Designing and building cars that drive themselves will require a massive shift not only in road legalisation, but vehicular design practices.
That’s according to Maarten Sierhuis, director of Nissan’s Silicon Valley research centre in Sunnyvale, California, who told AutomotiveNews (subscription required) that automakers need to focus less on hardware and more on software development in order to bring about the dawn of the robotic car.
Speaking yesterday at a the Automated Vehicles Symposium in San Francisco, Sierhuis said automakers need to design self-driving cars that can not only understand how humans drive, but also mimic it. Only then, he says, will autonomous cars be able to coexist with other road users.
“What the auto industry has to come to is a shift from thinking about the car as a physical, mechanical system,” Sierhuis explained. “Autonomy, autonomous systems, is about understanding how humans do that, and then replicating it with software.”
And that, says the former NASA software engineer, means more than just today’s smattering of passive safety technologies like Lane Assist, Adaptive Cruise Control, and Emergency Brake Assist.
While Nissan has already developed a self-driving Nissan LEAF prototype capable of tackling typical highway routes, Sierhuis says the key to developing autonomous driving further is developing computer software that can assimilate and process data like a human.
On first glance, that may seem like a counter-intuitive measure. After all, one of the advantages to autonomous driving systems is that they’re not human. But think about driving for a second, and it’s clear that humans do have some tricks that computer systems need to learn.
Be it a temporary change in route or an emergency situation, human drivers are are able to receive a whole range of sensory data simultaneously and use it to make the best possible decision given the circumstances. Moreover, they can use past experience from similar situations to figure out what to do next, even if they’ve never encountered this particular scenario on a particular piece of road.
Essentially, autonomous driving software needs to not only understand how to treat each piece of data from its sensors, but the context behind that data. While that challenge isn’t exactly new — software giant Google has been working on autonomous driving technology for years — it’s certainly a big one.
“It’s a matter of also understanding the roads, understanding the situation, understanding other objects and knowing what to do with that information,” said Sierhuis. “To plan your path around it needs deliberation — A.I. thinking.”
In other words, autonomous driving systems not only need to be able to follow a set of hard and fast rules, but also to be able to learn and adapt as new situations present themselves.
To develop these kind of systems, Sierhuis says, automakers need to stop thinking so much about mechanical, physical engineering, and more about software processes.
That’s something Nissan is already doing, Sierhuis admits. In fact, the Japanese automaker is already mid-way through developing a special computer system that can interpret and react to data in a very human way.
Combine this quasi-human way of thinking with the ability to communicate with other autonomous vehicles via interconnected wireless networking and the Nissan’s autonomous car of the future could not only help eliminate congestion, pollution and accidents but also cut commute times dramatically.
Would you want to own an autonomous car capable of learning new things through artificial intelligence processes, or does it just sound a little too much like HAL9000?
Leave your thoughts in the Comments below.
You can also support us directly as a monthly supporting member by visiting Patreon.com.