New Details About Fatal Tesla Crash Emerge

[Onkel Neal]

http://gas2.org/2016/07/02/new-details-fatal-tesla-crash-emerge/

A woman driving on the same highway and in the same direction as Brown claims she was driving 85 mph when Brown’s Model S flew by her at a high rate of speed. That information is not included in the official traffic accident report filed by the Florida State Police, but is surely something known to Tesla, as it has access to all of the data stored in the car’s computer system. It is included in a story at Teslarati.

The driver of the tractor trailer, Frank Baressi, age 62, told the Dallas Morning News in a telephone interview that the Tesla driver was “playing Harry Potter on the TV screen.” He added, “It was still playing when he died.” Baressi said “he went so fast through my trailer I didn’t see him.” He didn’t see the video playing, but claims he could hear the sound track still playing when the Tesla finally came to a stop several hundred yards up the road. Tesla Motors says it is not possible to watch videos on the Model S touch screen

Did I miss something, wth is "autopilot" and why is it legal in a passenger car? :down:

[August]

And WTH kind of autopilot system requires the user to keep his hands on the wheel at all times? If I have to do that I might as well do the driving.

My guess is the victim had fallen asleep which is going to be a much bigger problem than the proponents of "driverless" cars are currently admitting.

[Catfish]

The BMW and Audi brands have a similar system, but it is intentionally not that autonomous. They already keep the distace to a car ahead automatically, and keep the car between the white lines, helped by GPS. They also have a kind of automatic braking system and lots of other little electronic helpers, that work for appx. three years from new.
Seems to be all the rage, everyone buys it and pays lots of money for that.
Just like in the new planes like Airbus and Boeing i might say.

I guess i'd never buy a used car, no one knows which of the electronic gadgets still work, and the repair shops are unable to cope with system errors.

[Platapus]

I don't think these cars should be allowed to engage any type of "autopilot" until the technology is mature, tested, and certified to be safe.

In this mode, this car was, in my opinion, essentially an experimental vehicle being driven uncontrolled on a public road.

What was the speed limit on the road? I would imagine that the first thing any type of autopilot would do is not exceed the speed limit.

Interesting technology but way too immature to be used on public streets.

[Torplexed]

AP & Reuters reported that the driver had a portable DVD player with him and was watching a Harry Potter movie. After passing under the truck the vehicle traveled another quarter mile before hitting a telephone pole. The driver had 8 speeding tickets. Friends describe him as fearless and a speed demon.

Apparently, this guy thought he had an autonomous car already and was "driving" like he did. You do get the impression that a lot of people out there are anxious to shed all personal initiative and responsibility even before the technology that they want to surrender it to is mature.

[Skybird]

What if Carl Benz 150 years ago would have bee confronted with the calculation of how many traffic kills his invention would score per anno, worldwide (far over one million, says the WHO, btw), and if somebody would have calculated for him how his invention would contribute to the emmission of climate-critical pollution? We probably would have no cars today, for his invention would have been stopped-to-death in its infant shoes. This is how a German TV anchorm commented on the issue yesterday.

Tesla says that statisctically every 150 million kilometers driven in a normal car somebody dies due to traffic accident cause dby car. The Tesla model now had already driven 225 million km.

We also know how notoriously irresponsible and incompetent especially young male drivers in the age group 18-25 can be if you let them take control of a wheel. Lets face it: a good ammount of young men from that age group should be banned from driving until they reach the mid-20s.

Finally we know from other transportation, espoecially train, that the majority of accidents get caused not due to failing technology, but failing humans.

Personally I feel little sympathy for automatic car driving - but if it is technologically possible, the development has plotted course for it in the future for sure. Its always like this with technology. What can be done, will be done.

Still I plead to leave this accident in relation to all car traffic there is - and the accident numbers that traditional carcv traffic causes: and almost always due to human error. I wonder how the description of the killed Tesla driver fits into that last sentence again.

[Oberon]

Tesla autopilot helps save a mans life:

http://www.bbc.co.uk/newsbeat/article/37009696/tesla-car-drives-owner-to-hospital-after-he-suffers-pulmonary-embolism

[Bilge_Rat]

I have been driving over 40 years and have seen way too many crazy/moronic/stupid/distracted/enraged/dangerous/blind drivers to ever, ever put my faith in an automatic system.

I never text while driving, concentrate 100% of my attention on the road around me. I even turn off the radio when approaching a particularly tricky merge in my daily commute to maintain full concentration (anyone else do that?).

[Buddahaid]

Yes, only I rarely listen to anything save a ballgame.

[Skybird]

There was an essay in German media recently, raising an interesting issue. If cars drive autonomously, and end up in a situation where an accident under whatever circumstances is inevitable, may it be due to the autonomous car, may it be to another car, bike, bicycle or pedestrian - which human should the car decide to kill then?

[Oberon]

I have been driving over 40 years and have seen way too many crazy/moronic/stupid/distracted/enraged/dangerous/blind drivers to ever, ever put my faith in an automatic system.

