So We Know This Won't Have a Manual Transmission Available - Page 5 - Dorkiphus.net
Navigation » Dorkiphus.net > Technical & Track Disussions > Other Technical Discussions » So We Know This Won't Have a Manual Transmission Available

Other Technical Discussions A place for technical discussions NOT related to Porsche or BMW. Other makes, home DIY, etc.

Reply
 
Thread Tools Display Modes
  #41  
Old 03-29-2018, 10:08 PM
trytryagain's Avatar
trytryagain trytryagain is offline
 
Join Date: Jul 2014
Location: Northern Virginia
Posts: 655
trytryagain
Default

Can the technology of automated cars really work or are we being sold a bill of goods? Obviously, visual wavelength recognition of important hazards or pedestrians does not work. Way too much computational intelligence is needed. Does radar work? Does Lidar work? What about reflection and scattering of active sensing wavelengths in a cluttered environment? Did scattering off the bike attentuate the signal to the point of “non detection?” What about the truck in the middle of the road that was not detected by the Tesla? It seems clear that shortcuts are being taken, perhaps for the reason of seeking a competitive or marketing advantage. This whole thing may crash and burn, and I won’t shed a tear.
__________________
Jim C.

1986 911 Targa (Aka "Dusty")
2014 Cayman (Aka "Rubber Ducky")

1971 914 (back in the day)
Reply With Quote
  #42  
Old 03-30-2018, 12:48 AM
N0tt0N's Avatar
N0tt0N N0tt0N is offline
 
Join Date: Sep 2013
Location: DC
Posts: 4,628
N0tt0N has five HoF threadsN0tt0N has five HoF threadsN0tt0N has five HoF threadsN0tt0N has five HoF threadsN0tt0N has five HoF threads
Default

I know Dave... And I'm no Dave!

__________________
Martin
2011 Cayman S (Track Car) - Hardtop Blechster
2006 Cayman S (DD)
2016 Mazda CX-5 (Her DD)
2002 Boxster S (Gone) - Ragtop Blechster - Pura Patina!

Dorkiphus: I buy it for the articles
Reply With Quote
  #43  
Old 03-30-2018, 10:09 AM
Handlebar's Avatar
Handlebar Handlebar is offline
 
Join Date: Sep 2014
Location: Herndon, VA
Posts: 708
Handlebar
Default

Quote:
Originally Posted by Vicegrip View Post
We die in/from cars about 3 an hour in the US. Danes seem to have figured out how to die at a rate per mile far less than us Murkin's and for Dogs sake don't drive in Mississippi if you are risk adverse.
Mississippi is worse than South Carolina? Wow.

We just drove north on I-95. Georgia: 250 traffic fatalities so far this year (10 million population), South Carolina was just under 1,000 fatalities (under 5 million population).
__________________
-----
Ken

1988 911 Carrera 3.2
1995 911 - #121 - Track Toy
2002 BMW 330i - Newt’s when he is home
2008 Subaru Outback
2008 VW Touareg 2
2010 Subaru Outback
Reply With Quote
  #44  
Old 03-30-2018, 10:49 AM
Vicegrip's Avatar
Vicegrip Vicegrip is offline
 
Join Date: Jan 2003
Location: Fallsgulch
Posts: 12,516
Vicegrip has six HoF threadsVicegrip has six HoF threadsVicegrip has six HoF threadsVicegrip has six HoF threadsVicegrip has six HoF threadsVicegrip has six HoF threads
Default

Quote:
Originally Posted by Varejao17 View Post
Hi Hugh - I was actually suggesting that just better than humans is not nearly sufficient...

Ed
Better is not where you stop. It is simply a good start. Traffic deaths per mile are going up after going down for a good while. We humans seem to have reached our limits in the present setup so lets make the cars and roads better. The car based systems are just getting started and will only get better and better.

I don't think we should be using public roads for un-manned system testing. I also think that the human should not have a phone in reach.
__________________
http://vimeo.com/29896988203100watts
Keepers of the cog. Rule#5#10#43

You Matter.
Until you multiply yourself times the speed of light squared. Then you Energy.

Science, Please.
Reply With Quote
  #45  
Old 03-30-2018, 12:51 PM
HughA44s HughA44s is offline
 
Join Date: Aug 2014
Location: Woodbridge, VA
Posts: 592
HughA44s has one HoF thread
Default

"I don't think we should be using public roads for un-manned system testing. I also think that the human should not have a phone in reach." - Concur.

Ed I also agree with your comments in regard to a suitable "death" spec being better than than what we experience now.
Reply With Quote
  #46  
Old 03-30-2018, 01:20 PM
smdubovsky's Avatar
smdubovsky smdubovsky is offline
 
Join Date: Mar 2004
Location: Silver Spring, MD
Posts: 5,219
smdubovsky has three HoF threadssmdubovsky has three HoF threadssmdubovsky has three HoF threads
Default

Quote:
I don't think we should be using public roads for un-manned system testing.
I do.

They don't just write a bunch of code and send the first version out on highways to test. There is TONS of testing on private courses before Uber/Google/etc graduate to public road trials. There is sufficient computing power available to do what they need. I'd bet the AI today is actually beating the average (distracted) driver.

