Self-Driving Cars in Four California Accidents

Yes, the world will be completely safe and no one will ever die again when these cars come out.../S

Meanwhile we will have given away yet another one of our ever dwindling pleasures.
Where will it end?
Until we are like optimal weight maggots, fed with a government accepted health sludge, sitting in cages that do all the thinking and doing for us with computers monitoring our swallow reflexes in case we choke on a chicken bone?

Sounds like hell.

Why not accept that things can be dangerous and educate accordingly,then just let humans be humans? It's called life.

a reply to: peck420

But, I have seen the meat sack inside the car react in a way that saved lives. I have been driving for forty five years and probably logged almost a million miles on the road interacting with other drivers. Now if every car was controlled by computers and the roads never got icy and if ice and snow never built up on the sensors, they might work.

Those sensors are not going to detect a small piece of metal laying in middle of the road that can get sucked up under the car and slash the tires. How is it going to react to my road where you need to wander all over the roads to miss pot holes.

Only under ideal conditions will that work. You did not address how the car was going to respond to the things I said in my last post, things sensors cannot detect.

a reply to: Soylent Green Is People
Ah, so the coding...that will be cross checked hundreds of times by hundreds of different people, tested hundreds of times by hundreds of different people, and require live testing per DOT existing rules is the hiccup?

Versus the once in your life driver exam.

Yup...that's logical and reasonable.

a reply to: FamCore

I do think this is expected because the technology is in it's infancy.

I think it will get much better and if the accidents with these cars can reduce the accidents and fatalities that are caused while humans are driving then of course this will be more efficient.

originally posted by: rickymouse
Now if every car was controlled by computers and the roads never got icy and if ice and snow never built up on the sensors, they might work.

My sensors work year round, from Yellowknife to Calgary, to Vancouver. No issues.

Those sensors are not going to detect a small piece of metal laying in middle of the road that can get sucked up under the car and slash the tires. How is it going to react to my road where you need to wander all over the roads to miss pot holes.

You must be confusing modern sensor suites with human eyes. Modern active suspension already "sees" road deformations and adjusts the suspension accordingly...it just doesn't tell you about it. I wouldn't doubt that it sees the metal as well, it just allows the human to decide to avoid or not.

Only under ideal conditions will that work. You did not address how the car was going to respond to the things I said in my last post, things sensors cannot detect.

What would that be? You are attempting to claim that the "sensors won't detect" an obstacle, but disregarding the fact that these cars have more than one sensor, that analyzes more than just the visual spectrum.

a reply to: peck420
Will those sensors detect a two foot deep hole filled to the brim with water up to road level and adjust to miss the hole if it means swerving into the other lane to do so? One simple question.

I have backup sensors on my explorer, they do not work when covered with sandy salty slush and ice. There is no possible way they can with the mineral content of the sand and salt taken into consideration. I also feel that this is true for the sensors you are talking about. A sensor has to be clean to work right. Everyone can't be washing their cars everyday, we have a shortage of water happening in many places.

Now, if you want to believe in this, it is your right, but you won't be able to convince me. I have been doing mechanic work for a long time and know that corrosion on contacts is the number one reason that problems occur on electronics in cars. Corrosion caused by salt on the roads used to control ice.

The gas mileage on cars goes down as the connections get corrosion on them. This is well understood by mechanics. The computer compensates for it and the car runs fine but efficiency drops. It is impractible to be changing a sensor every time there is corrosion. Old cars just needed a tune up now and then and some of the new technology is all right but most time it just causes increased trouble. A friend of mine drove a wrecker for years and he said there has been a big increase in accidents caused by people being distracted by technology most times. Getting rid of distractions is a better solution than inventing vehicles that can drive themselves. I owned a Lincoln town car with all the fancy gadgets and protected connections and constantly was working on that thing. The more electronics and sensors, the more to go wrong. Keeping things simple is the best thing. They will be recalling those cars for all sorts of problems in areas where salt is used to control snow and ice. What will be the response? They will invent something twice as expensive and twice as bad on the environment to spray on the roads.

originally posted by: neoholographic
a reply to: FamCore

I do think this is expected because the technology is in it's infancy.

I think it will get much better and if the accidents with these cars can reduce the accidents and fatalities that are caused while humans are driving then of course this will be more efficient.

Ah the old "If we can save just one person ..." approach.

I prefer freedom to absolute safety.

a reply to: FamCore

They must have programmed them to text message!

originally posted by: peck420
a reply to: Soylent Green Is People
Ah, so the coding...that will be cross checked hundreds of times by hundreds of different people, tested hundreds of times by hundreds of different people, and require live testing per DOT existing rules is the hiccup?

Versus the once in your life driver exam.

Yup...that's logical and reasonable.

