A whole bunch of automakers are trying to make self-driving cars happen. GM has been giving it a red-hot go with Cruise, but the company’s autonomous taxis have been off the road for some time. Now, it seems the company is finally ready to resume testing in public.
Cruise’s issues date back to last October, when one of the company’s cars hit a pedestrian that was knocked into its path in San Francisco. The vehicle then tried to pull over, dragging the individual a further 20 feet, causing significant further injuries. The company’s license to operate in California was revoked amidst allegations it had tried to cover up the incident for several weeks. As reported by The Autopian, GM ended up stashing a bunch of Cruise vehicles at a plant in Michigan in the aftermath.
A full 8 months later, Cruise is ready to turn over a new leaf. It’s not returning to California; not yet, anyway. Instead, it will first resume operations in Phoenix, Arizona. Cruise vehicles will hit the roads on Tuesday.
Cruise is taking a gentle approach as its vehicles return to the streets. It won’t be going straight back to fully autonomous operations. Instead, it will employ safety drivers behind the wheel so they can take over in the event of any incidents.
This move makes sense – particularly as a trust-building measure, given the current level of public faith in the company’s vehicles. Cruise also highlighted its safety report form with which members of the public can submit feedback on any concerns regarding the driving of the company’s vehicles.
However, this decision also comes with limitations. The Cruise AV vehicles are based on the Chevy Bolt, and they have conventional controls so they can be operated by a human driver in the front seat. However, the Cruise Origin vehicles, which are more pod-like, do not have conventional driving controls like a steering wheel or pedals. These vehicles will have to remain off the roads until Cruise is ready to resume fully autonomous operations without safety drivers present.
In the lead-up to this decision, Cruise had been operating human-driven vehicles on the roads of Phoenix for data collection. In time, Cruise plans to expand operations to Scottsdale, Paradise Valley, Tempe, Mesa, Gilbert and Chandler, assuming its operations meet stringent safety benchmarks.
Ultimately, Cruise is still dealing with the fallout of the accident last October. People have always expected that autonomous vehicles will be a danger to the public; when that happened last year, those fears were realized. Overcoming that will take some time. Cruise executives will be hoping that the return to operations goes as smoothly as possible, allowing the company to fly under the radar for some time.
GM remains confident that Cruise can deliver autonomous driving technology, and is backing the company with billions of investment dollars. Ideally, it will continue towards that goal with less controversy from now on.
Image credits: Cruise, Nick Thomas
Phoenix has been a favorite testing location for years. We used to see autonomous Ubers, Cruise, Waymo, all testing there. It’s just a giant grid layout wise, almost no inclement weather, lower regulation, makes sense for them to test. I’ve been baffled by how quickly they thought they could move away from backup drivers though so good to see that back for sure, will be a very long time before I trust a fully autonomous vehicle.
What happened to the pedestrian? Did they get compensated? Did GM or tax pacover it? Is this big news that GM is again putting the public in danger and agreed to terms that were the only acceptable terms? TELL US MORE?
All the huge strides they’ve made in automated driving technology amount to nearly nothing when you really take an objective look at it. These automated cars can make safe trips during dry weather in limited geographic areas with close supervision. Unless someone does something unexpected. Unless there is a road closure. As long as the roads have the expected markings. Unless there are emergency vehicles present. As long as the street lights are functioning properly. Unless it snows. Unless it’s foggy. Unless it rains too hard. As long as it has internet connectivity… the list goes on. Meanwhile even the best hardware and software malfunctions from time to time. I had a Hyundai where the ABS only worked when it was warm and dry, i.e. the time when you were least likely to need ABS. Apparently the wheel speed sensor didn’t work when it got wet. I don’t expect technology companies to be better than that. I have a Ford Maverick. It’s built by a debatably respected car company that has been around for more than a century. 5 of the 6 recalls I have had on it are software issues. If they can’t program it to operate the tail lights correctly the first time, how much faith is it reasonable to have that they can ever program automated driving correctly?
If they hadn’t tried to hard to look good after the accident last year, they probably wouldn’t have had to go through all of this. But having a person in the driver’s seat is a good step back. I’m wondering why they couldn’t get approval for the Origins to drive around a tourist trap, like in Vegas.
So they’re on the road again, but not playing star again.
Would it be fair to say GM was able to Turn The Page?
Only if they operate on a long and lonesome highway, east of Omaha.
Self driving is just not going to happen. They can roll out any update they want to it and a new situation will arise that the robot car cannot figure out and another bad thing will happen. Having the safety driver present will help temporarily but the companies will get comfortable again and take the drivers away. As has been explained by the writers here, you just can’t account for every possible situation with software.
Not realistic.
Self-driving is coming, but it’s not coming nearly as fast as some who want it, and it won’t be nearly as profitable as anyone thinks it is.
Self-driving will start in city centers and limited access areas, and will expand from there. There may even be some dedicated limited access highways designated for it. It’ll expand beyond that, get pushed back, and then expand again.
But it will come, slowly and gradually. Too many people see the value of self-driving cars.
By the time it becomes accepted, it will be a commodity. That’s why I think that much of the money invested in it is foolish. It has to be a commodity to be broadly accepted, and once it’s a commodity, there’s no way to maximize profits through enforced scarcity. I expect it to be profitable, but not in a revolutionary way.
But it’s definitely coming.
I’m not convinced it’s as inevitable as you are, but I agree that there’s a huge disconnect in terms of the financial viability of the technology. The way money is being thrown at it today, by the time it becomes commercially viable it will have to be so obscenely expensive to justify the investment that no one will be able to afford it.
Exhibit no 1 is parked over there attached to the charging station.
Just because something is deployed doesn’t mean that it’s actually functional and practical to the point of adoptability. For example we’ve had turbine cars. Fully deployed, worked without issues. They never made it beyond the 1963 test fleet despite experimentation beginning in 1946 and continuing until 1980. Same with flying cars. There’ve been multiple deployed models, but only sold in single digit numbers. Same with amphibious cars. Same with modular body cars.
Self driving isn’t going to go beyond that stage of “being deployed”, because there are far more problems it has to solve compared to the problem it was designed as a solution for. That problem being that for some reason Americans hate rail.
If this “strict supervision” doesn’t include nuns with yardsticks it’s not enough.
From now on, all self-driving vehicles must have sensors that allow them to feel pain. And we’ll have to teach them shame (and Catholic guilt). It’s the only way we can effectively train them.
Got to go visit the penguin…
https://www.youtube.com/watch?v=sqkowVU5mZI
Stop trying to make
fetchself-driving happen.Comment of the Samuel Jackson-ing Year right here