Now it's happened: First accident caused by an autonomous car.

But why did it accelerate into them?

Volvo ' s statement was that the human avoidance system , even if installed on this vehicle, could be overridden by aggressive acceleration as seen in the video. Although you may think you would only accelerate aggressively at a person if a self-righteous journalist was standing in front of you, Volvo's engineers have to consider the case of one of their upper middle class patrons finding themselves accidentally in the middle of a riot. Someone gets dragged out of a Volvo and beaten to death, it looks bad on tv.

fritzsch wrote: But why did it accelerate into them?

Based on my understanding of the story, it was because the person behind the wheel stepped on the gas, hard.

"Aggressive acceleration" could have been as simple as a wadded up floor mat.

It didn't ncessarily have to be either the autonomous car, nor the driver.

Doesn't matter. Still a black eye for Volvo (and still a variable that the autonomous car could not predict).

The variables are infinite. That's why it will be tough to bring to fruition, and why the predictions of a safer world for our children will likely fall flat.

That's OK. We the sheeple will still have our self-driving cars, and will not notice when we have to pay 4X for the purchase price, 3X for our insurance, and 2X for any repairs.

Car ownership will become a luxury. Like strawberries in "Soylent Green". Those that have cars will contract for them to be used continuously to generate revenue (like Uber), the rest of the sheeple will be content riding a crowded autonomous bus.

That's hilarious. Repairs will be less (no driver abuse, the new gear will be in well-protected locations, plus the cars will be EVs), insurance will be negligible compared to today's prices, and the car will cost only a little more, at most. Technology gets cheaper - if it didn't, why aren't we all driving Model Ts?

You are a dreamer.

I like that about you.

GameboyRMH wrote: Hahaha no. Autonomous cars will not only save hundreds of thousands of lives (and to the manufacturers, that means many potential lawsuits) by making crashes extremely rare, but they'll be valuable to other industries too, like the trucking and taxi industries who will be able to lay off drivers en masse. I think autonomous cars would have to kill at least as many people as human drivers before any manufacturer might think it's a bad idea.

Remarkable that there is even one person that thinks like this. Methinks your glasses are distorting your vision.

In reply to 1988RedT2:

You'd be silly to think outfits with lots of drivers aren't looking forward to that day.

Edit: left out obligatory Chinese Volvo Tienenmen Square Package joke.

In reply to GameboyRMH:

Honestly, I'd be perfectly happy driving a 1930's ford daily. Screw automation.

these videos remind me of the debut of ED209.

https://www.youtube.com/watch?v=A9l9wxGFl4k

From what I understand, the next generation of driver aids is being programed by having actual human beings drive fully instrumented cars in normal situations. The sensor data is then run through a neural network that learns how experienced human driver react to the data presented. In theory, more data means more a refine program.

As I write this, I wonder if maybe they should have extra ordinary talented drivers in the cars. That be an amusement park ride. Take a trip around Road Atlanta with the Lewis Hamilton program.

Keith Tanner wrote: Imagine the uproar on GRM if the various autonomous devices could NOT be overriden by the driver.

One of the things I like about my Volvo is that I can turn the DRLs off with a Konami Code-like manipulation of the ignition switch and headlight switch. So I did.

Datsun1500 wrote: People won't get on a train with no one at the controls, and those are on a track.

I think they do though, at airports between terminals. Might be at chicago I am thinking of

GameboyRMH wrote: Volvo's essentially set an industrial robot loose on the streets by offering this setup.

Any industrial robot with wheels that I've seen has tons of safety systems to not run anybody over. I once toured a GM assembly plant using them, if you got in the way it would stop and honk at you.

Read up on Volvo's history with autonomous driving cars. It's a remarkable tale of failure from a remarkably safety conscious company. They keep making accidental predators.

Wally wrote: You'd be silly to think outfits with lots of drivers aren't looking forward to that day.

I think the complete opposite is true. I think outfits with lots of drivers would want this less than anyone.

To a company, drivers are not machines. They are not steering wheel holders. They are the face of the business, and the front line sales people. Companies will become faceless machines, which is not a positive thing.

