CES sees Vegas traffic grind to a stop every single year. Unfortunately, 2020 has proven no different. Adding to the crush of regular cars, limos, buses and pickup trucks is a throng of autonomously driven ‘robotaxis’ shuffling the hordes around the city, with the ‘driver’ simply supervising proceedings. A cheeky trick when stuck in Vegas gridlock known to every Vegas cabby and late EFTM crews is to cut in front of these autonomous taxis: they stop and let you in every single time! They have to stop for you; the computer insists on it. But what if that wasn’t always the case?
Mobileye, an Intel company, has used CES 2020 to propose a fascinating idea. What if we can teach autonomous vehicles to drive in a way that is appropriate to their context?
Professor Amnon Shashua, senior vice president at Intel Corporation and president and chief executive officer of Mobileye, leads Intel’s advancement into the field of autonomous driving.
The professor explained that in Tel Aviv, for example, a very conservatively programmed autonomous vehicle simply won’t get anywhere. In this context, the autonomous car’s ‘hesitancy’ will be interpreted by other drivers as permission to cross it’s path.
Meanwhile, an aggressively dominant program would plainly be entirely inappropriate in suburban streets. Basically, it is a case of training the software to understand its context. I wonder if manufacturers could tweek this feature.
- BMW 3 Series: lane cutting, horn blowing and punching red lights.
- Toyota Camry: permanently in the fast lane 50kph below the speed limit.
- Ford Raptor: constantly maintaining a half metre gap to the car in front.
I think Intel is really onto something with this.
Seriously though, it is easy to forget that Intel do much more than just supply processors. Professor Shashua explained that Intel currently supplies the suppliers of the automotive brands we know and love; however, Intel is now proposing a complete suite of hardware and software to enable highly autonomous and fully autonomous driving solutions; a one-stop-shop for manufacturers who want to include autonomy functions in their vehicles. Intel identifies autonomous vehicles as an area for growth and wants a piece of what will no doubt be a huge pie.
Mobileye currently has L1/L2 ADAS solutions, such as autonomous braking, L2+/L2++ solutions, such as driver monitoring and driver redundancy and by 2022 sees themselves able to provide full autonomy solutions, including true robotaxis. By 2025, Mobileye hopes to offer a chauffeur function in passenger vehicles.
Years ago, CES was buzzing with the prospect of autonomous cars. More excitingly, now we are seeing the bugs being ironed out and meaningful solutions being presented. Exciting times.