squirtdad

except history shows the automotive business had to be legislated to add seatbelts and shoulder belts and airbags and reverse cameras as standard.

tyrion

Nevertheless, the company that makes a safer self-driving car will have a big business advantage over those that make less safe self-driving car.

A few bad accidents could put a company in this sector out of business, and everyone knows it.

squirtdad

that has not deterred elon musk who by all accounts has the least capable sensor and technology pack and is the most hypeing the capabilities and has had the most serious incidents

bottom line is it will take government intervention do define what is self driving and what is not and the details around the whole thing

history has shown time after time this is the case...pollution rules, mine safety, work safety, the list goes on

CliffordK

If Autonomous Vehicles become a big thing, then I would hope they come up with a good light/transponder system that they'll offer cyclists at a low cost.

Still, the cars will have to avoid kids, dogs, and deer, which may not have the transponders, but I'd be happy if they came out with a device that could substantially reduce an already low risk.

Road Fan

This makes sense - the radar was not pulsed. Pulsed radar directly reads back the distance to a metal object in front of it. Doppler, at least "pure" doppler directly reads back speed. I'm not sure either one is good at detecting pure transverse motion. Hybrid types of radar are now more common in industry, as are systems which blend radar systems with laser or passive optical sensors, and combine the data from multiple sensors to create a comprehensive picture of target motion. A fully-vetted International Standard defines the minimum performance required for detection of possible forward collision objects, and it did not specify what kind of sensor had to be used, rather they specified what the sensor or system of sensors had to detect and what the system had to accomplish. To determine what design, what kinds of sensors are to be used to satisfy the safety requirements, is left to the car companies, to enable the best AND most cost-effective design to be chosen. The car company designs the car, not the ISO International Organization for Standardization. The fact that Tesla failed at that design is, to me, a sign of deficient engineering, and of not finishing the design before releasing it to the public. They also did not realize that owners usually do not read their owners' manuals.

Leisesturm

The woman that the Uber car hit was jaywalking at 2:00am! The car that hit her was doing 37mph in a 45mph speed zone. The back-up driver being on their phone looking at a movie was irrelevant. When pilots hand off control of an aircraft to one another there is a period of transition while the first pilot remains in control and the second pilot gets their bearings. If humans are to serve as 'back-up' drivers for AV systems then they cannot be allowed to disengage their awareness of what the car is doing anymore than the driver of a standard car can disengage their awareness. It simply takes human beings too much time to react to a fast developing situation to have any hope of being usefull. AV back-up of human drivers? Ya think??!! Now there is a niche ripe for exploitation. AV reactions are inhumanly fast and in many situations that might overwhelm a human driver the AV would stay calm and could bring a vehicle back from a potential disaster. But not this one. That woman was doomed the nanosecond her feet hit the roadway. Nothing could have saved her and that is why Uber was not sanctioned.

I think its wrong to stick people in the cockpits of these thing and then tell them they can actually save it if the AV screws up. Yeah, right. Only a human is likely to have that much hubris as to think such a thing is actually true. You drive a car or you don't. When you are in a dual controlled car with an instructor s/he is ENGAGED. When you are flying in a dual controlled aircraft with an instructor, s/he is ENGAGED. When that Budweiser semi-truck made that 120 mi. run from Wisconsin to Colorado (?) the human hung out in the back of the cab. What were they going to do if the truck got into trouble from back there? Their job was to navigate the local streets to the Interstate on-ramp and take control again to make the last leg to the distributor.

AV's are going to have accidents! Many will argue that that is a sign of their unfitness. Maybe, but probably not. The physics of very large masses is pretty concrete. The Tesla driver who was NOT driving a fully autonomous vehicle but was not ENGAGED was at fault in his own demise. That is not a mark against Tesla. It should have 'seen' the truck. Maybe. I'm sure it does now, and more besides. These things probably see in x-ray as well as infrared, LIDAR, RADAR, SONAR and plain old 'eyesight'. When they really are perfected they will be BETTER than you. That's just the way it is. If you don't want to spend the second half of your life starving to death you will support Progressive politicians that are arguing strenuously for taxes and other compensations from the companies that are poising themselves to completely automate every possible human function that allows ordinary people to earn a living. Every single one. Not even doctors are safe. Robotic scalpels and diagnostic algorithms will put thousands of doctors on the unemployment lines in the coming decades and unless we act now to insist that they pay into a general fund for a Basic Income for all Americans you will be very hungry.