Additionally, transportation companies and delivery companies do not have the technical proficiency to manage this well. They will have to contract with technology companies, which represents a lack of control and a need to trust outsiders with components of the business which are mission critical. Again, not a good thing.

It is highly likely that automation will not lead to greater efficiency and lower labor costs and increased profits for existing companies. It is more likely that new companies better equipped and structured to handle the technology will move in, which will result in a loss of market share for existing companies. Google will replace Yellow Cab. Again, not a good thing.

Even big companies that are capable of handling the technology could lose market share. Amazon's delivery drone program will not expand opportunities for UPS.

I don't see it.

SVreX wrote:

Wally wrote: You'd be silly to think outfits with lots of drivers aren't looking forward to that day.

I think the complete opposite is true. I think outfits with lots of drivers would want this less than anyone. To a company, drivers are not machines. They are not steering wheel holders. They are the face of the business, and the front line sales people. Companies will become faceless machines, which is not a positive thing. Additionally, transportation companies and delivery companies do not have the technical proficiency to manage this well. They will have to contract with technology companies, which represents a lack of control and a need to trust outsiders with components of the business which are mission critical. Again, not a good thing. It is highly likely that automation will not lead to greater efficiency and lower labor costs and increased profits for existing companies. It is more likely that new companies better equipped and structured to handle the technology will move in, which will result in a loss of market share for existing companies. Google will replace Yellow Cab. Again, not a good thing. Even big companies that are capable of handling the technology could lose market share. Amazon's delivery drone program will not expand opportunities for UPS. I don't see it.

Guess who wins the post for "pure speculation presented as fact award" today?

In reply to z31maniac:

Umm, absolutely everything on this subject is pure speculation.

And your point is...?

In reply to z31maniac:

First two words of my post- "I think..."

Sorry you feel threatened by my opinion.

SVreX wrote: I think the complete opposite is true. I think outfits with lots of drivers would want this less than anyone. To a company, drivers are not machines. They are not steering wheel holders. They are the face of the business, and the front line sales people. Companies will become faceless machines, which is not a positive thing.

I think I'd like you as a boss, but the competition would see this attitude as a weakness to be exploited, because most companies have no problem at all being seen as faceless machines (although I agree it's not a positive thing). Some of them actively work towards it. Uber treats their drivers as easily replaceable steering wheel holders and cuts their pay "because we can." They're also getting involved in autonomous car development.

SVreX wrote: Additionally, transportation companies and delivery companies do not have the technical proficiency to manage this well. They will have to contract with technology companies, which represents a lack of control and a need to trust outsiders with components of the business which are mission critical. Again, not a good thing.

How is this different than their current relationship with existing automotive OEMs? They already rely on Delphi etc for ECUs and other electronics, to say nothing of the rest of their trucks and cabs. I'm surprised a lot of trucking companies even contract out vehicle tracking systems when it's something they could fairly easily do in-house.

SVreX wrote: It is highly likely that automation will not lead to greater efficiency and lower labor costs and increased profits for existing companies. It is more likely that new companies better equipped and structured to handle the technology will move in, which will result in a loss of market share for existing companies. Google will replace Yellow Cab. Again, not a good thing. Even big companies that are capable of handling the technology could lose market share. Amazon's delivery drone program will not expand opportunities for UPS.

You could very well be right, but to me this is inconsequential in the grander scheme of things - I don't see anything inherently wrong with new companies replacing older ones. The effect is the same whether the autonomous cabs and trucks of the future have a familiar logo or some new vowel-starved name on the side. Anyone who uses UPS would probably much rather have Amazon drones flying around - the UPS drones might just dive-bomb an attempted delivery notice at your front door without slowing down

GameboyRMH wrote:

Curmudgeon wrote: The story does prove that there are a bunch of pitfalls to autonomous cars far and above what we humans 'think' there are.