And you can't compare Tesla accidents to any other self-driving cars. They have *VASTLY* fewer and inferior sensors than the real self driving cars.

Fatalities/mile: IMO, The yardstick for implementation is to beat humans. It needs to be no better at the start. We didn't wait for planes to be 10x better than human pilots before implementing - it was phased in. Progress will be at an exponential rate once systems are rolled out en masse and feedback gathered. The safety goal line will be moved over time (just as car safety requirements have ratcheted up over the years.)

Once AI cars reach critical mass, they will be able to crowd source data from one another and reach a level of safety simply impossible for a human to compete with. At that point we will see human drivers legislated off certain major roadways.

This is all happening today. Not a pipe dream. The technology already exists.
__________________
Stephen
www.salazar-racing.com
1970 914/6 - 3.0L GT
1983 911SC - 3.32L IROC
1984 930
2008 S2R1000
2015 Mustang GT (Microsoft SUCKS!)
Sold: 2001 Boxster (hers),
2003 996tt x50 , SpecE30, 1996 E36M3 GTS2 racecar
Reply With Quote
  #47  
Old 03-30-2018, 04:13 PM
Vicegrip's Avatar
Vicegrip Vicegrip is offline
 
Join Date: Jan 2003
Location: Fallsgulch
Posts: 12,516
Vicegrip has six HoF threadsVicegrip has six HoF threadsVicegrip has six HoF threadsVicegrip has six HoF threadsVicegrip has six HoF threadsVicegrip has six HoF threads
Default

Quote:
Originally Posted by smdubovsky View Post
I do.

They don't just write a bunch of code and send the first version out on highways to test. There is TONS of testing on private courses before Uber/Google/etc graduate to public road trials. There is sufficient computing power available to do what they need. I'd bet the AI today is actually beating the average (distracted) driver.

And you can't compare Tesla accidents to any other self-driving cars. They have *VASTLY* fewer and inferior sensors than the real self driving cars.

Fatalities/mile: IMO, The yardstick for implementation is to beat humans. It needs to be no better at the start. We didn't wait for planes to be 10x better than human pilots before implementing - it was phased in. Progress will be at an exponential rate once systems are rolled out en masse and feedback gathered. The safety goal line will be moved over time (just as car safety requirements have ratcheted up over the years.)

Once AI cars reach critical mass, they will be able to crowd source data from one another and reach a level of safety simply impossible for a human to compete with. At that point we will see human drivers legislated off certain major roadways.

This is all happening today. Not a pipe dream. The technology already exists.
Spot on. Can't over emphasize the part I put in bold. Few things drive advancements like putting new tech into many hands.

The aircraft systems have a much higher bar than cars. The present safety levels for cars per mile is dismal compared to com air. Even the best and brightest get feet and meters mixed up or install an sensor upside down. Remember the early automation augmented "Fly by wire" 320 Airbus that mowed a forest for a bit before plowing mother earth? With a cabin full of reporters, at an airshow.
__________________
http://vimeo.com/29896988203100watts
Keepers of the cog. Rule#5#10#43

You Matter.
Until you multiply yourself times the speed of light squared. Then you Energy.

Science, Please.

Last edited by Vicegrip; 03-30-2018 at 04:20 PM.
Reply With Quote
  #48  
Old 04-19-2018, 08:23 AM
Vicegrip's Avatar
Vicegrip Vicegrip is offline
 
Join Date: Jan 2003
Location: Fallsgulch
Posts: 12,516
Vicegrip has six HoF threadsVicegrip has six HoF threadsVicegrip has six HoF threadsVicegrip has six HoF threadsVicegrip has six HoF threadsVicegrip has six HoF threads
Default

Not true autonomous car data and not a large sample but compelling.

https://electrek.co/2017/01/19/tesla...topilot-nhtsa/


Interesting that 40% seems to be a number that has come up before. In this case for rear ender reduction with reactive braking.

http://www.iihs.org/iihs/ne..."
"For the study, researchers looked at police-reported rear-end crashes in 22 states during 2010-14 involving Acura, Honda, Mercedes-Benz, Subaru and Volvo vehicles with optional front crash prevention. The crash rates of vehicles equipped with the technology were compared with the crash rates of the same models without front crash prevention."

"The analyses show that forward collision warning alone reduces rear-end crashes by 23 percent, while forward collision warning
with autobrake reduces them by 39 percent. The reduction for City Safety is 41 percent.The study also shows that autobrake reduces injuries. The rate of rear-end crashes with injuries decreases by 42 percent with forward collision warning with autobrake and 47 percent with City Safety. Forward collision warning alone is associated with a 6 percent decrease in rear-end injury crashes, though that finding isn't statistically significant."

ADAS and EBA were already shown to be working, before fitted to the Model S. http://www.iihs.org/iihs/sr...
__________________
http://vimeo.com/29896988203100watts
Keepers of the cog. Rule#5#10#43

You Matter.
Until you multiply yourself times the speed of light squared. Then you Energy.

Science, Please.

Last edited by Vicegrip; 04-19-2018 at 08:35 AM.
Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump