Race start monitoring system

[Big Mangalhit]

I think there's a slight misunderstanding, the blue graph is the change in signal strength measured by the system, the orange line labeled "Bottas movement" is just highlighting how far he moved before 'green' - the tolerance itself is basically a distance X they are allowed to move between the 5th red and 'green' (the vertical lines), it is not a threshold for "the car has actually started"

In my shitty sketch he does NOT pass the tolerance at all.

Let me try it this way: any movement within the yellow rectangle is allowed...

1) is a normal start, the car starts moving after green with a certain reaction time
2) is a preempted start, movement starts before green but graph does NOT leave yellow area before 'green'
3) false start since it does leave the yellow area before 'green'

http://i.imgur.com/tW9ewHu.png

those are idealized btw, there's probably some noise in the signal and change in signal strength from engaging gear

Yes I agree 100% with you, that is the way i see it.

I am saying since bottas did a start like your number 2 I think that the 0.201s we saw in the screens comes from the time the blue line took to cross the black dotted line, counting after the green line. does this makes sense? I am clearly way worse than you at expressing myself

[RZS10]

That is a possibility but i doubt it, that way the teams could have long figured out the tolerance, that is if the FIA really doesn't want them to know and their secrecy argument is valid.

They know when their car started moving from their own data, with the 'reaction time' they could easily calculate the distance they travelled.

Unfortunately none of the articles specified where the FOM got the reaction times from...

[hardingfv32]

I think that what is being discussed is an over simplification.

1) It is being assumed that the car is coming to a stop over the staring box loop that provides a position along the signal strength graph that has been illustrated. This is clearly not the case for all the cars on the grid. The strength of their signal could put them anywhere on the curve. It would be pure chance that they would find their stopped position on the curve where it is being discussed. Remember that the car only have to stop approximately within their assigned box.

2) The traditional signal strength curve has a much rounder base than the curve being used above. There is rapid decay as you move away from the peak. If that is the case then the tolerance that has been illustrated is going the vary greatly depending on your stopping (starting) point location on the curve.

The tolerance being discussed is going to be much smaller if the staging (starting) point is near the the top of the curve as compared the the bottom of the curve.

I still think that the FIA is hiding a very inadequate starting system.

Brian

[Edax]

Preempting the start must be quite risky when the light system delay is random.

The FOM graphic showed Bottas had a reaction time of 0.201s. Is that from the light going out, or the signal for the lights to go out?

Just wondering. What is the reaction time? Normally I would say engaging the clutch. But on a downward sloping start straight like Austria, I guess you also have to release the brake pretty much at the same time.

Could the initial movement of the car be caused by releasing the brakes and if so is that counted in the reaction time?

I think the reaction time has to be the time from green light to the transponder passing the 'box starting line'. From the onboard footage Bottas is very clearly moving within 1 frame which corresponds to 20 ms.

I think your right, so that would mean they have average about 15 cm of grace movement (30 cm box) before passing the virtual starting line.

Still makes me wonder whether the movement was caused by letting the brakes go on a slope, or the clutch starting to bite. The former you can more or less predict ( control) the latter likely not.

Anyway I guess this now holds the record for the most analyzed 200 us of F1

[deltaecho5]

I am not in Silverstone this weekend due to North American commitments, but wanted to put in two cents worth just the same.

Sadly, I would like to challenge the validity of the "pictures" of the (theoretical) "race start system" depicted several replies in this thread, the so called "loop" placement, the images of data and analysis of the information.

Not to rain on the parade, but the thread is getting off the legitimate true technical path.

If this is just guesswork, please so state as not to confuse others reading the post. If it is not, please provide the origin of this information, as it is direct contradiction to the current FIA start system. (unless they just changed it at Silverstone).

How it really works for anyone interested:

The FIA delegate within Race Control analyzing the start of the race utilizes broad view high speed camera images to determine any car leaving the start box before the electronically registered time stamp of the last red light being extinguished.

This information is reviewed at the time of the start both utilizing the S/F passing system and also the DVR of the race start. The time stamp is compared to the camera feed capture time, and the movement is determined via calculation: thus Bottas' .2 sec reaction time.

There are no loops in each start box. And in case anyone missed it, there are no loops in the start boxes.

At minimum, if we are just guessing if such as system would be used in the future, this would require 20 additional decoders placed at the start boxes tied to a proprietary piece of software that would calculate the time of each car leaving it's box and comparing a TX hit (which one of the thousands received but the parked F1 car I don't know) to the Start lights time stamp. Not to mention the 2+ hours of time per decoder to setup, calibrate, test and certify it's function.

Since the transponder SNR is not accurate enough to depict movement, only signal peak, the data would ultimately be useless and a poor use of resources. (Let's not get into the 140+ hits per millisecond being transmitted by the TX being picked up by the loop before the start.) (However, all the personally drawn artwork depicting loop crossings is fairly accurate, but only when related to a car traveling over a sector, pit or track data loop.)

