I would like to address this particular objection as it is one that pops up with regularity whenever this topic is discussed. As confidence in whatever system we develop grows, and it will, personal distrust will diminish. It’s a natural part of technology evolution. Your grandchildren will be amused, and maybe a little impressed, by the fact that you drove yourself everywhere.
you don’t need connectivity to have a reactive traffic light system. This should have been done years ago given the computing power we have to work with.
There are certainly some well organized traffic light systems in various cities but I think this is one of those areas where the Federal Government could spend tax money on something with a high ROI. “High tech” traffic light systems should be a priority because it saves time, money and the environment. I’m going out on a limb and suggest that if the technology isn’t already on the shelf it’s a small contract away from happening. It would add considerably to fuel economy across the board regardless of vehicle.
There is no excuse for driving from red light to red light to red light to red light to red light on a major road yet I’ve experienced this in cities large and small.
My 2C.
The problem with an autonomous system built into a car that allows for close quarter driving is that it can’t be used in all driving conditions. If people get use to seeing their vehicle driving in a linked manner to the car in front then they will not recognize this as a dangerous condition when operating the car themselves.
I don’t know if it can be done without connectivity or not but it is indeed an early target in the new connectivity standards.
Well it makes sense to have worldwide standards but it’s not necessary for traffic lights. Cameras are being used to replace magnetometers now so it’s something that we can be doing without the need for cars that communicate with the traffic light system.
That would be the dream, I think. At that point, cars become like mass transit. Just on-demand, personalized mass transit.
Well at least the news stories will finally be accurate. “SUV kills family of 4”.
Pretty amazing how quickly the technology is maturing.
My prediction is that once autonomous vehicles get underway (and I think it’s going to happen surprisingly soon) in urban areas the price of unmanned taxis will fall - and their convenience will rise - to the point where most urbanites won’t have a car.
The National Highway Traffic Safety Administration starts to take notice.
I predict the liability of an autonomous car will keep it off the street.
That’s not likely. I would buy an autonomous car if the cost of the driving unit was less than 5k over a similar car.
There are plenty of people just like me, and dollars will make it happen. That said, I think this is what the government should be subsidizing.
I don’t think you understood what I was referring to. It doesn’t matter what you want or how much it costs if the product is vulnerable to liability issues. If it crashes autonomously then the maker of the car is the driver.
An externally mounted speedometer always seemed like a good interim solution
http://www.google.je/patents?id=7C8lAAAAEBAJ&printsec=abstract&zoom=4#v=onepage&q&f=false
I’m not entirely sure you are wrong. However, almost all damage caused by motor vehicle crashes is currently borne through motorists’ insurance, the cost of which is spread across motorists through premiums. If part of that cost is instead built into the purchase price of a vehicle - on the basis the manufacturer will now often be liable - then so be it, it isn’t a new cost. It will just fall somewhere slightly different.
And given that the vast majority of MVA’s happen through human error, the accident rate may well go down dramatically with autonomous cars, and if so, so will the overall financial cost of accidents, whether borne by motorists or built into the price of vehicles.
Which brings the thread full round back to the claims of the quote in the op - the claimed potential for these systems to eventually improve safety to such a degree that car weights can even safely decrease.
Connected autonomous systems cannot completely eliminate things like trees falling down, or the actions of others who do not participate in or who disable the systems. But they can make it such that a claim of negligent driving, of being “at fault” in an accident, is highly unlikely to prevail. True, anything that can go wrong will so quality control and redundancy of safety systems will be vital and likely highly regulated, but while IANAL I would think that exposure in a rare case of an accident due to the “one in million” case of all systems failing at the same time (not due to manufacturing or design negligence) would be small, especially spread out over the complete system.
Again, the penetration of these systems is likely to be staged. The first applications will focus on ones at least risk - interstate highway platooning especially for trucking (really just a step up from cruise control), creeping along in heavy traffic jams, automatic parking in lots, information on suggested speed to hit all the lights, wider adoption of automatic safety systems that prevent accidents by having the car take over if one is about to occur (slowing down or braking if approaching another vehicle too quickly for the conditions, identifying the lane drift of someone falling asleep at the wheel and correcting the drift while also dinging an alarm, stopping a car backing up if something has entered the area, avoiding a collision by preventing a rapid lane change into another vehicle in the blind spot …). Then wider applications, greater intervehicle coordination, and increasing autonomy will take off.
Time course? Well ten years ago when I bought my last car I stated that my next car would be one I plugged in. Many of my family and friends scoffed at me. (Even knowing I tend to own cars longer than average.) Tomorrow I am going to order my C-Max Energi (and I’ll be declining the automatic parallel parking option) and once it arrives will indeed power all my daily commutes and a majority of my future yearly miles off the grid just plugging in at home at night. I today state my next car after this one will at least mostly drive itself. We’ll see.
motorist insurance only covers the risk of the owner. If the car is driving itself then the manufacturer and any intermediaries will take on additional liability. lawsuits will follow the deepest pocket.
You completely miss the point. All that would change is the path by which MVA liability is funded. Motorists would pay more insurance built into the price of a car and less insurance directly.
Overall, the total societal bill for MVA’s won’t go up merely because manufacturers get sued instead of motorists.
Every product is subject to liability issues; every innovation creates unknown exposures. The simple truth, as this expert analysis posits, is that “nobody knows anything” about how it will play out. He takes a very interesting historical perspective though and hazards some informed speculations. He concludes:
What we can say with some confidence is that the total number of accidents and road fatalities would decrease by something like one order of magnitude (given that over 90% of accidents are felt to be due to human error and negligence). And that there are models for encouraging actions that have societal benefit but that have potential individual company liability exposure. In my field the National Vaccine Injury Compensation Program comes to mind right off. Suits, usually without merit but still costly, against vaccine manufacturers for alleged harm from the old pertussis vaccine, were driving manufacturers out of the business. The National Childhood Vaccine Injury Act of 1986 responded to that by simultaneously creating the means to better monitor/enforce vaccine safety standards, a Federal “no-fault” compensation system for those whose injures were by a relaxed standard of evidence possibly caused or triggered by vaccines, a means for deciding such claims, and a funding mechanism by way of an excise tax. The system works.
Again no such system would be needed for those initial phases. Initial phases are more semi-autonomous driving systems that are safety innovations but that in no way abrogate the responsibility of the driver. New cars coming out with forward collision and lane departure warning systems, even when coupled with automatic braking, do not allow a driver defense that the car should have worked to avoid a collision, nor are those injured in an accident involving such a car likely to go after the manufacturer because the car still somehow managed to be in an accident. “Car trains” as envisioned by the SATRE project still involve a professional driver a the head of the queue. Even ones that I envision, ad hoc car trains of “connected” vehicles which also have their own semi-autonomous safety overrides, will have application only once in a highway condition and require active disengagement from the system and the queue, limiting any manufacturer exposures. Even traveling in a hyped up cruise control in which the vehicle maintains the lane and the safe distance as well as the speed for an individual vehicle and lower speed traffic jam creep, and automatic parking in lots will be low risk for serious accidents. Those sorts of applications will drive the initial roll out, innovations, and acceptance.
One other thought - manufacturer may want to build in a black box of sorts, saving all data from one minute prior to any collision, which could then be used to provide evidence of their vehicle’s lack of fault.
But even without I still aver that the safety, fuel saving (and commensurate decreased CO2 and increased energy independence), and decreased congestion societal advantages of even those initial roll-out applications should incentivize potential solutions to any imagined tort exposures.