I was going to make the argument that most GPS and navigation software is absolute rubbish at navigation... but then again, so are most drivers.
I was going to make the argument that most GPS and navigation software is absolute rubbish at navigation... but then again, so are most drivers.
Ironically, I edited my post to address that before seeing thisOriginally Posted by BehindBlueI's
Basically, see the 2A analogy at the end.
I have not dealt with anything of this kind personally. However, a guy I used to work with now works for a reconstruction unit in my PD and he had one very recently where a Toyota (I don't know what kind) began to accelerate at random intervals due to a computer issue in its throttle by wire system. It was an older female operating the vehicle and she did not notice it becoming more pronounced, and eventually was trying to get off a highway when this issue kicked in and she hit about 60 on an off-ramp, rolling and flipping several times. She didn't make it, nor did the two pedestrians who were waiting for the light to turn green (both were runners and had their ear phones in...).
Anyway, he pulled the ECU out and ran it through a bunch of programs his unit uses to figure out what happened, he kept seeing this second "ACCEL" data point which he then contacted Toyota about in reference to explaining the read outs, and they said that that second "ACCEL" is a non-user (operator) generated factor and it should not be there (normally only found in testing of the engines via computer before they are installed or diagnosing an issue at the factory for factory purposes).
He went back and saw that it would randomly appear in the last few weeks and he associated it with speed increased even when the "BRKE" data point was executed by the operator. He ruled it an accidental and passed everything to the DA's office who probably made it available to the families, would not be surprised if we see this on the news in a "Toyota SUED for KILLING a driver!" type headline.
There should be strict liability for those providing the vehicles and those operating them with special insurance provisions.
-All views expressed are those of the author and do not reflect those of the author's employer-
Everyone should drive manual (standard) transmission cars and trucks with manual windows and a basic stereo.
The biggest thing to me with accidents is road design and how generally shit it is in the US, there are some civil/traffic engineers I would like to slap. When you have intersecting roads that have multiple lanes and directions of travel accidents will occur. Same with 24 foot wide roads with two lanes flowing opposite directions and no buffer, stuff is going to happen no matter computer or human in my opinion. I think we should look more into the infrastructure of our road systems but that cost money that governments and tax payers don't want to spend, but new cars with all the gadgets and features people will get a 10 year 50% interest loan on so there is that. The feat of re working and zoning roadways is also extremely immense and complicated.
I'm not LE, but I'm going to break protocol and suggest that the mature version of this tech will make dickheads safer and self-reliant, responsible people less so.
Put another way, the 70 percent of at fault drivers who also fail to carry insurance will be a lot better off in public Google cars (along with everyone else as career fools get cyber-nannies), and folks who've driven for 35 years with no at-fault accidents, no DUIs, and who maybe have an old motorcycle safety course or .mil driving training under their belts will see a reduction as a few of them watch in horror as the computer charges into a situation that they saw coming through the glass.
But who knows. Car and Driver mag once wrote about seat belt laws making the most responsible people (the ones already buckling up) less safe as the belts became a comfort compromise, but we also later got airbags, so maybe it'll eventually pan out.
Now that's an idea!
212
What really baked my noodle was the question of what does the computer-car do when its options are either to crash and perhaps kill the driver or crash into a crowd of pedestrians?
Also, from what I have read, the vast majority of the computer-car crashes are attributed to it not being able to negotiate the bad move by the human driver. If the rules are followed explicitly, the computer-cars do even better.
The rest of it is simple math to me. Once computer-cars start beating actual drivers by a country mile (and they are already close to doing so), you roll the dice and take your chances, just like every other time you get behind the wheel.
It consults with its internal AI simulation of Isaac Asimov's ghost.
What do most people do when faced with this decision? I doubt that it involves rational thought processes. Chances are a computer car wouldn't do the stupid crap that put the car in a position to crash into a crowd of pedestrians to begin with. A computer car isn't going to get drunk, smoke K2, be elderly and confuse the pedals, flee the scene of another crash, or show off and losing control of a burnout.
I've lost count of the number of wrecks I've worked where people have swerved to avoid hitting a wayward orange construction barrel, basically a 35 gallon Rubbermaid trash can, and ultimately lost control and slammed into a retaining wall.
Personally if it gets to the point that a car is literally driving me around, I'll take the just bus or train.
Reply With Quote
