Self Driving Audi RS7 to Lap Hockenheim at Race Speed

[kenny]

» Click image for larger version

Mark your calendars for October 19th because Audi will be showcasing their latest autonomous driving technologies by having an RS7 do a lap around the Hockenheim race track as part of the German Touring Car Championship (DTM) season finale. However, what’s important to note is that the RS7 will not just be doing a slow sunday drive (hah!) but instead will be doing the lap at full race speeds and will likely challenge the top lap times set by human drivers by driving to itself to its physical limits with millimeter precision.

http://www.beyond.ca/self-driving-au...eed/44825.html

[killramos]

Originally posted by kenny

» Click image for larger version

Mark your calendars for October 19th because Audi will be showcasing their latest autonomous driving technologies by having an RS7 do a lap around the Hockenheim race track as part of the German Touring Car Championship (DTM) season finale. However, what’s important to note is that the RS7 will not just be doing a slow sunday drive (hah!) but instead will be doing the lap at full race speeds and will likely challenge the top lap times set by human drivers by driving to itself to its physical limits with millimeter precision.

http://www.beyond.ca/self-driving-au...eed/44825.html

But unless tesla does it it isn't innovative don't you know? Not even newsworthy.

[Sugarphreak]

...

[Fly Fishin']

I will use a self driving car for road trips but a good old dumb car for everything else. This should be a cool display the watch.

[finboy]

Originally posted by killramos


But unless tesla does it it isn't innovative don't you know? Not even newsworthy.

Beat me to it

[Xtrema]

Originally posted by Sugarphreak
Will self driving cars be better at driving than people? How long until driving a car will be considered reckless and dangerous! They might become self aware!?

Damn toasters!

I think if it is alone on a sunny day, dry track, it can probably out drive 99% of human population.

[FixedGear]

Originally posted by Xtrema


Damn toasters!

I think if it is alone on a sunny day, dry track, it can probably out drive 99% of human population.

Definitely. And even (especially?) in dark, wet conditions too.

[94boosted]

I sincerely hope it can't beat a human set lap time, if it does that's scary

[JRSC00LUDE]

I gotta start smoking piles of weed so I can justify paranoia of a technologically autonomous society....

[revelations]

Originally posted by 94boosted
I sincerely hope it can't beat a human set lap time, if it does that's scary

It IS missing about 100 KG of weight, so on a track with turns its going to make a big difference in a longer race.

Nascar, not so much!

[Xtrema]

Originally posted by 94boosted
I sincerely hope it can't beat a human set lap time, if it does that's scary

May not be this time, eventually it will. Right now I assume its programmed by man but once it can learn and combine everyone's telematic data and results, it will have a complete data set to attack each piece of tarmac in every environment variables. The robots down fall right now is processing visual sensory inputs. That will be solved one day with better algorithms, sensors and processors.

The key is I doubt robots can race with humans because I doubt they can include the human factor in the calculation. And human errors are what racing exciting to watch.

If anything need to be fast, efficient and reliable, humans need not apply.

[FixedGear]

^^lidar is really good these days, especially for a simple case like a car track. in fact, lidar is far superior to the human eye in dark or foggy conditions. my $300 vacuum can navigate my house just fine, just imagine how well a robotic car can navigate when the sensor alone is > $100k.

[Sugarphreak]

...

[revelations]

Originally posted by Sugarphreak
I've never found Lidar to be that great from an accuracy standpoint, it is good for getting a lot of rough information that is easily processed, but I don't think it compares to the human eye when it comes to precision. Laser scanning on the other hand would be far better.

(And yes, I know that modern day Lidar use lasers, not the same thing)

I dont follow ....

Last time I worked on airborne systems, LIDAR = Laser Scanning

[Sugarphreak]

...

[Milk2%]

Originally posted by JRSC00LUDE
I gotta start smoking piles of weed so I can justify paranoia of a technologically autonomous society....

[revelations]

Originally posted by Sugarphreak


I've used information from aerial LIDAR, I've never found it that accurate compared to the land surveys. The function is similar to ordinary radar, except that it send out narrow pulses of light rather than broad radio waves. For that reason we use it as rough information only.

High res laser scanning on the other hand is different in the sense that you take it from multiple locations, and compile it to generate a very accurate 3D object.

Personally I don't think Lidar is as good as the human eye for applications like driving.

Ok, i see where youre coming from - youre comparing a LIDAR scanning unit from an aerial platform (dynamic, sub decimeter accuracy) to a static LIDAR scanning unit mounted on a tripod or solid object (mm accuracy)

They are both laser scanners and the very basic principles of bouncing light and measuring time/phase shift/power etc. are the same.

[dubhead]

Originally posted by Sugarphreak
[B]
Will self driving cars be better at driving than people?

Well based on the stupid shit I see on the roads nearly everyday this bar isn't really all that high.

[94boosted]

Originally posted by revelations


It IS missing about 100 KG of weight, so on a track with turns its going to make a big difference in a longer race.

Nascar, not so much!

Yah but I'd be curious to see how much all of the additional sensors, scanners, cameras and computers weigh.

Originally posted by Xtrema


May not be this time, eventually it will. Right now I assume its programmed by man but once it can learn and combine everyone's telematic data and results, it will have a complete data set to attack each piece of tarmac in every environment variables. The robots down fall right now is processing visual sensory inputs. That will be solved one day with better algorithms, sensors and processors.

The key is I doubt robots can race with humans because I doubt they can include the human factor in the calculation. And human errors are what racing exciting to watch.

If anything need to be fast, efficient and reliable, humans need not apply.

I know it's inevitable I guess I'm just not ready for it haha.

[FixedGear]

Originally posted by Sugarphreak


I've used information from aerial LIDAR, I've never found it that accurate compared to the land surveys. The function is similar to ordinary radar, except that it send out narrow pulses of light rather than broad radio waves. For that reason we use it as rough information only.

High res laser scanning on the other hand is different in the sense that you take it from multiple locations, and compile it to generate a very accurate 3D object.

Personally I don't think Lidar is as good as the human eye for applications like driving.

That's because you're working with shitty data. plus i think it's pretty clear your opinion is pretty much worthless when it comes to science and technology.