This isn't some exotic new problem, this is a dunce move that should have been obvious to anyone with two brain cells to rub together. You can order the car with the capability to park itself but WITHOUT the capability to see people? What could possibly go wrong? Also a car doesn't need to be as smart as a person to drive - it has to be about as smart as a computer-controlled competitor in a racing game. An early '90s PC could run several of them at once. The sensors are what need to catch up, in the game the "AI" gets neatly formatted information about everything around it, while the real world is a mess that needs to be made sense of first.

Thank you for making my point. At the current level of technology, AI simply cannot do that properly. Hell, a lot of people can't do that properly and it's generally accepted that people are still smarter than computers.

Google's true autonomous cars (as opposed to the retrofitted Lexus SUV's etc) won't have a steering wheel or pedals, i.e. no way for the meat inside to override it. They see this as good; I don't. Having seen the number of glitches possible in something so simple (comparatively speaking) as a laptop computer and the ease with which viruses can be introduced (and yes there's a couple of people who have cracked existing safeguards on car systems) I don't trust the completely autonomous car at all. As I said, I'm happy I won't be around to see them take over.

While I'm thinking about it, the Amtrak crash up near Philly shows a couple of problems, not the least of which is the difficulty and expense associated with making transportation systems safe. Sure, the crash appears to have been human error but I'm talking about the electronic safety override system being installed. It was only operational on one side, Amtrak cited expense and long lead times for equipment installation as the reasons it wasn't installed yet on the other side. There's one glaring flaw I see with the whole thing: the train depends on signals from trackside to make its decisions. All it needs to fail dramatically is a couple of bored kids with a battery powered Sawzall. But we are supposed to depend only on all this technology? I don't think so.

I'm still sorta hoping that an autonomous system does get a rollout before I croak. The reason: one day I'm pretty sure I'll be driving my impossibly old fashioned clunker down some remote forest road and come across someone whose autonomous car has just given up, it won't go at all because it doesn't see GPS signals or signals from roadside sensors. They won't be prepared for the weather or have a backup for them to manually operate the car or for that matter the basic knowledge of how to do such a thing. If I'm lucky it will be cold, rainy (or snowy) and I'll just laugh, tell the occupants that their technology got them into this, let's see it get them out.

In reply to GameboyRMH:

I didn't say it was a good or bad thing (one company phasing out when another steps in).

My comment was in response to Woody's, suggesting there are lots of existing companies with lots of drivers looking forward to autonomous cars.

I don't think so. I think many of them dread it.

We have people on this very board who employ drivers as a significant component of their business who could probably never consider transitioning to autonomous cars. If it happens, it will put them out of business.

In reply to GameboyRMH:

Regarding "faceless machines"...

I appreciate being seen as weak. But I really think companies value their employees more than many employees understand. They just suck at communicating it.

Companies need sales. It's their lifeblood.

If a driver is just a steering wheel holder, it is easy to replace him.

But if he is an intricate part of the sales force, it's a lot harder.

There are millions of receptionists across the country in love with their UPS guy. Same thing for FedEx.

Convenient stores depend on drivers not just for delivering, but for inventorying, and stocking too.

Uber passengers kind of like that their drivers are regular guys. If Uber can't see that, they have a big problem with their business model and management.

Anyone who has every worked in a shipping/ receiving department can tell you the value of a good driver, who gets out of the truck and drives their forklift to unload the truck when necessary.

My own little medical transport company was absolutely dependent on the personal touch of our drivers to hand the customers respectfully, and appropriately.

The list goes on...

I bet that within a generation we will have pilot-less cargo planes. Think about the pilot’s job – all the skill is in take-off and landings. The well paid pilot puts the cruise on (autopilot) after 10 mins in the air. You have guys flying sophisticated drones with advanced weapon systems all over the middle east from some base thousands of miles away already. UPS or other large cargo carriers can have an air crew do nothing but take-offs and landings all day long. One air crew that probably costs well over $600k in salary and benefits who are limited to the hours they can work due to fatigue from watching a plane fly itself would now take care of 30+ planes. Pretty tempting ROI.

BTW: it doesn’t matter how much a receptionist loves their UPS guy, the decision to use a carrier probably doesn’t fall to the person manning the front desk anymore.