The phrase, “Drivers, start your engines,” will have a new meaning in October 2021 when 36 universities from around the globe compete with their team-designed autonomous vehicles (AVs) on the Indy 500 Speedway as part of the Indy Autonomous Challenge.
The high-tech challenge is being led by Energy Systems Network (ESN), an Indianapolis-based nonprofit and initiative of the emerging technologies-focused Central Indiana Corporate Partnership, in close collaboration with the Indianapolis Motor Speedway. Sponsors for the event include Pittsburgh-area companies, Ansys, a simulation software developer, and Aptiv.
Designed to be a collaborative effort, the STEM-centered challenge requires teams to develop “neural nets, computer vision and other artificial intelligence systems” for racing and maneuvering at speeds up to 200 mph. The teams, which were required to register this past February, are modifying a Dallara IL-15 in the hopes of outracing other teams.
According to the organizers, this is the first AV, head-to-head race at the Indy 500. To take on professional racing conditions, including instantaneous decisions to avoid collisions and accounting for wind stream and other factors, a simulation race will be held in February. The teams will use simulation technology from Ansys to test their algorithms. The company has also offered $150,000 to the top simulation performer.
“We can create, with physics, multiple real-life scenarios that are reflective of the real world,” said Ajei Gopal, Ansys president and CEO. “We can use that to train the AI, so it starts to come up to speed.”
The simulation testing involves two parts. First, the modified vehicles must complete 10 virtual solo laps in under 15 minutes. Second, a head-to-head simulation race with other teams will occur. It includes 20 virtual laps to be finished in under 30 minutes without interfering with other racers.
Those who pass will compete in qualification races scheduled for October 21-22. To pass, the AV is required to complete 10 solo races in under 15 minutes, with one lap being completed in under 75 seconds. The final race will be held on October 23.
“There is no human redundancy there,” said Matt Peak, ESN managing director. “Either your car makes this happen, or smash into the wall you go.”
While the heart of the challenge may be to engage future innovators, organizers also hope the competition will shed more light on the possibilities AVs may have in regard to commercialization. The teams will have to consider extreme operation parameters, such as avoiding obstacles at high speeds while maintaining control, as well as spark new innovations in making AV software more economical while also being safe.
Interested in more AV innovations? Check out New Autonomous Driving Simulator Uses End-to-End Reinforcement Learning and When Making Autonomous Vehicles Smart and Safe, It Helps to Have the Right Tools.