GM brings AI to the track
2024-10-11
Auto racing is the ultimate analog competition: driver against driver, car versus car, all unfolding in real-time. The goal is the same as it was a century ago: First one across the finish line wins. So, you might be surprised to find out that artificial intelligence and machine learning play an increasingly important role in General Motors’ racing efforts. From NASCAR to IndyCar, AI is helping GM dominate the racetrack.
No driver races alone. Every top-tier racing team has dozens of experts monitoring the car, the driver, the competitors, and the track conditions, using this data to craft a real-time racing strategy. Some racing series even allow onboard telemetry equipment in the cars, transmitting vehicle information to the race team for a second-by-second view of how the car is performing. Drivers keep in constant contact with their teammates in the pits via onboard radio, and analysts track lap times and monitor for traffic and crashes.
The result is a constant, massive flow of data, more than any human could evaluate in real-time. So where the rules allow it, GM makes AI a part of the racing team.
At GM’s Charlotte Technical Center in North Carolina, racing engineers and motorsports analysts keep tabs on Chevrolet and Cadillac race cars competing hundreds or thousands of miles away. Thanks to AI, these experts have the most important data available almost instantly, empowering them to make informed decisions at the speed of racing.
At the Charlotte Technical Center, AI transcribes drivers’ radio conversations instantly, allowing analysts to stay informed and react to feedback from the athlete behind the wheel. Off-the-shelf transcription software couldn’t handle the raucous background noise and specialized terminology heard on race-day radio, so GM developed its own proprietary AI custom-built for the rigors of racing. With every word committed to text, analysts can keep track of numerous topics at once - or monitor multiple drivers’ radios for updates on changing race-track conditions that might require adjustments to the cars.
In racing series that allow car-to-pit telemetry, AI keeps track of the data to help craft a pit strategy. By monitoring fuel consumption and using simulations to predict tire wear, AI helps race teams pick the best moment to come in for more fuel and fresh tires. A badly timed pit stop can push a front-running car to the back of the pack, so nailing the pit strategy is crucial for victory.
AI doesn’t just read the string of digital data coming from the car—it can evaluate the car’s condition in a flash. Race team photographers stand at the edge of the track, snapping photos as their car runs past to document crash damage or mechanical malfunctions. It used to take precious minutes for each photo to reach the team and be analyzed. Now, every photo is AI-tagged in seconds, and engineers get alerted whenever damage is detected in an image. The tech also works in reverse, letting teams know when a car escaped unscathed after a crash on the track—and helping them avoid losing time to an unnecessary pit stop to check for damage.
Racing is still a human-led sport, with drivers and strategists vying against the clock and the competition to cross the line first. With the help of AI, those humans can focus on the most important data in any one moment, making informed decisions that lead to better, more exciting racing.
William Byron, driver of the #24 Valvoline Hendrick Motorsports Chevrolet Camaro ZL1, in the pit at the NASCAR Cup Series Hollywood Casino 400 at Kansas Speedway. Artificial intelligence developed by General Motors helps race teams refine their pit-stop strategy in real-time during a race.
Auto racing is the ultimate analog competition: driver against driver, car versus car, all unfolding in real-time. The goal is the same as it was a century ago: First one across the finish line wins. So, you might be surprised to find out that artificial intelligence and machine learning play an increasingly important role in General Motors’ racing efforts. From NASCAR to IndyCar, AI is helping GM dominate the racetrack.
No driver races alone. Every top-tier racing team has dozens of experts monitoring the car, the driver, the competitors, and the track conditions, using this data to craft a real-time racing strategy. Some racing series even allow onboard telemetry equipment in the cars, transmitting vehicle information to the race team for a second-by-second view of how the car is performing. Drivers keep in constant contact with their teammates in the pits via onboard radio, and analysts track lap times and monitor for traffic and crashes.
The result is a constant, massive flow of data, more than any human could evaluate in real-time. So where the rules allow it, GM makes AI a part of the racing team.
In command rooms like this one at Charlotte Technical Center, AI selects the most important racing data and shows it to GM analysts on a second-by-second basis.
At GM’s Charlotte Technical Center in North Carolina, racing engineers and motorsports analysts keep tabs on Chevrolet and Cadillac race cars competing hundreds or thousands of miles away. Thanks to AI, these experts have the most important data available almost instantly, empowering them to make informed decisions at the speed of racing.
At the Charlotte Technical Center, AI transcribes drivers’ radio conversations instantly, allowing analysts to stay informed and react to feedback from the athlete behind the wheel. Off-the-shelf transcription software couldn’t handle the raucous background noise and specialized terminology heard on race-day radio, so GM developed its own proprietary AI custom-built for the rigors of racing. With every word committed to text, analysts can keep track of numerous topics at once - or monitor multiple drivers’ radios for updates on changing race-track conditions that might require adjustments to the cars.
In racing series that allow car-to-pit telemetry, AI keeps track of the data to help craft a pit strategy. By monitoring fuel consumption and using simulations to predict tire wear, AI helps race teams pick the best moment to come in for more fuel and fresh tires. A badly timed pit stop can push a front-running car to the back of the pack, so nailing the pit strategy is crucial for victory.
AI doesn’t just read the string of digital data coming from the car—it can evaluate the car’s condition in a flash. Race team photographers stand at the edge of the track, snapping photos as their car runs past to document crash damage or mechanical malfunctions. It used to take precious minutes for each photo to reach the team and be analyzed. Now, every photo is AI-tagged in seconds, and engineers get alerted whenever damage is detected in an image. The tech also works in reverse, letting teams know when a car escaped unscathed after a crash on the track—and helping them avoid losing time to an unnecessary pit stop to check for damage.
Racing is still a human-led sport, with drivers and strategists vying against the clock and the competition to cross the line first. With the help of AI, those humans can focus on the most important data in any one moment, making informed decisions that lead to better, more exciting racing.