In a way, that's one reason to put more faith in an automatic system, since a computer isn't moronic (mostly), doesn't get distracted or enraged (yet) and isn't blind, so as long as the computer is left to its own devices and isn't interfered with by the user it should deliver the driver safely to their destination without incident. The machine has reflexes that we can only dream of, and reaction times to match. Even at your most aware and attentive, you cannot match the system in something like a Google car, the human body isn't built that way.

There was an essay in German media recently, raising an interesting issue. If cars drive autonomously, and end up in a situation where an accident under whatever circumstances is inevitable, may it be due to the autonomous car, may it be to another car, bike, bicycle or pedestrian - which human should the car decide to kill then?


That is a question that many people have been asking since autonomous cars have started to become a thing. The honest answer is that it will do what it is programmed to do, but at this point in time, no-one knows what to program it to do. I imagine right now it would probably crash both figuratively and literally trying to decide what to do, but it's a problem that is going to have to be addressed, preferably by a consensus.
Reducing the chance of that no-win scenario happening is probably the best thing we can do, and that's really up to removing as much human involvement in the process as possible, but ultimately that's only do-able to a point and sooner or later, a Tesla car is going to have to face its Kobiyashi Maru.

[Buddahaid]

I wonder just how true that is. How far ahead will the computer sense? Will it be able to anticipate a risky situation developing based on experience or just be quick to respond when it's too late.

[Mike Abberton]

Far enough down the road (pun intended), one idea is that all the cars would be communicating with each other in some sort of radio link. That way when one car starts to stop (let's say because of an unanticipated obstruction), all the cars behind it can start to stop before they (or a driver) could even sense the problem. Even with human drivers, such a system could propagate a warning back up the highway for all the other cars to slow down. Such a system also has the potential to ease traffic congestion in high-travel areas as well (assuming most/all of the cars are equipped with the system).

The whole "how an automated car responds to a car-bike-bicycle-pedestrian dilemma" question is pure baiting. How does a human pick in the same situation? The implication is that the human would make a "better" choice, but would they? An equally relevant (probably more relevant really) is how much better than a human is an automated car at avoiding those damned-if-you-do type of situations altogether. Studies have already shown that the assisted-emergency-braking systems are waaay better at stopping a car in an emergency than a human driver is.

Mike

[Oberon]

I wonder just how true that is. How far ahead will the computer sense? Will it be able to anticipate a risky situation developing based on experience or just be quick to respond when it's too late.

A lot of that would depend on sensors, now if you were to perhaps have each car ping its current location, speed and travelling direction to an orbitting satellite, then some sort of traffic control system could be created in which each car would know where the other cars in both its immediate and non-immediate surroundings are, this would help to lower congestion by keeping traffic flowing at a smoother rate (much as a similar system in London on the buses attempts to avoid bunching by keeping tabs on where the buses are and telling drivers to slow down if they're getting too close to the bus in front of them on the timetable), thus it would be able to direct traffic in a manner that would try to avoid the increase of major accidents taking place. Of course, that wouldn't do much to prevent immediate close range problems, but that's where the sensors on the car would come in...and, thinking about it, as time goes on then the system will learn about particular blackspots and trouble areas and modify the behaviour of the cars to cope with this. Is a certain place known for the cars having to slam their brakes on because of pedestrians crossing the road at a certain time of day? Then the system will adapt by telling the cars to travel down that road slightly slower or just avoid it at that time of day.
Of course, this is probably at least twenty to fifty years in the future, and there will still be accidents, it's a statistical certainty, and there will be quite a few instances where the cars will unfairly get the blame for it, such as in the case of the OP, who is at fault, the car for having the auto-pilot, or the human for abusing the system? Ultimately though, when it comes to human drivers and computer drivers, we're the most dangerous one.

[Skybird]

http://moralmachine.mit.edu/

What kind of killer are you? Any preferences for your ideal victims? :)

I have a principle problem with that "test", for it interprets meanings into details that I did not even care for when running it. But okay. Lets talk about your psychopathy, not mine. LOL

Mostly I seem to comply with a decision-making behaviour that is more prominent in the Anglosaxon world than in that of the continental European West: mostly I decide by mathematics, means: number of victims. I weighted also age and family status. When I think about it, I always have ticked this way, all my life. Legality, social importance , ethnicity/race did not even catch my attention - I was surprised to see them being included in the final "verdict". Maybe the mention these details in the instruction - which i did not care to read. :D

Ethics vary between US and England on the one side, and the rest of the West on the other, the Anglosaxon world by trends is closer to what is called utilitarianism. Germany for example, the Zeitgeist of its society and population, I would expect to be found ticking quite differently from that.

The research done in that MIT project is done in connection with the so called trolley-problem, which is examined in moral philosophy as well as social psychology and related branches since the 50s. Its gaining importance in computer sciences and AI development.