Author Topic: A.I.-Generated Land Speed Record Cars  (Read 2117 times)

0 Members and 1 Guest are viewing this topic.

Offline TD

  • Full Member
  • ***
  • Posts: 153
Re: A.I.-Generated Land Speed Record Cars
« Reply #45 on: April 03, 2024, 02:30:17 PM »
The North American Eagle alleged records were timed and certified by the World Land Speed Racing Association (WLSRA). WLSRA is owned (according to their trademark application) by E&D NORTHAMERICAN EAGLE SERVICES, INCORPORATED, DUH!

WLSRA claims: The days of timing lights and questionable accuracy are no longer
necessary. Global Navigation Satellite System (GNSS) ? based timing is
the methodology embraced by the World Land Speed Racing
Association? (WLSRA?), is state of the art, and is being used on the
North American Eagle? land speed record challenger project.

The WLSRA BS was discussed at length in landracing.com in 2016.
So, NAE allegedly "timed" themselves. No details ever given.

It is the duty of genuine landspeed racers to maintain the integrity of our sport!
Garbage in - garbage out!

Having worked with GNSS-based precision time and frequency distribution in another life, I looked for and found an article describing the NAE's on-board GNSS receiver and its integration with an on-board computer: 

https://insidegnss.com/vehicle-positioning-with-gnss-at-the-speed-of-sound/

I also reviewed a couple of papers which discuss the accuracy of GNSS-based velocity measurement, including these:

A couple of things I took away from all of the above:
  • The Topcon receiver on the NAE had a nominal sample rate of 20 / s, or one sample every 50 ms
  • None of this research was conducted at speeds at which an absolute LSR contender might run
  • For conventional automotive applications, reasonable velocity accuracy (< 0.1 m/s) might be achieved
  • In a dynamic environment, eg high acceleration and deceleration, accuracy of velocity estimation suffers.

There are several different ways the receiver could be used to estimate speed.  The most straightforward way would be identify, after the run, the time and position samples captured by the on-board receiver at or near the start and end of the mile or kilo.  One would then calculate the change in position (the distance), and divide by the change in time - just as timing lights and a clock would.

At 600 MPH an LSR vehicle is moving at about 880 feet per second.  In 1/20th of a second (50 ms) - the GNSS receiver sample rate - the vehicle moves 44 feet.  The best one might say is that at time t0 the vehicle was at point x0 on the track (as much as 44' before the start of the mile or kilo), at time t1 the vehicle was at point x1 on the track (as much as 44' past the end of the mile or kilo), and so the average speed over the mile or kilo is about (x1 - x0) / (t1 - t0).  Points x0 and x1 would be very unlikely to be the same as the fixed start and fixed end of the mile or kilo, so they'd have to be selected before the start and after the end.

The position accuracy of x0 and x1 would be whatever the receiver could obtain while moving 600 MPH.  Topcon claims good positional accuracy (a couple of millimeters) for static measurements, but what the receiver could achieve while moving at unlimited LSR speeds isn't published.   As an aside, a GNSS-based transit station would probably be fine for positioning the timing lights; it would not surprise me if this is how things are done in practice.

So, I'd say that an on-board GNSS receiver could be useful for data logging and post-run analysis ("at this time and position on the run we were going about X MPH when the front fell off").  But I think there's enough inaccuracy, and obvious non-compliance with the regulations (the measured mile and kilo aren't in fixed positions on the track; half a mile back we were going way faster so let's count that; we were going at least 765 MPH over the mile, etc.), to disqualify it for consideration as the primary measurement method.   Never mind the FUD created in the minds of competitors and the community. 

Maybe this is why they added lights.
Not to trigger Mayoman, but see also the Budweiser Rocket and the USAF radar fiasco... :dhorse:

Thanks for listening
Tim





Offline MAYOMAN

  • Sr. Member
  • ****
  • Posts: 450
Re: A.I.-Generated Land Speed Record Cars
« Reply #46 on: April 03, 2024, 07:21:39 PM »
If you want to set, and claim, the absolute world land speed record, it is simple.
FIA is the internationally recognized body governing motor sports.
The FIA Code spells it out in their Appendix D - World Land Speed Record Attempts.
The FIA World Land Speed Record is the fastest FIA-recognized land speed record. In this case, it would be the faster of the FIA Flying Start Kilometer or Mile records.
The record speed is the average speed of 2 runs in the opposite direction through the same course within one hour.
Timing devices must be of the automatic type with an accuracy of at least 1/1000th of a second, the recording being produced directly by the passage of the Automobile without any human intervention.
Them?s the rules. If you don?t follow those rules, don?t claim your automobile was faster than The Blue Flame, Thrust 2, or Thrust SSC. They specifically followed those FIA internationally-recognized rules that have been followed over a century.
SIMPLE!
The road is long - Life is short - Drive fast

Offline Stainless1

  • Administrator
  • Hero Member
  • ***
  • Posts: 8980
  • Robert W. P. "Stainless" Steele
Re: A.I.-Generated Land Speed Record Cars
« Reply #47 on: April 03, 2024, 10:53:37 PM »
Yep it is just that simple.... not easy... but that simple.... :cheers:
Stainless
Red Hat 228.039, 2001, 65ci, Bockscar Lakester #1000 with a little N2O