Read time: 8.8 minutes

AI – Artificial Intelligence – has enjoyed a rapid rise to prominence in recent years, exploding across the wider world and developing even faster than people’s understanding of how to use it. Unlike many inventions, it comes with no ‘one-fits-all’ user manual and so many possibilities, the potential for which remains largely untapped. 

In this series, in partnership with Dell Technologies, we’re exploring how it’s used by our race team... 

Time. Whether at the factory or the track, F1 teams never have enough of it. It’s always been intense – but in recent years, the screw has been turned tighter, with shorter practice sessions, compressed Sprint weekends, fewer testing days and more races back-to-back. This is where Dell Technologies’ AI Factory delivers a strategic advantage for McLaren.

Creating better sims

The standard F1 race weekend has three one-hour practice sessions spread across Friday and Saturday during which time, the team will hone the car’s mechanical and electronic setup. The best way to ensure performance is optimised is to have a good car to begin with. Getting this baseline specification right is a job that consumes a lot of time at the factory in the lead up to the race.

There’s a lot of data flying around for them to work with: historical records, offline aero scans, the simulator runs, and weather forecasts, to name just some of it, and that’s a lot for the team to assimilate – but AI can never have too much data.

Anjum Sayed is our Lead Data Scientist, responsible for training AI set-up models. “We will run simulations to see how each part of the car behaves – but the various systems don’t operate in isolation. The current generation of cars, for example, are very sensitive to ride height – but rear ride height is affected by front ride height, which is affected by what front wing we have fitted.

“Tying all of these together requires thousands and thousands of simulations, most of which won’t be anywhere near the sweet spot we’re trying to locate – but using Dell Technologies’ AI Factory to join the dots helps us narrow down the possibilities and puts us in the best position to give Lando and Oscar a car that’s in the right area when practice begins. They can spend more time on track, and less in the garage, having settings adjusted and bodywork changed.”

It isn’t simply a case of being able to plug in more data. Dell Technologies’ AI Factory helps us weigh that data to generate the best baseline. Historical performance and settings are the starting point for race prep but can be contradictory and difficult to judge… Can the previous cars be used as a good guideline? Should the team rely more on recent years, or similar conditions? Or would we be better prioritising races from this season on similar tracks? When people are making these judgements, there’s only so much information they can integrate – but the AI Factory doesn’t have those limitations.

It’s not just about us

Most of the team are focused on our own performance across a race weekend, but some are tasked with understanding what our competitors are doing. Often shorn of context, the data accrued about cars can be difficult to understand – but AI helps turn that into something a little more useful.

This sort of opposition research isn’t just about number crunching. AI can help with some of the more physical tasks too. The primary task of set up isn’t to make a car that laps as quickly as possible but rather to make a car that laps as competitively as possible. Those sound like the same thing – but they’re not. Oscar’s victory in Baku is a good example.

On a track where overtaking is possible, the MCL38 will not have been set up for ultimate lap-time, because that wouldn’t be useful if it can’t attack or defend on the long main straight. That means set up isn’t simply a case of getting the best out of our car, but also understanding what our competitors are doing.

“There are photos that might give us an insight into what others are doing – but looking through them is extremely time-consuming, so we use AI to analyse them and flag up the interesting ones,” says Andrew McHutchon, Head of Data Science.

“Understanding the trade-off between drag and downforce is a good example. For instance, if we turn up at a circuit where everyone else has a much higher end-of-straight speed. We need to rapidly determine what’s causing that. Are they running with less drag? Do they have a better engine? Are they deploying energy in a different way? We can answer the first of those questions by looking at photographs.”

This research is not, however, all tasked for immediate consumption. “Consider something like GPS,” adds McHutchon. “Data from all 20 cars is shared between the teams, and we can try to understand what our competitors are doing from studying their GPS trace, assessing where they are faster. That will inform some of our set-up decisions – but in the longer term, it will also tell us where we need to focus our development efforts.

“If we see a competitor can go through a low-speed corner faster than us, we know there is a solution out there to improve our performance in that area, and can task a development team to go looking for it.”

The AI tyre whisperer

Finally, the value of AI to the trackside team doesn’t end with practice. Tyre management remains one of the great dark arts of F1.

There are too many different variables: racing lines, temperatures, drivers, track changes, and tweaked set-up options can all create wildly different results. There is useful data gathered every time the car hits the track, but extrapolating it from the background noise is difficult – especially when that data is needed in real time. It’s a job for expert tyre whispers – but even they benefit from the support of a trained AI tyre model.

“Just as you can use an AI model to forecast weather, we can use it to forecast tyre behaviour. It helps our race engineers and strategists understand if we can push for another five or 10 laps, and when the optimal window to pit will be.”

The key phrase here is ‘helps’. Just as you cannot go racing with a team of one, AI cannot compete alone. AI thrives as one of the team, working alongside our experienced, knowledgeable engineers to make decisions faster, based on better data. AI is never tired, and never overwhelmed by the sheer volume of information it has to sift through. Together, with AI, the team is stronger, together the team can beat the time pressure.

As Official Technology Partner of the McLaren Group and McLaren Formula 1 Team, Dell Technologies brings its expertise as an innovation catalyst empowering McLaren to accelerate AI outcomes. Find out more here.