Robotics
New adaptive system lets autonomous racing cars boost pace safely in real time
Researchers deployed a performance management module that gradually increases racing speeds while monitoring vehicle dynamics and safety metrics.
A new system helps fully autonomous racing vehicles find the fastest safe pace during a race by continuously adapting how hard they push through each section of the track. Rather than trying to measure grip in real time—a risky approach for critical safety systems—the method adjusts performance targets sent to the vehicle's motion planner and controller.
How the system works#
The Mission Performance module begins each run with preset warm-up routines to bring the tires up to operating temperature. From there, it monitors longitudinal speed, lateral cornering, and combined acceleration on a sector-by-sector basis. When vehicle dynamics and safety metrics look good, it incrementally raises the performance level; if something looks off, it holds steady or dials things back. This adaptive loop keeps the car progressively approaching its maximum safe capability without the guesswork.
Testing in the real world#
The team validated the approach on EAV-25, a fully autonomous Dallara Superformula race car, at the Yas Marina Circuit during Abu Dhabi Autonomous Racing League (A2RL) Season 2. The practical results demonstrate that the strategy effectively accelerates lap-time improvement while maintaining the safety margins critical for autonomous competition.