In addition, if such a system was utilized, this information would undoubtedly be available the viewing audience via TV gFX instantaneously, and would in all likelihood be telecast as data within seconds of the race start for all to see, so if it's like that, why aren't we seeing it? Because it's not the system we use in FIA.

Here's a start box in Fuji. Aside from the parallel expansion and loop line, there is no loop cut into the start box. You can however see the loop cut across the S/F.


Here's a start box in COTA. There is no loop cut into the start box.


Here is a closeup of the loop cut into on of the S/F at COTA


Sorry folks, it's just not the way we do it.

[RZS10]

I think that what is being discussed is an over simplification.

1) It is being assumed that the car is coming to a stop over the staring box loop that provides a position along the signal strength graph that has been illustrated. This is clearly not the case for all the cars on the grid. The strength of their signal could put them anywhere on the curve. It would be pure chance that they would find their stopped position on the curve where it is being discussed. Remember that the car only have to stop approximately within their assigned box.

2) The traditional signal strength curve has a much rounder base than the curve being used above. There is rapid decay as you move away from the peak. If that is the case then the tolerance that has been illustrated is going the vary greatly depending on your stopping (starting) point location on the curve.

The tolerance being discussed is going to be much smaller if the staging (starting) point is near the the top of the curve as compared the the bottom of the curve.

I still think that the FIA is hiding a very inadequate starting system.

Brian

mate just re-read the whole thread slowly ... i think you did not fully understand any of the explanations or you purposely misinterpret them

1)
- the 'zero' signal strength varies for every car

it's in the very first quoted explanation from the race thread
"[the] loop is (...) measuring the signal strength over time (...) it [the signal strength] will be different every time, but every car has a "zero" (...) strength. "

- the car's movement "creates" the curve - it's position does not "put [it] anywhere on the curve"

2)
see above ... the point you are trying to make is moot... and i mentioned that the illustrations are idealized

please provide the origin of this information

I really just put together what i gathered from articles, mainy this AMuS one which quotes Whiting, the article says that there's 'start position' sensors (plural) on the grid and transmitters/transponders in the cars etc ... + the video posted earlier in this thread

http://www.auto-motor-und-sport.de/form ... 05494.html
google translation:


Maybe AMuS completely misquoted Whiting or never talked to him and pulled all that info completely out of their arse or someone misinterpreted and/or misunderstood what they were told by some FIA PR person ...

I don't know ... to me it sounded plausible ... but maybe everything in this thread is a bunch of bull

[deltaecho5]

Regarding the article - The vague reference it makes is relative to the S/F only, and the article isn't in context honestly. The format referenced is true of P1 and P2 cars, since TX strength to S/F loop is degraded significantly from there for any of the other competitors. We've learned often that utilizing the decoder channel to do this is highly inaccurate and is used only as a redundant system to video scrutiny. In other sanctioning bodies, we also use a GPS emitter setting, and this aides in relative position to the loop, and is useful in auditing start position and out of order line passings at S/F at the start. This is also done in FIA via software.

In any event, there is not a loop in each box for the 20 cars.

On a side note, I did notice there are other inaccuracies in the article...TX's being at different points on the car since cars vary in length? Almost to the millimeter behind the mounting posts of the nose on each car...never found a constructor that would want it farther back in the car

No worries, there are many ways to interpret the very little amount of information about timing that is out there. The drawings are an accurate description of SNR fade to loop position of the loops that are used to score the race.

[AJI]

How it really works for anyone interested:

The FIA delegate within Race Control analyzing the start of the race utilizes broad view high speed camera images to determine any car leaving the start box before the electronically registered time stamp of the last red light being extinguished.

This information is reviewed at the time of the start both utilizing the S/F passing system and also the DVR of the race start. The time stamp is compared to the camera feed capture time, and the movement is determined via calculation: thus Bottas' .2 sec reaction time.

What is the frame rate on the on-board camera's on the cars? Surely the car camera (assuming it can see the lights and the front wheels) would be the easiest way to judge a jumped start rather than a broad view with a time stamp from lights out? Even at 25fps you get a 40ms resolution.
I still don't understand where the 201ms comes from? If Bot was moving before the lights (as the video quite clearly shows he was) the RT should be 0ms or a negative number? Assuming he didn't move before the lights (which we know to be incorrect) 201ms is actually a crap RT. Also, I saw reported somewhere that RAI had an RT of 141ms. Why wasn't he also investigated?

[RZS10]

But if the calculation gave the .2s reaction time ... where did the .06s come from?

If there's no specific system in place it kinda makes the whole thing weird given that they looked at the DVR (i assume that is the recording) and have that high speed cam and still determined it wasn't a false start ...

