A2RL Returns to Yas Marina

A2RL, The Abu Dhabi Autonomous Racing league, is a showcase of AI technology in motorsport. Returning to Yas Marina this week, A2RL tried their latest developments against former F1 star Daniil Kvyat and, for the first time ever, exhibited six driverless racing cars on-track at once.

Last year was impressive from a technology standpoint, but the racing was a failure, so it was incredible to see huge advancements in both this year as 11 teams representing 10 countries fought for a share of the $2.25 million prize pot.

The ‘season finale’ began with a Silver time trial event with Dallara SF23’s run by Code-19, FRVIAV, TGM, Fly Eagle, and Rapson. TGM took a convincing victory around the Yas Marina North layout.

The next feature was a return of last year’s ‘Human vs AI’ race, where Daniil Kvyat once again faced an autonomous counterpart – A2RL reigning champion, TUM – in a head-to-head over 10 laps. With a 10-second head-start, it was TUM’s job to keep Kvyat at bay.

Last year Kvyat could easily stretch a 10-second-per-lap advantage over the AI, however this year his task was very much different, with TUM lapping within a second of the seasoned racer. They crossed the line at the end of lap 10 with no more than a second between them.

“What was once thought of as science-fiction has now become reality,” remarked Kvyat following his race, pointing out the huge development since April last year.

“Now I can push properly, I can chase [and it will lap] within a second and it’s a great achievement and that’s impressive. I congratulated everyone at A2RL, that’s great progress.

“Who knows where we’ll be next year. I’m super excited about this technology, it’s a fantastic platform for the people to develop this sort of technology right here.”

Behind-the-scenes developments across all areas have built up the final event on the A2RL schedule: a highly anticipated six-car race over 20 laps of Yas Marina. The ‘autonomous stack’ you see replacing the drivers of these Super Formula chassis are all supplied by A2RL, but it’s down to the teams to input the most efficient and effective coding to make the fastest racer. 

Challenges from last year arose in all areas: braking, accelerating – even using full throttle on the straights – besides other more procedural tasks such as yellow flags and safety car periods. These had been rigorously tested and smoothed over the last 18 months, providing the chance for a tense race to end the day.

This year’s six-piece showdown involved TUM on pole, followed by Unimore, who both spent most of the event pulling away, culminating in a sensational overtake by Unimore on lap 11 to take the lead. However, shortly after, the two found a lapped Team Constructor at turn one, attempting to move out the way. Unimore clipped Constructor, causing terminal front-right suspension damage and ending their race in the barrier. 

That collision on lap 12 began a red flag period. When they restarted, TUM convincingly won the race for a second consecutive season. TII Racing and PoliMOVE completed the podium with Kinetiz last of the running cars to finish.

Despite the misfortune to end their day, Unimore was positive about the successes of A2RL this year. Head of Unimore racing, Marko Bertogna, said: “I was very, very happy with the performance that we showed, we were very stable reaching the professional level that we did… The overtake we showed was at a professional level.”

Stephane Timpano, CEO of ASPIRE (A2RL organiser and part of the UAE government’s Advanced Technology Research Council) closed the season, saying: “What happened tonight is not only amazing – it’s the result of very hard work that has been done by all the teams, and by the technical team of A2RL who’ve pushed the technology so hard over the last 18 months. It was great to see Abu Dhabi being so engaged, and the public feeling the emotion that you see in professional racing.”

Given the level of development over the past 18 months, it would be no surprise to return to Abu Dhabi next year and see the AI drivers racing at speeds equivalent to professional human counterparts. You will just have to see.