Yes.

I'm all for this becoming the future standard. I think someday it will. However, it is NOT a true statement to say "computers controlled cars are TODAY/RIGHT NOW better than human driver solely due to the fact that computers can react more quickly".

As I explained, there is much more to a computer being safer at driving a car than simply fast reaction times. Fast reaction times are great, but just because a computer-controlled car can react faster, that doesn't automatically mean that it is safer.

As you pointed out, the software (written by PEOPLE who can make mistakes) needs to be checked and double checked. There needs to be a proven track record of test track testing and real-world testing phases that need to be done to work out any bugs prior to this becoming "the standard".

Fast reaction times are great, but there is much, much more to making it safe than just fast reaction times. These computer-controlled cars are only in their infancy, and I think it's possible that we will find (through real-world implementation of the system) that the people who wrote the software telling the computer HOW to react in certain situations may have not told the computer the best ways to react.

The computer may react quickly, but it can only do what it's told to do by the software designer. That "only doing what it is told to do" is what I think needs to be closely scrutinized. It may react quickly, but end up quickly doing a "dumb" thing only because the software designer didn't consider the situation properly.

originally posted by: Soylent Green Is People
As you pointed out, the software (written by PEOPLE who can make mistakes) needs to be checked and double checked. There needs to be a proven track record of test track testing and real-world testing phases that need to be done to work out any bugs prior to this becoming "the standard".

Which, funnily enough, is exactly what they are doing now and you seem to have such an issue with it, despite it being so far proven to be safe and reliable.

You seem to have a major hang up with regards to automation, but most modern cars (especially descent, non-American cars) are full of electronics constantly monitoring systems, adjusting power to wheels, adjusting suspension for different road surfaces, adjusting the balance of the car going into a corner etc etc - some even have crash avoidance systems and some have systems that constantly monitor the road ahead with radar.

All of which is done so quickly it is barely perceptible to the human driver and, if left to the driver to do, would be impossible and without such adjustments, the car would be a whole deal less safe.

originally posted by: JAY1980
a reply to: FamCore

So the accidents took place at around 10mph? Doesn't sound like very stable technology if it gets in accidents at 10mph. Let alone barreling down the highway at 70mph. My opinion is it shouldn't be on the roadways yet if the technology isn't sound. It's jeopardizing others safety.

So if you are going 10mph and someone hits you from behind you should not be allowed behind the wheel of a car anymore? Would love to understand that logic.

Can you tell me what the Google cars did exactly that led you to your conclusions?

originally posted by: rickymouse
Only under ideal conditions will that work. You did not address how the car was going to respond to the things I said in my last post, things sensors cannot detect.

By having a person behind the wheel for those rare occurrences.

originally posted by: stumason

originally posted by: Soylent Green Is People
As you pointed out, the software (written by PEOPLE who can make mistakes) needs to be checked and double checked. There needs to be a proven track record of test track testing and real-world testing phases that need to be done to work out any bugs prior to this becoming "the standard".

Which, funnily enough, is exactly what they are doing now and you seem to have such an issue with it, despite it being so far proven to be safe and reliable.

You seem to have a major hang up with regards to automation, but most modern cars (especially descent, non-American cars) are full of electronics constantly monitoring systems, adjusting power to wheels, adjusting suspension for different road surfaces, adjusting the balance of the car going into a corner etc etc - some even have crash avoidance systems and some have systems that constantly monitor the road ahead with radar.

All of which is done so quickly it is barely perceptible to the human driver and, if left to the driver to do, would be impossible and without such adjustments, the car would be a whole deal less safe.

Who said I have a major hang-up with the technology in general? I think this technology is great, and someday cars will be automated, and the road will be a safer place because of it.

My current hang-up was with the assertion that simply THIS computer-car's "fast reaction times" already makes THIS car better than a human driver, as if fast reaction times are the "be-all and end-all" in automobile safety. My hang-up is that just because this car can react quickly right now, that does not automatically mean that it is currently safer than a human driver is, right now.

This computer-controlled car's fast reaction times are certainly impressive, but it STILL only reacts in a way that has been told to do by its human-written software -- no matter how fast that reaction is.

All I'm saying is that the software is in its infancy, and there are probably many bugs in the software that need to be found, and perhaps "bad reaction decisions" that are currently written into the software that need to be weeded out.

As I said repeatedly, this car in its current configuration can in fact react more quickly than a human, but considering safe driving is about MORE than just being about to react quickly, that does not automatically make it safer than a human just yet.

originally posted by: peck420
My daily driver is 10 years old. Here is a list of all of the components that I have direct and absolute control over, without any "computer intervention":



Let that sink in for a bit.

Vehicle components being optimized by a computer system is not the same as a car that operates autonomously.