Road Fan

I've read in several literature sources that such transponders are under development, using transmissions similar to WiFi. These have range of 100's of meters, if not a km or two. So perhaps kids, dogs, pedestrians and cyclists can be electronically identified to the AVs as objects of interest.

Kids and adults, fer sure. Dogs and deer? Hmmm.... Imagine being a member of a tagging party that is going to sift through the wilderness and attach transponders to all bear, moose, wild boar, elk, deer, wolves, and wild felines. And which has to go out after every birth season to get the new generations. But it may not be necessary - passive thermal cameras have a much longer range than radar and lidar.

Road Fan

[QUOTE=

AV's are going to have accidents! Many will argue that that is a sign of their unfitness. Maybe, but probably not. The physics of very large masses is pretty concrete. The Tesla driver who was NOT driving a fully autonomous vehicle but was not ENGAGED was at fault in his own demise. That is not a mark against Tesla. It should have 'seen' the truck. Maybe. I'm sure it does now, and more besides. These things probably see in x-ray as well as infrared, LIDAR, RADAR, SONAR and plain old 'eyesight'. When they really are perfected they will be BETTER than you. That's just the way it is. If you don't want to spend the second half of your life starving to death you will support Progressive politicians that are arguing strenuously for taxes and other compensations from the companies that are poising themselves to completely automate every possible human function that allows ordinary people to earn a living. Every single one. Not even doctors are safe. Robotic scalpels and diagnostic algorithms will put thousands of doctors on the unemployment lines in the coming decades and unless we act now to insist that they pay into a general fund for a Basic Income for all Americans you will be very hungry.[/QUOTE]

The Tesla should also have detected and assessed the state of the driver and tried to make the driver alert or brought the car into a safer condition (perhaps stopped!) than driving at a fatal collision speed on a public road, or it should have been prevented from activating into semi-auto control at speeds above, perhaps 30 miles per hour (called "geofencing" or dynamic fencing). Same for the Uber design. In the case of at least one of the Teslas, it was driver sleep. In the case of the Uber, it was driver inattention.

tyrion

My point is that there are strong market pressures to produce safe vehicles. People will avoid taxis with a bad safety record. Insurance will be high (or unattainable) for self-driving trucks with bad safety records. A couple more lethal crashes and Tesla could be out of business. Boeing rues the day it cut corners on 737 Max.

CliffordK

They do tag a lot of wild animals (and could vaccinate them at the same time against things like Rabies). Of course, they never get all of them, except for some critically endangered species (I think every California Condor is tagged including fledglings).

Dogs get tags, and could have an option for an E-Tag.

Hunters, of course, could benefit from hacking the transponders.

As far as kids, it certainly would be an advantage to know where they are at the theme parks, malls, and walks home from school.

Still, I'd anticipate the accident rates to drop without the addition of transponders. It is hard enough to get a lot of cyclists to use quality lights. Then just reduce accident rates when using the transponders.

Of course, we're getting close to this technology with cell phones.

CliffordK

The autonomous vehicles will be off the road in a decade if they can't stop running into homeless people and the broad side of trucks. Or, they won't be licensed for general use.

AV's that you have to have your hands on the wheel?

As far as the 737.... So Boeing is now predicting that they'll be flying again by the end of the year??? That is 6 months away. I'd hate to be one of the customers counting on them.

Perhaps in the future, we won't have critical safety features sold as "optional".

holytrousers

I am beginning to realize that we so distrust each other that we have become ready to give away public space to robots.

jon c. 

Many would consider that a feature rather than a flaw.

rossiny

Read thru some of these post and comments. and all I can say is complicated..!@ The question maybe should be , just because we can, should we??. What would be so great if AV, were completely accident free and safe. The absolute best case scenario. Is there really a good reason to have them? We would still have the pollution from their exhaust, manufacturing them, road building, traffic jams, . I always wonder, when will that tipping point come in when the earth can no longer keep up with the amount of natural rescources needed, and conversely the amount of pollution produced in manufacture. When will the economy be unsustainable , because the earth can not sustain the those two forces. 1. Natural rescources used, 2.pollution produced in manufacture and production? . I dont know the numbers but from what I have read car manufacturers are basically looking at the USA as a stagnant market. The real growth is in China , because China does not have a car for each man and woman . So that is what they are eyeing. I dont think corporations self regulate themselves based on sustainability, rather they regulate based on growth and profit. I have never been able to understand huge corporations being able to undertake massive projects and projections. Running a small business myself it seems you have to do- it-yourself , or else no profit or it does not get done. I guess I am in awe and also fear. How can they control such a huge operation. I know they delegate power and authority , so responsibilities are shared , but still kinda scarey.

