Cars will turn to more of an "as a Service" thing in the sooner-than-you'll-expect-it future, so it makes sense that the manufacturers are doing this. And with data analysis it can probably predict that you're going out tonight, just like 10 other people in your area, so it will make sure to dispatch enough cars to wait in your area, so your Vwuber shows up in 5 minutes instead of needing 30 minutes to get from the giant parking structure outside of town.
> And with data analysis it can probably predict that you're going out tonight, just like 10 other people in your area, so it will make sure to dispatch enough cars to wait in your area,
Uber already does this. I recommend more closely following the technology industry.
Leaving aside VW's liein' and cheatin' ways for a moment.
Adaptive cruise control on my Golf is a big thing for me right now. It's one of the reasons I bought that car.
For the long journeys I do, not having to have foot on the gas is great. (Normal cruise control doesn't cut the mustard for me as I'd spend the whole drive worrying about the cars in front of me).
The current models also have lane assist and a crawl mode for traffic jams. Can't wait to get a car with that.
In the short term, before fully autonomous vehicles are wide spread, I can see loads of ways AI could make transport a more pleasant experience.
Autonomous cars drive better in traffic though, since they're not as easily scared by a brake light or caught off guard. I read some research that said that if we'd replace 10% of the cars on high traffic highways with autonomously driving ones, traffic jams would get much, much better. Way less rubberbanding, for example.
I test drove an Audi with near complete autonomous driving in traffic jams recently and promptly ordered one with those features as it is simply brilliant in traffic. Much smoother than I can drive in traffic myself, and in Belgium traffic jams are practically unavoidable.
You can place cars driving in a circle and spontaneously traffic jams will occur without the apparent presence of a bottle neck. If a non-significant number of vehicles are autonomous, it'd be easier to control for this.
If autonomous vehicles can communicate with one another (as I think will eventually happen) then we'd be able to see something very special where vehicles would be able to travel at speed while maintaining very small marginal distances because machines react faster than humans.
>we'd be able to see something very special where vehicles would be able to travel at speed while maintaining very small marginal distances
Braking distance alone (excluding the reaction time) is going to be over 100ft at highway speeds. Considering how often human drivers tailgate AVs will probably have more space between them on average.
The following distance for two autonomous vehicles is going to be very small, much smaller than the distance a human can maintain. There is no point in guessing, the above paper is very quantitative. Look at the section titled "effects of autonomous vehicles".
Autonomous vehicles (AV's) might allow for a smaller headway, by removing reaction time, but this benefit will be minimal at higher speeds when cars requires hundreds of feet to stop after the brake has been applied. Reaction time becomes a smaller portion of the headway as speed increases.
Most human (American) drivers drive 80 MPH. That means they need 300 or 400ft to stop, without including reaction time (these distances get even higher in poor weather). Most people don't do this, as a result lots of people die, rear-end collisions are extremely common. AV's will need to leave a few hundred feet minimum between cars to operate safely at highway speeds, which is more than most human drivers on a well trafficked road where I live.
This is where you see the safety benefits of AV's: conservative driving.
Since you couldn't bother even opening the paper I linked to you, I will post an excerpt from it:
This analysis, with the help of the macroscopic traffic flow models, shows that, in principle, a significant increase in capacity can be expected from using autonomous vehicles and that this would also enable a more efficient use of the existing transport infrastructure. Along with the expected increase in capacity for existing traffic infrastructure, traffic jams and lost time are reduced, which in turn improve the quality of traffic flow. In particular, two factors are responsible for the increase in capacity:
(a)
One factor is the shortening of headways between autonomous vehicles. In this context, it is significant that ride comfort is maintained, despite the short time gaps, by anticipating the actions of the preceding vehicles and thereby enabling lower acceleration or deceleration. This could also be important for column stability. The intercommunication of vehicles and infrastructure appears to be an important prerequisite for this.
(b)
In addition to the duration of the time gap, the speed of the vehicle group is very important. The higher the speed at a constant density, the higher the traffic volume over a cross section. However, achieving high speeds while maintaining traffic density is possible only in purely autonomous traffic. A single human-driven vehicle in the column would lead to slower speeds and reduce the capacity gain.
tldr; fully autonomous traffic allows for the possibility of high density high volume traffic flow.
Agree about VW ACC. I hired a 2016 Golf in France with it and it was fantastic. Set a target speed and make sure it stays in line and the car did the rest.
Fun (and unsafe) trick is using it around roundabouts...
Time will tell if this is enough. It certainly is difficult to turn them into software-focused companies. One reason is that everyone who works there right now, went there to work because she/he wants to work on ICE engines, and not all this "modern stuff".
ICE will probably have to go because of climate impact. But mechanical cars, and other kinds of engine can still be done, albeit not as the 99% monopoly ICE have now.
Bosch is (and always was) a company heavily focused on (1) making things electronic (2) making things programmable (3) R&D, they're very modern. While one of the biggest automotive suppliers they never supplied engines or most of the heavy mechanics (gearbox etc.), but rather: [engine] control, which was iirc invented by Bosch, sensors, electronics, firmware, ...
The new Audis also have pretty nice Assistant Systems and full digital Cockpits. They surely lack behind on electric engines and full autonomy, but i think they understand what the market will look like in 10 years.
Audi will introduce the world's first mass production car with lidar-based automation features in a few months. They will introduce their first fully electric SUV next year. I would be very careful when throwing shade on them right now.
