How Indy Autonomous Racing Got Real at Las Vegas Motor Speedway

• Officially this was the second round of the Indy Autonomous Challenge (IAC).

• The first round held last October at the Indianapolis Motor Speedway was only time trials, while Friday's event at Las Vegas featured two cars at a time in head-to-head racing.

• Once the cars were lapping at over 150 mph it was not that dissimilar from cheering for two “normal” race cars running neck and neck to the checkered flag.

“Ladies and gentlemen, start your software.” And with those history-making words, uttered by Karen Chupka, EVP of CES, Consumer Technology Association, the world’s first ever race for autonomous race cars got booted up at the Las Vegas Motor Speedway.

Officially this was the second round of the Indy Autonomous Challenge (IAC) but as it turned out the first “race” held last October at the Indianapolis Motor Speedway was only time trials where nine autonomous race cars drove one-at-a-time around the famous oval.

This time the organizers created a head-to-head elimination race with two cars on the track at a time—somewhat similar to how drag races are conducted.

“The software algorithms required to make two cars race side by side at up to 170 mph takes a considerable jump in resources compared to a solo run,” said Paul Mitchell, president and CEO of Energy Systems Network, the company that produces the IAC, before the racing action got under way on Jan. 7.

Halo, a driverless Kia remotely controlled by an engineer sitting in front of a monitor in the media room, served as the official pace car, leading each of the teams from pit lane so they could complete warmup laps at speeds of 65-80 mph before the start of each round. The Kia’s engineer/pilot communicated with the modified car using T-Mobile’s 5G network.

All of the autonomous race cars are based on a modified Indy Lights racer built by Dallara and powered by a 2.0-liter, turbocharged 388 bhp Honda (K20C) engine. Each cockpit is filled with identical electronic guts including three Lidars, three Radars, six cameras, computers and sensors. Co-incidentally the weight of all this electronics closely matches the weight of a human driver, which meant Dallara did not need to do much suspension re-engineering.

The race cars were programmed to run laps at increasingly faster speeds. On each lap the car in front was the defender and it was up to the second car, the attacker, to safely pass the defender. If it succeeded it then became the defender and the other was designated as the attacker. Passes could only be made along the front straight.

It sounds boring, but in actuality the final race, which pitted the two cars (PoliDRIVE and TUM) that had also proven themselves as being the fastest at Indy, was far from boring.

Once the cars were lapping at over 150 mph it was not that dissimilar from cheering for two “normal” race cars running neck and neck to the checkered flag. In the end the race came to a finish when simple robot error caused TUM, the car that had just been passed, to spin out on to the grass right at the start/finish line. Amazingly it managed to control the spin without heading off into the wall. Afterwards one of the team members, after viewing the video, said TUM had simply inputted slightly too much steering correction which caused it to spin at almost 170 mph as it exited turn four.

The winning car went on to complete the lap but came to a stop on turn four when it sensed the other car was off the track. The two cars were then allowed to make their way to the start/finish line to receive their prizes—a $150,000 check for PoliDRIVE, the winner and $50,000 for TUM who finished second. TUM won $1,000,000 by taking first place at Indy and now adds $50,000 to its winnings.

A jubilant Prof. Sergio M. Savaresi, founder of the winning team from Politecnico di Milano in Italy, said “We’re Italians—racing is in our blood” What’s the betting some of these students end up working for Ferrari’s F1 team? Several students and faculty from the University of Alabama also assisted the Italians.

TUM, which one the big prize at Indy, was developed by students from Technische Universität München in Germany.

Okay. so, what does this all really mean in the real world?

Software and mechanical engineers have been working on autonomous vehicles for over three decades. However it was not until 2005 when the true capabilities of autonomous vehicles finally came to fruition in the DARPA Grand Challenge when Stanley, a VW Touareg modified by students from Stanford University, completed an off-road course just south of Las Vegas.

Since then dozens of companies have worked on designing an autonomous vehicle for every day use. Needless to say the task has proven much harder than expected for anything other than low speed deliveries and usage in controlled environments.

These IAC race cars are helping engineers develop ever-faster reaction times and the ability to sense numerous inputs, such as surface changes, wind and proximity of other vehicles at much higher speeds.

As Mitchell says “this brings auto racing back to its roots as a great arena for pushing the envelope with new technologies.” He even sees some crossover to regular human-driven race cars. “Some of these algorithms could be adapted to help make F1 cars faster and safer by giving real time input.” Perhaps this was why we saw Sam Schmidt, IndyCar Series owner of Blueprint Racing who lives in Las Vegas, in the pits keenly watching proceedings. We bet there were other less recognizable race team personnel attending as well.

On the other hand maybe we are witnessing the beginnings of real multi-car autonomous racing? Perhaps. Surprisingly, it was pretty epic watching two driverless cars passing each other at 160 mph on an oval track.