Take it from the expert here, guys.
Self-Driving Cars & How much bad driving is NOT distraction?
- Thread starter Eriol
- Start date
He has been observing them for much longer than the rest of us, I feel like he has at least somewhat of a developed viewpoint the rest of us lack.
Dave
Staff member
I don't know whether I should be offended by this or not.
Frankly, I'd settle for all the cars coming out on the market to have the sensors that apply the brake and/or keep your car a safe distance. and a way to retrofit older cars with it. I know that's probably impossible (the retrofit), but until there are self-driving cars, I'd feel a lot safer.[DOUBLEPOST=1407349889,1407349775][/DOUBLEPOST]
I think SpecialKO just called you one of The Watchers.
Damn, I should have short stopped on the new Mercedes that was tailgating the shit out of me, just to test his sensors...
Nope. Quite the opposite.
We're guys familiar with technology. And not just familiar with it, but intimate with it to a degree where we understand some of the challenges involved. How does a car tell whether that obstacle up ahead is a log, a downed human being, a speed bump, an animal, or cracks/debris from an earthquake? A human driver can instantly recognize the difference (given good lighting) and make a judgement call, but trying to teach a computer how to react to every possible instance of "object > 2in tall detected" is a very difficult engineering task. And this is just one of the many challenges that face the self-driving car. Computers are better drivers than people...IF the conditions match those they were programmed for. Anything outside of scope will cause the computer to throw an exception whereas a person would just make a decision. One of the biggest things about the mission to Mars is that sending a human being along would negate sooooo many of the programming challenges. A human being can function as an amazing general purpose computer, the sticking point is that, early on, it would likely be only a one-way journey with no guarantee that PersonOne would even survive to arrive at the destination, and nobody wants to be responsible for that.
--Patrick
And Google is thinking very hard about these things. http://io9.com/when-a-self-driving-car-meets-an-indecisive-cyclist-1568733226
Necronic
Staff member
The thing is though that this is the kind of tech that we rarely see. Not that it doesn't exist, but we just rarely are exposed to it. It is feature limited but has to be insanely robust, crashes are unacceptable and all exceptions should be handled and it all has to be done fast without lag. Most of us are used to seeing stuff like our smartphones and computers, which are really almost the exact opposite, feature freaks but not robust at all. This is not to say that the style of tech I described doesn't exist. This kind of technology would be the consumer equivalent of buying missile guidance software.
That was an accident.
You should have a self typing program write your posts to avoid those
This is my point though: no they don't. They freeze up, and usually hit it. Drivers SUCK on average. Most are completely unable to deal with ANYTHING even slightly different than what they deal with day-to-day.
I wish I had the degree of faith that you do in the human driver, but I've already seen too many counterexamples to be with you on this. The computer isn't going to be worse, it's going to be better.
A human driver may be paralyzed with fear, but his brain can instantly discriminate between log/deer/human/hole/etc. Teaching a computer this will be difficult. When I am driving, I assume all other drivers will be self-absorbed and/or self-prioritizing, and react accordingly. Usually, I am not disappointed, which unfortunately makes me disappointed.
--Patrick
In good lighting conditions, without rain or fog or blinding sun, when they're not distracted or tired, etc etc.
A computer doesn't have to use our limited band of light. Infra-red or heat detection will tell the difference between a log and a human who's fallen over with far more certainty than a human eye, especially in bad conditions.
Thing is, yes, occasionally a computer car will crash or run over a human being. The question is if any driver besides Michael Schumacher at the top of his game could have avoided that person. Frankly, as far as I've seen auto-parking, auto-stopping, auto-following a lane, auto-driving in traffic jams (you can already buy cars that'll do that perfectly well for you up to about 40 mph), computer engineers seem to go for "err on the side of safety" anywhere and always (as in, not turning when you can because there's a pedestrian looking out towards the crosswalk). With good reason, of course, but still.
I see this being huge for the shipping industry as well as mass transportation, like buses. Imagine if the Bus driver rarely had to focus on driving, and could instead focus attention on students.
