Alphabet's Waymo sharpens self-driving car tech, expands testing lead over rivals like App...
Even as Apple tests remain shrouded in secrecy, prototype self-driving cars by Waymo -- formerly under Google -- are dramatically improving their skills, data from the California Department of Motor Vehicles revealed on Wednesday.
While Waymo's test fleet in the state drove 635,868 miles in 2016, 50 percent more than in 2015, safety-based disengagements fell from 0.8 per 1,000 miles to just 0.2, something highlighted by Waymo's head of self-driving technology, Dimitri Dolgov. The executive credited progress to a "more capable and mature" mix of hardware and software, and operating on "complex urban or suburban streets," helping to build experience dealing with complicated situations.
In all Waymo dealt with 124 disengagements. The company blamed most of these on "software glitches," but "unwanted maneuvers," "perception discrepancies," and "recklessly behaving road users" also played a part.
Crucially, in no case did Waymo cars crash or otherwise get into an accident.
Waymo is believed to be well ahead of its rivals in testing self-driving cars, having kickstarted the modern rush by showing the technology could work. The company is transitioning away from self-designed test vehicles and should soon deploy modified Chrysler Pacifica minivans.
Apple has expressed interest in testing a car on public roads, but is thought to have temporarily shelved the idea of designing its own vehicle until late 2017, if ever. The company could choose to partner with an existing automaker for its self-driving efforts, known as "Project Titan."
In the meantime Apple is thought to be testing systems in virtual reality, and experimenting with augmented reality for purposes like navigation.
While Waymo's test fleet in the state drove 635,868 miles in 2016, 50 percent more than in 2015, safety-based disengagements fell from 0.8 per 1,000 miles to just 0.2, something highlighted by Waymo's head of self-driving technology, Dimitri Dolgov. The executive credited progress to a "more capable and mature" mix of hardware and software, and operating on "complex urban or suburban streets," helping to build experience dealing with complicated situations.
In all Waymo dealt with 124 disengagements. The company blamed most of these on "software glitches," but "unwanted maneuvers," "perception discrepancies," and "recklessly behaving road users" also played a part.
Crucially, in no case did Waymo cars crash or otherwise get into an accident.
Waymo is believed to be well ahead of its rivals in testing self-driving cars, having kickstarted the modern rush by showing the technology could work. The company is transitioning away from self-designed test vehicles and should soon deploy modified Chrysler Pacifica minivans.
Apple has expressed interest in testing a car on public roads, but is thought to have temporarily shelved the idea of designing its own vehicle until late 2017, if ever. The company could choose to partner with an existing automaker for its self-driving efforts, known as "Project Titan."
In the meantime Apple is thought to be testing systems in virtual reality, and experimenting with augmented reality for purposes like navigation.
Comments
Self driving cars will still mostly transport one passenger per vehicle which isn't mass transit by any stretch of the imagination.
Self driving cars are computerized taxis.
I myself find the notion of self driving cars unsettling. There are a whole host of ethical and legal issues to still be worked out. Testing autonomous vehicles without passengers in semi-controlled conditions is one thing.
Riding in one as a passenger knowing that I am in the hands of another software engineer/programmer is another. The software will be built to the ethical standards of someone else. And how will the software react to a child running into the street chasing a ball vs. runaway grocery cart. In one case, I would accept the vehicle steering into a wall. In the other case, I would not. Unless the cart had a child in it.
While the technology is serious, the flagrant promotion of self driving technology by Google is frankly, quite silly. I don't want it. And neither do the vast majority of my friends and colleagues.
The only people who do want it? People who have no business operating a motor vehicle in the first place. And do we really want a blind, demented elderly person being transported as the only passenger in a self driving vehicle?
I certainly don't. And if you believe it should be the case, then I would invite you to take flight on a commercial airplane without human pilots.
We'll look back on posts like this in a few years chuckle at the quaint thinking it represents. Is "a few years" 3 year, 5 years, or 20 years? I don't know, but I'm pretty sure that in my lifetime I'll be saying "I remember when everyone drove and owned their own cars."