Car manufacturers have been knee-deep in electronics and sw for decades now. 50% of all engineering work on a new car is sw and virtually no function in a modern car is not at least partially controlled by sw. The industry is in the process of switching onboard communication to TCP/IP on ethernet and switching the electronic architecture to pc-like nodes running linux. Linux is the state of the art in infotainment already.
And no offense, but your point about the people in the automotive industry is nonsensical. Only a tiny fraction of engineers are working on the drivetrain and most of even those people would take offense at being called ludites. In fact, electric drivetrains are (and have been for years) one of the big hypes in the industry.
How about they just release Free drivers for the latest cards instead? Having seen how wonky their proprietary drivers are I'd rather have something auditable if they're to get anywhere near controlling thousands of Kgs of metal on public roads.
They consider the low level information about how their GPU works a trade secret. Source code won't be as helpful as documentation, but it's more work than reverse engineering the driver binaries.
How is their progress on EVs? That's what I'm more interested in. Before I'd be interested in their self-driving solutions I'd also want to know they're not cutting corners and cheating their way out of strong security architectures for their cars.
28 comments
[ 3.8 ms ] story [ 26.8 ms ] threadRed light starts pulsing in center console, voice synthesizer: "Michael, are you trying to mess with me?"
Cars will turn to more of an "as a Service" thing in the sooner-than-you'll-expect-it future, so it makes sense that the manufacturers are doing this. And with data analysis it can probably predict that you're going out tonight, just like 10 other people in your area, so it will make sure to dispatch enough cars to wait in your area, so your Vwuber shows up in 5 minutes instead of needing 30 minutes to get from the giant parking structure outside of town.
Uber already does this. I recommend more closely following the technology industry.
Adaptive cruise control on my Golf is a big thing for me right now. It's one of the reasons I bought that car.
For the long journeys I do, not having to have foot on the gas is great. (Normal cruise control doesn't cut the mustard for me as I'd spend the whole drive worrying about the cars in front of me).
The current models also have lane assist and a crawl mode for traffic jams. Can't wait to get a car with that.
In the short term, before fully autonomous vehicles are wide spread, I can see loads of ways AI could make transport a more pleasant experience.
Reminds me of the old saying "there can never be too much traffic. If there was people would stop driving"
I test drove an Audi with near complete autonomous driving in traffic jams recently and promptly ordered one with those features as it is simply brilliant in traffic. Much smoother than I can drive in traffic myself, and in Belgium traffic jams are practically unavoidable.
You can place cars driving in a circle and spontaneously traffic jams will occur without the apparent presence of a bottle neck. If a non-significant number of vehicles are autonomous, it'd be easier to control for this.
If autonomous vehicles can communicate with one another (as I think will eventually happen) then we'd be able to see something very special where vehicles would be able to travel at speed while maintaining very small marginal distances because machines react faster than humans.
Braking distance alone (excluding the reaction time) is going to be over 100ft at highway speeds. Considering how often human drivers tailgate AVs will probably have more space between them on average.
The following distance for two autonomous vehicles is going to be very small, much smaller than the distance a human can maintain. There is no point in guessing, the above paper is very quantitative. Look at the section titled "effects of autonomous vehicles".
Most human (American) drivers drive 80 MPH. That means they need 300 or 400ft to stop, without including reaction time (these distances get even higher in poor weather). Most people don't do this, as a result lots of people die, rear-end collisions are extremely common. AV's will need to leave a few hundred feet minimum between cars to operate safely at highway speeds, which is more than most human drivers on a well trafficked road where I live.
This is where you see the safety benefits of AV's: conservative driving.
This analysis, with the help of the macroscopic traffic flow models, shows that, in principle, a significant increase in capacity can be expected from using autonomous vehicles and that this would also enable a more efficient use of the existing transport infrastructure. Along with the expected increase in capacity for existing traffic infrastructure, traffic jams and lost time are reduced, which in turn improve the quality of traffic flow. In particular, two factors are responsible for the increase in capacity:
(a) One factor is the shortening of headways between autonomous vehicles. In this context, it is significant that ride comfort is maintained, despite the short time gaps, by anticipating the actions of the preceding vehicles and thereby enabling lower acceleration or deceleration. This could also be important for column stability. The intercommunication of vehicles and infrastructure appears to be an important prerequisite for this.
(b) In addition to the duration of the time gap, the speed of the vehicle group is very important. The higher the speed at a constant density, the higher the traffic volume over a cross section. However, achieving high speeds while maintaining traffic density is possible only in purely autonomous traffic. A single human-driven vehicle in the column would lead to slower speeds and reduce the capacity gain.
tldr; fully autonomous traffic allows for the possibility of high density high volume traffic flow.
Fun (and unsafe) trick is using it around roundabouts...
Nvidia and Bosch team up on self-driving car AI supercomputer - https://techcrunch.com/2017/03/16/nvidia-and-bosch-team-up-o...
Just recently Bosch, Mercedes and Porsche joined forces with local universities in the "Cyber Valley" initiative: https://a9t9.com/blog/facebook-cyber-valley/
Time will tell if this is enough. It certainly is difficult to turn them into software-focused companies. One reason is that everyone who works there right now, went there to work because she/he wants to work on ICE engines, and not all this "modern stuff".
The new Audis also have pretty nice Assistant Systems and full digital Cockpits. They surely lack behind on electric engines and full autonomy, but i think they understand what the market will look like in 10 years.
And no offense, but your point about the people in the automotive industry is nonsensical. Only a tiny fraction of engineers are working on the drivetrain and most of even those people would take offense at being called ludites. In fact, electric drivetrains are (and have been for years) one of the big hypes in the industry.