The FIA has stopped using electronic detection loops to monitor track limits infringements at most corners after deciding they are not accurate enough.
The loops were added at corners on many circuits where drivers frequently ran beyond track limits. However following the farcical situation at the Austrian Grand Prix, where over 1,200 potential track limits infringements were identified, the FIA has stopped using the loops in most cases.Tim Malyon, the FIA deputy race director and head of its Remote Operations Centre which provides support to race control during grand prix weekends, said the ROC has “made a state change” in terms of the data it relies on following the experience of Austria.
“We used to have three competing data sources in terms of how a potential track limit would be identified,” he said. “On-car detection – which was the car estimating its position relative to the track – you would get a loop detection, or you would get a manual detection via someone’s eyes.
“Essentially, what we concluded after Austria is that those three data sources were all sending potential reports and if you took out all the ones that were being erroneously reported by the loops, and then did the analysis to say all the ones that were correctly reported by the loops, were they also captured by the human, the answer was yes.
“So we basically concluded that the loops were insufficiently accurate. And that by far, our most accurate solution was having a data analyst looking at the video itself. In fact, that’s an interesting element of the story as currently, through loop positioning, through GPS positioning etc, the human still wins.”
The FIA’s single seater head of information systems strategy, Chris Bentley, said the loops are now only used in some specific cases.
Advert | Become a RaceFans supporter and
“We’ve turned off loops now for every circuit unless there’s a chicane, because it just gets in the way of what we’re trying to achieve,” said Bentley. “And ultimately the rule of thumb is that if it’s too close to call, then the benefit of the doubt goes with the driver.”
“We’ve changed the approach significantly between then and now,” said Malyon. “As an example, at the recent Qatar Grand Prix we had eight people working on track limits instead of the four we had in Austria and between them they monitored 820 corner passes, which were then whittled own to 141 reports sent to race control and of those race control elected to delete 51 laps.
“Because of Austria and improvements made to the software, we can deal with those checks and turn them into 150 reports. Now it’s simple a case of clicking down a list of reports and saying yes or no.”
The number of people working in the ROC will also increase next year, from four to eight, in order to examine potential violations more quickly. At this weekend’s Abu Dhabi Grand Prix the FIA will run a trial of a new system which uses artificial intelligence to weed out false positives from the potential track limits breaches they identify.
“At the moment we’ve ‘brute forced’ the situation by saying ‘we need to make thousands of checks, how do we do that’? Well, we throw people at it, because that’s the most accurate solution. What we’re looking to do now is introduce a level above ROC, and that’s AI software.
Advert | Become a RaceFans supporter and
“Again, it might sound strange but the methodology with this AI has a lot of parallels with discussions going on in medicine at the moment and the use of Computer Vision, for example, to scan data from cancer screening. What they’ve concluded is they don’t want to use the Computer Vision to diagnose cancer, what they want to do is to use it to throw out the 80% of cases where there clearly is no cancer in order to give the well trained people more time to look at the 20%. And that’s what we are targeting.
“So, as we said, currently it’s 800 down to 140, down to 50. What we’re targeting with the AI is to take that 800 down to 50 – to remove the ones that clearly don’t need a human review. So we have two layers of check now and we’ll add the extra Computer Vision layer upstream. And that will allow the expert users in the ROC to look at a smaller number of potential infringements, which should further reduce the number of reports that go to race control, and overall increase the speed of processing.”
2023 Abu Dhabi Grand Prix
- Mercedes’ team photo shows we still have a long way to go on diversity – Hamilton
- Why luck was a factor in Perez’s penalty for “joke” comment
- Bottas and Zhou see encouraging signs from overhauled 2024 car
- Only Verstappen got the best from his car every weekend, rival team bosses admit
- How Sainz suffered the worst season finale for a driver since Schumacher in 1997
SteveP
23rd November 2023, 19:40
Wrong technique in use.
