Cruise Begins Driverless Testing In San Francisco (techcrunch.com) 39
Cruise Automation, the autonomous vehicle subsidiary of GM that also has backing from SoftBank Vision Fund, Honda and T. Rowe Price & Associates, has started testing what it describes as fully driverless vehicles on public roads in San Francisco, the first milestone required to secure a permit to launch a shared, commercial service that can charge for rides. TechCrunch reports: The company released Wednesday a video of its first ride -- with no human safety driver behind the wheel -- in the Sunset neighborhood in San Francisco. There was a human safety operator in the vehicle, sitting in the passenger seat, the video shows. Cruise's testing of fully autonomous vehicles is in limited geographic area and in arguably one of San Francisco's simpler environments; the video below shows the testing was conducted at night in a less congested part of San Francisco. However, it still marks progress by the company that had once aimed to launch a commercial service by the end of 2019.
For some in the industry, the caveats of having a safety operator in the passenger seat and launching in an "easier" and small geofenced area matter. Cruise says this is just the beginning and that it will eventually expand its driverless testing area, adding in more complicated environments over time, as well as removing the safety operator from the vehicle. "We recognize this is both a trust race as well as a tech race," Cruise spokesperson Milin Mehta said in an email. "Given that, during the beginning of our use of this permit, we will maintain a safety operator in the passenger seat. The safety operator has the ability to bring the vehicle to a stop in the event of an emergency, but does not have access to standard driver controls. Eventually, this safety operator will be fully removed."
For some in the industry, the caveats of having a safety operator in the passenger seat and launching in an "easier" and small geofenced area matter. Cruise says this is just the beginning and that it will eventually expand its driverless testing area, adding in more complicated environments over time, as well as removing the safety operator from the vehicle. "We recognize this is both a trust race as well as a tech race," Cruise spokesperson Milin Mehta said in an email. "Given that, during the beginning of our use of this permit, we will maintain a safety operator in the passenger seat. The safety operator has the ability to bring the vehicle to a stop in the event of an emergency, but does not have access to standard driver controls. Eventually, this safety operator will be fully removed."
21st Century Road Rage (Score:2)
"Damn this guy is going too damn slow!" (passes car and rolls down window to start yelling expletives) "Hey asshol..." (finds driverless car)...
Yeah, road rage is going to get interesting for a while with all these law-abiding "drivers". The COVID speed limit on my local freeway went from 70 to about 85...
Re: (Score:2)
On a related but different note, I wonder if deploying large numbers of these as police cruisers with dark tinted windows that make it difficult or impossible to determine whether there is an officer inside or not would help to reduce crime?
Re: (Score:2)
Why stop there? Force the car to pull over, cut the LTE antennas and then send it to a chop shop.
Re: (Score:1)
Re: 21st Century Road Rage (Score:2)
No need to open fire, just place something on the road in its path.
Re: (Score:2)
Re: (Score:2)
If a gang of thugs wants to carjack an SDC, all they'll have to do is stand in it's way and it'll come to a complete stop.
If a gang wants to steal stopped cars, there are already plenty of parking lots full of them. There are also cars sitting in driveways and parked on the street. Unlike the SDCs, these cars don't have cameras recording everything around them.
Re: (Score:2)
Re: (Score:2)
I specifically was talking about carjacking.
Do you really think an SDC will go where a random unidentified person tells it to go?
Re: (Score:2)
Re: (Score:2)
and how funny will it be for the other cyclist who runs into the suddenly stopping car?
Re: (Score:2)
Re: (Score:2)
The fact that these cars are full of cameras with logs will by itself reduce nearby crime. The odds of getting caught doing something illegal on the road, or even outside a business next to the road, will rise dramatically once camera cars proliferate.
Re: (Score:2)
One of the side effect of driver less cars especially ones without passengers in them would be the increased abuse they will take, from other people.
While I am not a large all person myself, while sitting in a seat of a car with my broad shoulders, I look like a much bigger and stronger person than I really am, while seated in a car. So there are many times where people are aggressive driving around me in my little Prius C, then when they see who is at the wheel, they normally back off. I rarely ever have
Re: (Score:1)
Re: (Score:2)
Re: (Score:2)
Re: (Score:1)
Yip, it's only misdemeanor to beat the shit out of a bot. And bot haters will cheer their videos, making them Youtube heroes.
Wow the name/word Cruise (Score:1)
When I read the headline, and saw Cruise, I thought Tom Cruise. :) I was going to say cue the song "Right into the Danger Zone" by Kenny Loggins. I had to read the article synopsis to see it was a tech venture. D'oh!
But soon we'll have a "Johnny Cab" alluding to the movie "Total Recall" based upon a novel by Philip K. Dick.
Re: (Score:2)
Re: (Score:1)
Driverless, well no drunk drivers. More programatic errors. Headless, well would be a perfect Hallowe'en marketing pitch.
