The Indy Autonomous Challenge (IAC) successfully held the world’s first race with multiple autonomous cars at CES 2025 – an important milestone in the development of physical AI and autonomous racing. The qualifying session took place in the SIMPHERA software-in-the-loop environment from dSPACE.
Four fully autonomous IAC AV-24 race cars, controlled by AI drivers, competed against each other in a 20-lap exhibition race. The race ended in a thrilling photo finish, with UNIMORE Racing crossing the finish line less than three-tenths of a second faster than Cavalier Autonomous Racing.
To ensure that the cars were safe and reliable in the real world at the Las Vegas Motor Speedway, the university teams qualified their AI drivers before the race in the virtual world in dSPACE’s SIMPHERA software-in-the-loop environment, as the multicar races are complex and demanding environments for the AI drivers. “In the run-up to the race, our aim was to achieve maximum validation between simulation and reality and to test the AI driver software in increasingly complex environments. SIMPHERA, our web-based, highly scalable solution for simulating and validating functions for autonomous driving, has proven to be ideal for this task,” says Robin Heveling, Project Engineer Engineering Services & Predevelopment at dSPACE.
The multi-vehicle race not only demonstrated impressive progress in the capabilities of AI drivers, but also illustrated how the IAC is developing the technology further. For the first time in the history of motorsport, four autonomous race cars completed a full race with multiple overtaking maneuvers and no accidents. This demonstrated how autonomous systems can navigate at extreme speeds while simultaneously safely managing complex interactions between multiple agents in real time.
“Since our first race in 2021, the IAC and our university teams have made history with the fastest autonomous race cars in the world – from setting speed records to introducing head-to-head autonomous racing to the world,” said Paul Mitchell, President and CEO of the Indy Autonomous Challenge. “However, our goal has always been multi-agent racing. To be the first to complete a race with all of our AI drivers, with five overtakes, no accidents, and a head-to-head finish, is testament to the progress of the global IAC ecosystem of highly talented university researchers, industry partners, and government supporters.”
Working with the dSPACE simulator – SIMPHERA – has made controller/planner development so much more fun. We’ve already made so many controller improvements and caught some crazy bugs, and just feel more confident in the Cavalier stack going into our Vegas race.
At CES 2025, nine university teams competed against each other in three races. In the Tier 1 individual time trial, the IU Luddy team from Indiana University won with the fastest lap with an average speed of 163.6 mph (263.29 km/h). The Tier 2 overtaking competition, a two-car overtaking competition in which the teams had to demonstrate strategic racing and AI adaptability in tight scenarios, was won by the Californian AI Racing Team. In the Tier 3 multicar race, the teams competed against each other in an open 20-lap race with five overtaking maneuvers. In use here for the first time: Autonomous push-to-pass, where the AI drivers predetermined when to use a 30-second 25-mile-per-hour boost.
dSPACE – Partner of the IAC
dSPACE has been supporting the IAC as a sponsor for in-vehicle computer technology since 2022 and provides the AUTERA AutoBox for all vehicles. This robust and powerful central computer processes data from lidar and radar sensors as well as from cameras and vehicle networks. At the beginning of 2024, dSPACE expanded its partnership and became the official software-in-the-loop technology sponsor of the IAC. dSPACE has provided the teams with its SIMPHERA simulation solution – a powerful framework that includes simulation models, test automation, test analysis, and the ability to integrate other components such as HIL simulators. Developed in close collaboration with the IAC engineering team, dSPACE’s Cloud Racing Services provide a virtual test environment that gives university teams access to a fully configurable framework. This solution, which is hosted in the AWS cloud, was developed specifically for applications in the automotive industry and enables flexible resource allocation and the parallel execution of test scenarios, significantly accelerating the testing process. For collaborative development and deployment, the IAC teams use a public project hosted on GitHub as part of a continuous integration/continuous deployment (CI/CD) workflow.
IAC at the CES 2025: The Future of Physical AI Presented
In addition to the races, the IAC had a strong presence at CES 2025, offering participants an insight into the IAC’s vision for the future of autonomous systems. The IAC AV-24, the fastest autonomous racing car in the world, was on display in the lobby of the West Hall, demonstrating the performance of AI-controlled vehicles to visitors.