There are drivers running over kids in the streets every single day. The standard can't be "when will self-driving cars be PERFECT?" They are rapidly approaching the ability to being better and safer than a good human driver, and that progress isn't going to reverse itself.
Are you ok with blind, demented elderly people riding in taxis?
You're letting the fire in your eyes at any mention of Google blind you. Within 3-4 years self-driving cars will be on streets in a city near you, and sold by several of those real car companies. Some may well be running Google software making it possible. Count on it.
EDIT: Since you appear to be completely unaware of how many companies are committing to autonomous transport and the progress they've made perhaps it might be a good idea to read this relatively short article prior to more comments.
http://www.businessinsider.com/companies-making-driverless-cars-by-2020-2016-10/#tesla-is-aiming-to-have-its-driverless-technology-ready-by-2018-1
We don't disagree on either of those points. Although "always" is a long time.
The automobile manufacturers were also big on HCCI and all sorts of technologies that never became reality. Just because several companies are looking at it doesn't mean that it will be feasible or that it will sell.
Not that it matters because it doesn't. But self driving technology as envisioned by Google will not happen in 3 years. And I will be happy to report back on Feb 1, 2020 to remind you of that.
Auto pilot technology has been available on the major airliners for several years. Yet they all still have human pilots.
Why is that? Piloting an aircraft is a far more controllable activity than driving a car.
Self driving cars have huge legal and regulatory hurdles to overcome. To think that they will be commonplace in three years is a stretch.
If fully automated transporting technology is feasible then why would I waste time on a car? Maybe for short trips to town.
Having that kind of technology for an airplane, however, would be much nicer. A four hour drive could be reduced to a 1.5 hour flight. With a computerized pilot doing the flying.
The idea of allowing a computer to do the flying is ludicrous to the majority but somehow a road vehicle is different.
I do understand what you're saying. I don't see it working out in the way you are envisioning.
Most people just want to hop in the car and go. Not wait. It's why people own their own vehicles. And for short jaunts, I personally would never engage the automated function. Maybe on long freeway runs when I am tired. But the feature will come at a significant cost. I would rather have other features to be honest for the same price. And I will still own my own vehicle. I don't really want to share it. I don't want to wait. And I don't want to get into a vehicle where someone may have just upchucked their last meal.
I really don't see the technology as any major advance and I really don't see it changing much other than allowing someone to get a little shut eye on long freeway stretches. Or to text a friend. It is going to come at a cost. A cost that only a small segment of the population would be willing to bear. The elderly person who is faced with the prospect of losing his/her driver's license.
I am not knocking the technology. But in the human world, it isn't going to change much really.
How did that technology pan out? How about the Chrysler turbine program?
Engineers aren't exactly good prognosticators.
Frankly, I would rather have a vehicle that can crash into a concrete barrier with minimal damage than a self driving vehicle. Perhaps the engineers would be better using their limited resources elsewhere than a technology that really is limited in its usefulness.
https://ot.to
An automated transit vehicle company is demoing to my city. They're like mini buses that carry people around.
Trust me, I get your point, and I don't disagree. I average almost 20,000 miles a year and have for more than 25 years. I love to drive. I would never buy a self driving car. If fact, I don't think many people will. It will be a service and I think it will go similar to Apple Music. 20 years ago, you bought an album, 10 years ago you might buy an album, but you probably bought a bunch of singles from iTunes too. Now, you pay $10 a month and stream whatever you want. I think now, a family owns 2 cars and drives everywhere. 10 years from now, maybe you only need 1 car and can have that computerized taxi take you to work and back. 20 years, you'll pay $10,000/year and be thrilled you don't need all the hassle of cars anymore.
If harm comes to the person, who is liable? If I am too old to drive, a human driver is still far preferable.
Things sound great on the surface but then there are always those pesky details that cause problems.
Self driving technology is still of limited benefit. AR technology is quite helpful.
I just don't see the technology amounting to much really other than a curiosity. Like Google glass. Interesting but not anything that will change much.
Autopilot 1.0 reduced accidents by 40%