Indy Autonomous Challenge racecar sets autonomous hill climb record (2024)

Posted by Kevin Jost | Jul 18, 2024 | Autonomy, Education, Motorsports

Over the weekend, the new self-driving AV-24 racecar from the Indy Autonomous Challenge (IAC) university student competition set the record for the fastest autonomous hill climb time at the 2024 Goodwood Festival of Speed. The car was piloted by an AI (artificial intelligence) driver developed by one of the IAC teams, PoliMOVE-MSU, a joint effort from Politecnico di Milano, Michigan State University, and the University of Alabama.

In a series of three runs, the record was broken during the third and final run up the Goodwood Hill where the racecar reached a top speed of 111.2 mph (179 km/h) and finished in 66.37 s. The previous record had been held by Roborace since 2019 when its DevBot 2.0 autonomous car clocked a 66.96 s time at a top speed of 101.16 mph (162.8 km/h).

Goodwood is known for its unforgiving narrow course lined with hay bells and limited run-off space. For the autonomous hill climb in the AV-24, the IAC and PoliMOVE-MSU teams worked closely with Bridgestone’s tire engineers, including traveling to the tiremaker’s European Technical Center and European Proving Grounds in Rome to optimize performance. Precision mapping and localization in the GPS-limited environment was a challenge, so the racecar had to rely more on its Luminar Iris lidars.

“This weekend we overcame significant challenges posed by a complicated track and temperamental weather patterns,” said Professor Sergio Savaresi, PoliMOVE-MSU Team Principal. “Thanks to the exceptional work and skills implemented by our engineering team and tire performance ensured by our Bridgestone racing tires, we made history by showing the public that safe and high-speed vehicular autonomy is possible even in adverse conditions.”

From AV-21 to AV-24

At CES 2024 in January, the IAC launched its AV-24 next-gen autonomous vehicle platform, which represents a significant leap forward in autonomous hardware and software technology. The organization says the platform and related simulation tools will accelerate the development of AI drivers capable of operating autonomous racers safely at speeds greater than 190 mph (305 km/h).

This year’s Autonomous Challenge at CES marked the tech transition from legacy to new IAC racecars. Nine teams from 18 universities participated in the event in different capacities, whether it was competing with legacy AV-21 racecars or demonstrating the capabilities of the new AV-24 racecar.

Six teams from nine universities competed using the AV-21 racecar at the Las Vegas Motor Speedway, with the TUM Autonomous Motorsport team from the Technical University of Munich winning over the second-place Cavalier Autonomous Racing team out of the University of Virginia. The racecars reached maximum speeds of more than 150 mph (241 km/h) despite cold weather affecting the tires and overall windy conditions. In a first for the competition, in a semifinal the TUM and KAIST (Korea Advanced Institute of Science and Technology) teams raced side-by-side and wheel-to-wheel through turns one and two, getting as close 1.5 m (4.9 ft) from each other.

The other AV-21 teams were Autonomous Tiger Racing from Auburn University; Black & Gold Autonomous Racing from Purdue University; and MIT-PITT-RW from the Massachusetts Institute of Technology, University of Pittsburgh, Rochester Institute of Technology, and University of Waterloo.

Three teams from nine universities demonstrated the AV-24 racecar at Las Vegas Motor Speedway, including AI Racing Tech from the UC Berkeley; University of Hawai’i; UC San Diego; and Carnegie Mellon University; PoliMOVE-MSU; and TII UNIMORE Racing from the University of Modena and Reggio Emilia.

“More than three years ago, the IAC took on the challenge of proving that autonomous vehicles could race head-to-head at high speeds,” said Paul Mitchell, President of the IAC. “After setting every autonomous racing speed record with the AV-21, our university teams’ AI drivers have found the limits of that system. Our next-generation robot, the IAC AV-24, will unleash the full potential of AI drivers to push limits that have not been attempted by man or machine with the goal of developing technology that can save lives and offer new high-speed mobility options.”

Better tech and tools

After three years of the AV-21, the IAC and its industry partners have evolved the AV-24 to become smarter and more capable, providing the world’s fastest autonomous racecar with increased functionality enabling the AI drivers to reach new performance levels. The robot is now vehicle-chassis-agnostic, capable of operating in multiple vehicle platforms, further expanding its versatility.

The IAC engineering team collaborated with autonomous vehicle technology suppliers to develop its AV-24 robotics platform. In addition to Bridgestone and Luminar, industry partners contributing to the development of the racecar include Cisco, Continental, dSPACE, Marelli, New Eagle, NI, and VectorNav.

The car’s sensor suite includes a “world’s first” 360-degree long-range lidar vision system using four Luminar Iris sensors, advanced 4D radar perception from Continental ARS540 sensors, and improved GNSS sensors from VectorNav for enhanced localization. The world’s first automated race control with two-way wireless communication and real-time software interface is powered by Marelli smart antenna and Wintax systems.

For AV-24 student team development, dSPACE provides the high-fidelity “digital twin” simulation that enables AI driver training and head-to-head SIM racing via its Simphera software-in-the-loop (SIL) simulation and, since 2022, provision of on-vehicle computer technology with the Autera AutoBox. A custom hardware-in-the-loop (HIL) test rig comes from NI.

The IAC provided two patent-filed contributions in the form of a custom-engineered drive-by-wire system including a first-of-its-kind independent actuation for front and rear brakes and the modular design of the AV-24 integrated robotics system allowing for competent interchangeability and integration into other high-speed vehicle platforms.

IAC says its deeper collaboration with dSPACE on a high-fidelity SIL tool for improved AI driver training will accelerate the transfer of learning from simulation to the real world. The partnership will create digital twins of the IAC AV-24 and historic racetracks where IAC competes.

Current university teams will have access to this simulation tool to develop, test, and validate their AI driver capabilities in head-to-head racing scenarios before they transition them to real-world AV-24 racecars. New teams wanting to join IAC will be able to use the simulation tool to determine their AI driver performance without the cost of real AV-24 racecars. The official IAC SIM racing platform, powered by dSPACE, launched in the spring with additional functionality being added throughout the year.

“After experiencing the enthusiasm of the teams using our pioneering hardware technologies in the first year of our IAC partnership, we are now taking the next step by providing simulation software to elevate the teams’ AI Drivers to the next level on our virtual racetracks,” said Dr. Carsten Hoff, CEO of dSPACE.

Next up: Indianapolis

The IAC held the “world’s first” head-to-head high-speed autonomous racing competition at the Indianapolis Motor Speedway (IMS) in October 2021. Since then, it has organized five races at iconic tracks including Las Vegas Motor Speedway and the Monza Formula One Circuit in Italy.

Along the way, IAC and its university teams have achieved a long list of world records for high-speed ground-based autonomy. They include the autonomous land speed record (192.2 mph), the top speed on track (180 mph), the fastest on-track head-to-head overtake (177 mph), and most miles of autonomous racing (7500 mi).

“Three years ago, we gathered the best and brightest minds from universities around the world to make history at Indianapolis Motor Speedway, proving that autonomous racing is possible and launching a global effort to accelerate the pursuit of high-speed autonomous mobility,” said Mitchell. “After setting many records and proving that autonomous vehicles can race at extreme speeds, we are returning home to the ‘Racing Capital of the World’ to attempt an even more ambitious head-to-head race.”

The IAC will return to the IMS in September featuring a head-to-head race of the world’s fastest autonomous racecars piloted by AI driver software developed by 10 teams representing 18 top universities from North America, Europe, and Asia.

Related