood was replaced by coal, which was replaced with petroleum, which we believe will be replaced by hydrogen.” Accordingly, the Toyota Research Institute, which officially opens later this month, will have a significant emphasis on material science in addition to artificial intelligence, in order to build stronger and lighter vehicles, and to bolster fuel cell performance. Toyota’s Prius hybrid and plug-in vehicles were not mentioned.
Finally, Toyota’s research is focused on expanding indoor as well as outdoor mobility, and the company is developing a range of home robots, which it says, “Could soon be prized as much in the future as cars were in the past.”
Originally published at energyfuse.org on January 5, 2016.