Trying to use a proximity system to determine relative position to a fixed line will always have a notable degree of error
If they tried contact sensing triggered when the car touched the limit line, they would eliminate the variability present in distance measurement.
Cheap pressure sensors in the line is what they need.
Rhys Lloyd (@justrhysism)
24th November 2023, 2:22
A properly positioned camera on each problematic corner would be far simpler—and portable.
Feeding into AI as Malyon describes makes perfect sense: AI can throw out the clear cut non-infringements so the humans can focus on determining true infringements.
When making a call, the ROC can then feed this decision back into the AI model and it will gradually over time get more and more accurate—likely to the point where every review it surfaces results in an infringement. I do however, think it’s important to keep the human factor reviewing infringements as AI isn’t very good at providing the benefit of the doubt.
SteveP
24th November 2023, 9:01
Massive over engineering. In fact, massive over engineering with a type of technology that isn’t particularly reliable unless a very large amount of human effort has been completed to teach it. Teaching that you need to repeat for every corner on every track.
Plus, you’re talking about moving the cameras and introducing a new error set very time you move them.
People need to stop over thinking the problem.
Define the line as the limit – nothing allowed on it or outside it.
Piezoelectric sensors in the line, you touch one, and you have exceeded limits.
Piezoelectric tech is cheap, very cheap. Use it to power a simple pulse code generator (like your car alarm key, or TV remote) and the code marks the location of the offence. Identifying the specific car is done using the current location system.
(or if that’s too imprecise, a coded passive sensor device in the car wheels – he says, realising the FIA probably didn’t do an accurate location system previously)
If you want fancier definition of which sensor, then each piezoelectric device feeds to (and probably powers) an Arduino based microcomputer board. That would be your test build, because the Arduino is a home development board intended to help you toward a more tailored end use design. ARM tech is low power, so one board monitoring a group of sensors could be powered from a button cell.
You can shorten that, AI isn’t very good. It’s heavily hyped. It’s an automation system capable of doing what it has been taught very fast.
Doing something it hasn’t been taught has a random outcome, with no “common sense” human judgement.
We’re a (non-Artificial) Intelligence system that has had a lot of teaching, at a deep level it is millions of years of teaching.
Jere (@jerejj)
23rd November 2023, 20:04
Good
Olivier
23rd November 2023, 21:04
I don’t really know how it is implemented, but in Formula E they have Attack Mode: Drivers temporarily boost their power by driving through designated zones on the track.
Perhaps Formula 1 could do the reverse: Driving with all wheels off the track, will reduce the 33 second power boost they have for the next lap with e.g. 5 seconds?
Keith Collantine (@keithcollantine)
23rd November 2023, 23:04
But in order to do that they’d need a system which can automatically and accurately detect when a driver has run too wide, and the fact they’ve largely stopped using the detection loops shows they don’t have that.
Asd
23rd November 2023, 23:35
OMG, that sounds disgusting. Basically Mario Kart.
SteveP
24th November 2023, 9:39
It might be the next feature of that short form event they do: “Sprint”, the one that ought to feature aspiring drivers in last year’s car running their own competition. Better prep for F1 than a FP1 wedged in on a weekend the team has nothing else of note happening.
Alianora La Canta (@alianora-la-canta)
24th November 2023, 11:27
The technique as stated in the article begged the question. If the verification method to determine if approach A is valid compared to approach B, is simply to use approach B, then of course the match will be 100%. What actually has to happen for a reasonable comparison to make is to compare to approach C or use some other metric that is not simply to use approach B as the yardstick.
It comes across as “Track limits should only apply if the stewards feel like it, and any technique that doesn’t agree must be wrong even when visual evidence says otherwise”, which will cement the feeling I see a lot where it is considered partial – and the risk of its getting manipulated.
AI often hallucinates data, so Approach C is going to be even less accurate – but there’s a sponsor keen to show what its AI can do, so Liberty probably had no choice.