I take it you like the Mikkado? I like the scene in the operetta when the Lord High Executioner arrives.
Where's the common sense? (Score:4, Interesting)
You mean that professionals that have spent a career dealing with bugs, careless code, outsourcing, patches, and arbitrary deadlines in computer code are skeptical about what is arguably bleeding edge technology? Are we supposed to think that after years of buggy code in cars that somehow automated cruise control features are somehow immune to these things?
Are we to think that computer code is somehow special just because it is driving a several thousand pound object in traffic with real people in it and around it? Are we supposed to forget that human life is always the greatest risk factor to consider? Does the author of this story spend any time talking to someone in technology that isn't working in the auto field? So many questions, so little common sense.
Re: (Score:2)
What it means is they matter because the gap between working 99% of the time and needing a safety driver, and working 99.9999% of the time and not needing a safety driver is more a giant chasm than a short hop.
If a company is at the 99% level they are years away from having a self driving vehicle at best, and their technology might not even pan out. If they are running a driverless taxi service and haven't been destroyed by liability lawsuits then maybe they will have a viable product soon.
That is why peopl
Re: (Score:2)
The fact that the vast majority of the world still hasn't moved on to driverless trains shows just how hard of a problem this is to solve. A lot of trains are are able to self drive, but even just for public perception they choose to put somebody in the driver's seat. I can imagine a lot of people taking up issue with having driverless cars on public roads.
Re:Where's the common sense? (Score:4, Insightful)
Re: (Score:1)
Take the number of daily trips made by automobile. Take that number and calculate to see how many trips are left over after 99.9999%. This may or may not be better than human drivers. At some point this will likely favor the computer.
However you have to consider things like weather. Automatic driving systems use queues like road lines or curbs to know where their lanes are. This works great in San Francisco, not so well in a Minnesota or Scandinavian road in the winter. Following GPS lanes is also not neces
Re: (Score:1)
I think driving in San Francisco or any dense urban environment (maybe not out where Cruise is, but major downtowns, narrow side streets, etc) is a harder problem. Dealing with skids from ice is a control problem that these types of systems could deal with very well, and there's been a number of improvements in the HW to deal with rain/snow. These things aren't relying on lane markers since in a lot of places they are inconsistent anyways, forget about the weather. There's other features humans use to
Re: (Score:1)
Just curious, have you got experience driving in winter conditions, especially in bad weather? I can assure that it is much more difficult than simply driving in heavy downtown traffic in a major city at rush hour. I say that having driven in rush hour traffic of pretty much any US city you can think of. Itâ(TM)s a completely different set of skills and things that you can do in warmer weather can kill you in the cold
Re: (Score:1)
Yes, and no I don't think it's more difficult as a control problem. You have visibility issues, which are different than loss of control from ice/snow on the road. But automated systems are very good at dealing with things like loss of traction (as much as they can be dealt with). Its a simpler problem than understanding human behavior - is that guy looking, is he going to cross, should I go or let the other person go, etc.
I will say from a novice driver's point of view, driving on snow is going to be mo
Re: (Score:1)
I agree it's harder, but do you have specific examples we can review? Being able to deal with sliding tires could be one, but maybe self-driving cars should have chains if the roads are slippery.
Re: (Score:2)
You mean that professionals that have spent a career dealing with bugs, careless code, outsourcing, patches, and arbitrary deadlines in computer code are skeptical about what is arguably bleeding edge technology? Are we supposed to think that after years of buggy code in cars that somehow automated cruise control features are somehow immune to these things?
Are we to think that computer code is somehow special just because it is driving a several thousand pound object in traffic with real people in it and around it? Are we supposed to forget that human life is always the greatest risk factor to consider? Does the author of this story spend any time talking to someone in technology that isn't working in the auto field? So many questions, so little common sense.
All excellent points, but, the common sense question is not how safe can the coding be made. Rather, can the computer coding be made safer than having live human beings controlling the same objects? Given the rather large numbers of bad drivers on the road today, I suspect that may not be a difficult standard to meet.
It will take a generation or so to move past the human driver experience. In the meantime, it's a very safe bet the general public will insist on judging mishaps with AI controlled vehicles t
Re: (Score:1)
Iâ(TM)m all for supporting an increase in driver training standards as well. Iâ(TM)m inclined to look at Germany for their standards on driving licenses and auto maintenance. That doesnâ(TM)t change or devalue the other concerns with automated driving.
Re: (Score:2)
testing/sf/homeless (Score:1)
Is testing these in SF really just an attempt to deal with the "homelessness" problem?
Re: (Score:1)
No, it just changes it from a 3D problem into a 2D problem.
Faking their way through 'testing'. (Score:2)
Can we get it to fake it's way through enough to convince technology-ignorant politicians to allow us to sell these deathtraps to the public?
No worries our legal department says we can tie up the wrongful death lawsuits in the courts until the plaintiffs give up for lack of money, LOL.