Daniel4

There's only one thing to do in order to mitigate the threat of autonomous vehicles: drive better.

Notso_fastLane

This premise falsely assumes consumers 1) have perfect knowledge (never the case); and 2) prioritize safety over all other considerations when buying a vehicle (also not the case).

I work in the aircraft industry, and I guarantee you, manufacturers would leave out a LOT of the mandated safety features if they could. 2/3 the cost of a new aircraft is showing that it complies with regulations (testing, analysis and FAA review) and we still get aircraft that have 'features' that slip through.

tyrion

No, useful knowledge is enough to influence people's decisions in a way that influences auto company business decisions.

No, safety must only be a significant consideration (which it is), not the highest priority consideration, to influence people's decisions.

Leisesturm

Low cost? You mean no cost right? But why should such a thing be necessary?! The only possible exceptions I can imagine are blind or deaf individuals, but all other categories of people should not have to wear transponders so they won't be hit by cars, autonomous or otherwise. If AV's cannot detect the presence of a pedestrian in or near the road then they cannot be allowed to use the roads. Some of you want these systems to be able to detect and track and extrapolate the future placement of vulnerable targets even when some distance from the roadway and this I think is a bridge too far. We don't expect such powers of prescience from conventional vehicles. Why are AV's being held to such an unreasonably high standard?

OBoile

I fully expect autonomous vehicles to, eventually, make the roads significantly safer for all of us. I can't see how anyone would think this would be the "wrong turn" for us to take.

Yes, there will be glitches, and things that were overlooked. But none of that is anywhere near as significant as something like texting and driving.

OBoile

Ridiculous comment. Technological progress has made various jobs obsolete throughout history. The idea that we shouldn't move forward just because some people will experience difficulty would keep us in the dark ages.

Leisesturm

Yes, technology has made 'various' jobs obsolete throughout history. There always was something else for a resourceful individual to transition to. This time we are literally looking at the 'end of work'. Some people? No, all people. And that is not an overstatement. The exceptions won't even fill the visitor section of the stadium at U of M, Ann Arbor. The technological advancement side is rolling out automated platforms across every sector of employment imaginable and it is our social awareness that is mired in the Dark Ages of 'by the sweat of thy brow shalt thou eat bread'.

Leisesturm

Depends on how widespread you believe texting and driving is. I mean ... just as a WAG, given the general consensus I would have expected vehicle deaths to have risen 100% at least, over ... say 10 years. Have they? I'm not sure but I don't think so. The data is certainly there if anyone is curious. But, to your point, I am not one of those who thinks AV's will be rampaging murder machines. I am concerned that given our penchant for insisting that our programmer class consist exclusively of individuals from tightly regulated societies that AV's will prove unable to function in a mixed culture of automated and non-automated vehicle traffic. Guess who will be asked ordered to cede the public spaces when the conflicts begin to rise in number?

base2 

I disagree with the premise.
We already know civilization took a wrong turn. Yet we ride anyway. The danger is the human driver with his very human limited attention, limited skills, limited information processing, and unlimited emotion capacities. We know & accept this as the normal state of things...Yet we ride anyway.

Autonomous vehicles have proven to be much much safer & will continue to become more so with ever improved AI & sensors.
The hold up currently isn't technology or even degree of corporate liability, it's regulation and a legal framework assigning responsibility.

I, for one, welcome our autonomous vehicle overlords.

Insurance companies would do well to offer autonomous vehicle discounts.

tyrion

This is inevitable. AVs will be significantly safer hence cheaper to insure.

I-Like-To-Bike

You have a peculiar definition for proof of safety for autonomous cars since none have been tested let alone fielded yet without human oversight/control for use on public roads except under strictly controlled and/or limited traffic and weather conditions.
Your faith/best wishes/wishful thinking doesn't cut the mustard as proof of anything and neither does the expenditure of a lot of money on the project.