Mark Fields, President and CEO, Ford Motor Company:
We’re announcing our intent to have fully autonomous vehicles in commercial operation for a ride-hailing or ride-sharing service beginning in 2021. This is significant. Ford will be mass producing vehicles capable of driving fully autonomously within five years. No steering wheel. No gas pedals. No brake pedals. A driver will not be required.
If someone had told you 10 years ago — even five years ago — that a major American car company would announce the mass production of a vehicle with no steering wheel, you would have said they were crazy.
This seems like a remarkably ambitious timetable for such a vehicle. Would you buy or ride in one of these in five years, assuming Ford can pull it off (which I don’t think they’ll be able to)?
I’d trust Ford to make a 1993 Taurus.
5 years seems likely. Not like they are starting today – as with most others they have been working on it for several years.
Google basically have this already – so another five year of development sounds about right for it to be a mainstream brand thing.
I’d almost suggest they have tagged 2021 as a good amount of time for public acceptance and get the legislation in place.
Maybe. I’d personally wait a year or two just to see but I don’t have an issue with the concept. It’s the execution that needs to be nailed. As I approach 60 I’m more and more aware of how valuable things like this could be in maintaining independence without being isolated. My Mom in her later years grew more isolated because she couldn’t drive. The ability to call a car, not worry about the driver (she passed before Uber etc) and go somewhere would have been huge for her. It will be huge for the next generation.
The thing that is holding back these cars is lack of infrastructure. GPS is only accurate to 4 meters at best, which is not enough for autopilot cars. They need something like centimeter resolution. Until then, this is all just a kind of game by manufacturers, boasting that they are going to be the one to solve the autopilot car. But the reality is, the solution has to come from infrastructure, not one private company. Today, the car has to do a ton of guessing to make up for its inaccurate GPS, and that is not a long-term solution because some cars will guess better than others. In the same way that we have built roads and painted lines, we need to “paint” next-gen GPS “lines” for autopilot cars to follow. Then the technology can scale.
If you compare the very-successful iPhone to the not-as-successful Newton, the biggest difference is that iPhone had wireless Internet (2G/3G and Wi-Fi) to take advantage of. Without that, iPhone would have just been another PDA, an accessory for your phone and PC, instead of a replacement for your phone and PC. Today’s GPS is like the first-generation cell network from the 20th century. It’s not ready yet to support “the iPhone of cars.”
Even if GPS was millimetre accurate it wouldn’t be enough for an autonomous car. For example, it still couldn’t tell the car that a person has just walked into the road a few metres ahead. Or that the traffic lights have just gone red. Or a million other things like that.
“This seems like a remarkably ambitious timetable”
Yep. And hindsight’s the fool-proof method of sussing out PR offensive(s) from press releases.
See you in 2021.
It’ll be huge for me before too long, as well.
I imagine they will use GPS as a rough guide to onboard sensors and intelligence. To get somewhere the care needs to turn right at, say, 20th Street. GPS will indicate that 20th Street should be 45 meters ahead and sensors on board will detect the sidewalk, other cars, etc, taking the place of what you and I do now when the GPS nav says “Turn right on 20th Street…” – we see the cars ahead, etc and don’t just blindly keep going if we shouldn’t.
This will also take care of @nizy’s issue. The key will be getting the sensors to accurately detect things and the onboard systems to react correctly and in time. We need to get over the “but it’s not PERFECT” hump too. People aren’t perfect now, so the hurdle for autonomous systems should be “is it better (or as good as) a person?”.
Lights actually can be optically detected at least in the US. Red is always on top, green on bottom. That’s actually what allows people with red/green color blindness to safely drive.
I imagine this will be push/pull – a car CAN detect lights today but it would be better if lights emitted a signal indicating state that the car could pickup. So, some standard will emerge, get deployed over time, etc.