AI, ML speed Toyota’s transition from track to road

For the most engaged motorsports fans, the action on the track is only ever part of the thrill. They know they’re watching innovation in fast-forward, seeing engineering jump forward from one race to the next. The teams they support are developing vehicles in a hothouse environment that raises R&D to high-adrenaline levels. It doesn’t just shape race strategy – it helps to shape the future of driving.

For over a century, racing teams have struggled for the upper hand in how they engineer their vehicles: more aerodynamic design, more powerful engines, more responsive suspensions. Today though, it’s a different type of technology that’s being pushed forward fastest. It’s helping to separate winners from the rest, and transforming the value of racing for the brands that sponsor it. Today’s track teams are racing to get more value from data. The AI models they develop to turn that data into innovation could be the most valuable legacy of motorsport yet.

The race to turn data into innovation

“It’s data that wins races,” says Dr. Marc Hilbert, Artificial Intelligence Strategy Lead for Toyota GAZOO Racing Europe, the winner of the last five editions of the iconic LeMans 24-hour race, and newly crowned 2023 FIA World Endurance Champions. “Every time we test or race, the sheer amount of data that we collect is just enormous. That volume of data is why the future for racing lies in AI.”

If you were to design an activity to drive the rapid evolution of ML models and the types of data they can work with, you’d probably come up with something like the World Endurance Championship, which Toyota GAZOO Racing clinched with a one-two finish in the title-deciding Bahrain race this month. The team uses the Amazon SageMaker ML platform to work with a vast breadth and depth of data, generated by everything from cars to the weather – and that data often provides the best option for optimizing a vehicle’s performance in the extreme situations the sport throws up.

“It’s not only time-series data like temperature signals that we’re collecting,” says Dr. Hilbert. “We’re also collecting audio data of the communication between the driver and the pit wall, and video signals like the images of the race from TV. This is going to spark a lot of innovation in motorsports.”

Toyota GAZOO Racing relies on the scalable compute infrastructure of Amazon EC2 to deliver that innovation – and not just within its racing team. “We’re using the AWS cloud to connect everything we do on track with our driving simulator,” says Dr. Hilbert. “We can then use simulations to test a lot of new setups for the cars before we actually go on track, getting the sense of how they might work and how the driver feels about them. We use the cloud for our full vehicle development – our race cars and our road cars as well.”

Bringing the spirit of racing to the road

Breaking down the siloes between the track and the road is a cornerstone of Toyota’s approach to motorsport. Thanks to the AWS cloud, the ML models designed to help win a LeMans 24-hour race are also available to engineers developing the next generation of Toyota’s road cars. “We could start a simulation at Toyota GAZOO Racing’s headquarters in Cologne – and then shortly afterwards our colleagues in Japan can look at the results, look at the parameters and rerun similar simulations, but for road cars,” explains Dr. Hilbert.

This collaborative approach to data and AI, enabled through the AWS cloud, is driving innovation in three crucial areas for Toyota. First, autonomous driving technology and the data it generates helps to enhance how race cars perform in traffic. Next, AI drives the optimization that enables more efficient aerodynamics, and use of fuel and tires, and makes both racing and road driving more sustainable. Finally, the simulators powering race car development have a valuable role to play in the road car buyer journey – engaging drivers with the driving experience that Toyota is developing.

“AI can help make driving and racing more accessible – and more fun,” says Dr. Hilbert. “When people drive Toyota cars, they’ll know that there’s machine learning within those cars that comes from racing – and that brings the spirit of racing to the road.”