Like others have already said, the current technology does not require special infastructure, a central decision making server, or for the majority of vehicles on the road to be self-driving. That being said as the technology becomes more widespread things will improve exponentially as more vehicles can "talk" to each other (i.e. the car in front of you tells your car that it's breaking, allowing for your vehicle to start braking before it's own sensors pick up the break lights or decrease in velocity).
This stuff is going to be amazing.
With today's car industry? Fuck no. Keeping the norm is key. Innovation isn't accepted. That's how they work. 10 year old screen tech in 95% of cars.
That's why they still run on fossil fuels and new guys like Tesla aren't welcomed because they "break the norm".
Self-driving cars is a pipe-dream that won't happen... at least in my lifetime. It's a great idea but I fail to see how it'll happen.
To be honest, it's probably not that I'm conservative about it, it's just that I'm someone who doesn't automatically think that because it's computer-operated (taking the decision-making away from humans), it must be better and therfore the answer to our driving problems. I've never been one who automatically assumes that our challenges can be solved by shedding responsibility, rather I believe that problems get solved by people who are better-informed and who are more involved their own decisions. In that respect, I'm probably more like @GasBandit . Now, if after careful evaluation, it becomes obvious that autonomous vehicles are a benefit, then I'll be all for 'em, and I'll happily get one (especially if my insurance rates drop as a result), but I'm not gonna just automatically swear my fealty to 'em "because computers."
If the bus drove itself, the bus ride could conceivably become 1st period.
--Patrick
Last edited:
Well, there's the Miniature Wonderland in Hamburg, the largest model railroad in the world.
They do have at least some kind of computer-controlled robot cars, though more or less on magnetic rails.
But it's pretty impressive to watch, we spent half a day of wonderment there in april.
Here's the fire brigade in action, german text and subtitle for additional fun:
They do have at least some kind of computer-controlled robot cars, though more or less on magnetic rails.
But it's pretty impressive to watch, we spent half a day of wonderment there in april.
Here's the fire brigade in action, german text and subtitle for additional fun:
If I may introduce a wrench, how will the money generated from traffic violations be generated when computer driven cars don't get into traffic violations? Will @Dave ever be allowed to visit his relatives in Springbank, Omaha?
Similar to the vein of since electric cars don't use gas, how does the government get to tax them if not at the pumps?
tldr: Bitches gotta get paid yo.
Similar to the vein of since electric cars don't use gas, how does the government get to tax them if not at the pumps?
tldr: Bitches gotta get paid yo.
The amount of money saved by not having to build/maintain the camera infrastructure (or even a network of traffic lights!) will make up for lost revenue.
When the car is self-driving, you won't need to have road signs, traffic signals, speed limits, or any of the other infrastructure which is there to communicate info to the human driver. Just a periodic transponder or two to give positional or regional updates and you're good to go.
Now if you need to generate revenue, you can make it so your car can deliberately violate speed laws, but you have to first agree to receive a ticket.
--Patrick
Problem: the signs, cameras, everything existing in infrastructure and equipment? It wasn't built on cash on hand. Many, if not all, states pay for stuff like that on credit based upon anticipated revenues. If that equipment suddenly becomes useless, the state still ends up paying for it.
You think the Tea Partiers are bad now, just wait until THAT happens.
You think the Tea Partiers are bad now, just wait until THAT happens.
The big savings will be not paying somebody $50,000 to $100,000 a year to keep you safe. Then you don't have to position one of these guys every 60 miles. You don't need to pay $75,000 per car to enforce traffic laws. And not have to put a dozen in each county, per agency that enforces traffic laws.
That all operates on the assumption that self-driving cars are the only things out navigating the streets, AND that they are only used to go on preprogrammed routes. Those who aren't using self-driving cars, or any vehicle at all, will still need the usual control, safety, and navigational aids.
...until self-driving cars become mandated, yes. Much like seat belt use, I see self-driving cars as the sort of thing where there will be a tipping point beyond which Government will up and say, "Welp, they're safe enough, so now everyone has to have one. Autonomous driving will become illegal after 2xxx."
--Patrick
And then we go into "A Nice Morning Drive" or Red Barchetta territory.
Sent from my iPad using Tapatalk HD