[AJI]

In other sanctioning bodies, we also use a GPS emitter setting, and this aides in relative position to the loop, and is useful in auditing start position and out of order line passings at S/F at the start. This is also done in FIA via software.

What is the sampling rate for GPS data on an F1 car?

[deltaecho5]

Surely the car camera [...edit...] would be the easiest way to judge a jumped start rather than a broad view with a time stamp from lights out?

Much of the discussion of these things turns to a binary system to make a decision, when only the decision needs to be binary. Technical limitations and variables make it necessary to use multiple sources of data. Consider onboard telemetry other than the camera...

What is the sampling rate for GPS data on an F1 car?

In my post I am not referring to F1 cars in this manner ("In other sanctioning bodies"). In any event the 'sampling' rate for GPS is sufficient to use for timing and also provides useful, for example, in a rolling start audit.

Something to think about...what has been considered, is the more precise the designated systems to determine a false start (centimeter movement) and installation of such at each track, the more accurate the cars positioning within the start box at the start of the race after the warm-up lap. This is currently not feasible to the rules.

Although it is capable of providing useful information, the current system was never optimized to determine precise reaction times to lights...only if per the rules a competitor moved out of the starting position prior to the lights being extinguished.

If movement to change into a gear is acceptable and that tolerance is included in the system, then the system has to accept that this could occur as the last light is extinguished...

-Cheers

[AJI]

Surely the car camera [...edit...] would be the easiest way to judge a jumped start rather than a broad view with a time stamp from lights out?

Much of the discussion of these things turns to a binary system to make a decision, when only the decision needs to be binary. Technical limitations and variables make it necessary to use multiple sources of data. Consider onboard telemetry other than the camera...

What is the sampling rate for GPS data on an F1 car?

In my post I am not referring to F1 cars in this manner ("In other sanctioning bodies"). In any event the 'sampling' rate for GPS is sufficient to use for timing and also provides useful, for example, in a rolling start audit.

Something to think about...what has been considered, is the more precise the designated systems to determine a false start (centimeter movement) and installation of such at each track, the more accurate the cars positioning within the start box at the start of the race after the warm-up lap. This is currently not feasible to the rules.

Although it is capable of providing useful information, the current system was never optimized to determine precise reaction times to lights...only if per the rules a competitor moved out of the starting position prior to the lights being extinguished.

If movement to change into a gear is acceptable and that tolerance is included in the system, then the system has to accept that this could occur as the last light is extinguished...

-Cheers

Nice reply but I'd like more specific answers if you would be so polite as to indulge me. I'll rephrase.

What is the frame rate of the onboard cameras? Assuming the on-board cameras run at 100fps we can see whether a jumped start was made to a 10ms resolution. If this is the case then why not use that system? It's only sensible to verify this up with other data, but a camera that is on-board the car that can see both the lights and the front wheels is pretty definitive...

Where does the 201ms figure come from? Is the 'broad view' camera that references the lights-out time stamp running at 1000fps? If it is I find it hard to believe it can capture a high resolution view of the first 4 rows let alone the entire field...

I don't really care about other sanctioning bodies. My specific question is: What is the sample rate the FIA (or FOM) use for GPS data on F1 cars?

[deltaecho5]

Nice reply but I'd like more specific answers if you would be so polite as to indulge me. I'll rephrase.

On-Board camera operation is not my speciality, I am familiar only with cameras used in the operation of timekeeping, and I am simply not at liberty (or willing due to provocation) to divulge the complete operation of FIA procedure, equipment and computations used to determine the officiating of a race start.

A reasonable person could conclude using data extrapolated from video systems, on-board telemetry and timing systems to provide an aggregate measurement if a competitor had a false start. This is the current process.

I have not indicated there are GPS systems on an F1 car used for the purposes of start box movement verification.

Again, to emphasize my reply to the assumptions that had been made: there are no loops in the start boxes of F1 tracks. If the information I have provided is of no use to the thread, my apologies...

[hardingfv32]

[quote=RZS10 post_id=707627 time=1500082348 user_id=20582
- the car's movement "creates" the curve - it's position does not "put [it] anywhere on the curve"
[/quote]

Disregarding the fact we are discussing a system that does not in use……

With your system does the position of the are car (distance from the transponder to the center of the loop) create a signal that varies depending on the distances? Could it not be stated that these signal/distance relationships form a curve?

Brian

[hardingfv32]

Again, to emphasize my reply to the assumptions that had been made: there are no loops in the start boxes of F1 tracks. If the information I have provided is of no use to the thread, my apologies...

Thanks for your information. I was having difficulty developing a jump start system that made sense with technical reality. So what we have is a system using human judgement from multiple data source. The statement of high accuracy numerical reaction times etc. are implying a sense of accuracy that does not exist.

Brian