The Indy Autonomous Challenge and ROS 2

Aug. 17, 2022
Apex.AI's Joe Speed explains how ROS 2 is controlling IAC autonomous cars.

This video is part of TechXchange: ROS: Robot Operating Systems and TechXchange Talks

The Indy Autonomous Challenge (IAC) is "organized by the Energy Systems Network. IAC university teams from around the world compete in a series of challenges to advance technology that can speed the commercialization of fully autonomous vehicles and deployments of advanced driver-assistance systems (ADAS) to increase safety and performance. The competitions are a platform for students to excel in Science, Technology, Engineering and Math (STEM)."

The 2022 Consumer Electronics Show (CES) hosted the autonomous race at the Las Vegas Motor Speedway. These specially built cars were programmed by a worldwide collection of universities using open-source hardware that included ROS 2, the Robot Operating System. ROS 2 can be used for almost any robot, but utilizing ROS 2 in cars can be a challenge, especially those targeting end-users. Apex.AI developed a version of ROS 2 that can meet the requirements, and the enhancements for ROS 2 have been returned to the open-source community. 

In this video, I talked with Apex.AI's Joe Speed about their involvement with IAC, ROS 2, and their work to make ROS 2 automotive-ready



Wong: One of the more interesting events at this year's Consumer Electronics Show was the Indianapolis Autonomous Race Challenge (IAC).

So, Joe, can you tell us a little bit more about what this thing was?

Speed: Yeah, happy to.

The Indy Autonomous Challenge is a program that I've been personally working on and helping since 2020. It's amazing. It's a university challenge that was explained to me in terms of the origin of it.

Sebastian Thrun, who's one of the godfathers of modern autonomous driving, was at the Indy 500 and he had mentioned that much of what's happening in autonomy these days is not that exciting, but if this racing was autonomous I would watch that. It would be very exciting. And so that's how it came about.

My friend Jean O'Connell, she's deeply involved in IAC. I've been working with her on advanced technology autonomous programs for a lot of years. She wrote me this idea and I got very involved helping to build 10 of these million-dollar fast robots. So, whatever robot you have, these are faster. We raced for the first time at Indy, Indianapolis Motor Speedway, in October.

COVID had an impact on getting enough track time. COVID had an impact on getting race cars built and delivered on time. The time trials were very exciting and very fun to watch. But it wasn't head to head.

At CES, they actually got out there and they did a passing challenge. You know, the first time they told me they're going to do a passing challenges, I thought that it sounds kind of boring. But when you see it in person, it's a real passing challenge.

What you have to understand is that each time I pass you, then you pass me, the speed goes up (by 10 miles/h) and I pass you and you pass me and the speed goes up again. We keep doing this until one of us calls chicken or one of us crashes, so it's actually super exciting.

Wong: Could you tell us what was used for the vehicle and the software, which is actually probably even more important?

Speed: Yeah. When this program started out, all of the teams were using some commercial software and a commercial DDS. By the time they got to Indy, every single team had switched to ROS 2 and specifically ROS 2 with the Eclipse Cyclone DDS that we helped contribute to.

By the time they got to CES, the top two teams had actually moved to the very latest ROS 2, ROS 2 Galactic, which has the Eclipse Cyclone with the Apex.AI-developed iceoryx zero copy built-in, and so that was rather exciting. So here, let me show you what that looks like.

You have TUM passing (see the main video), preparing to attack, preparing to pass a PolyMOVE, Polytechnic in Milan, and to actually hit 165 miles an hour while doing this passing. And then for the next lap, you have PolyMOVE setting up on the attack and they're heading 168, 169, 170. And then they actually got all the way up to 173.

Of course, everyone's eyes were diverted to TUM's spun out right in front of us, and that was the end of the race. Those were the top two teams. This was the last day of CES.

It was just an amazing thing to be there for this challenge. These are like my kids. I've been working with these students for quite a while now and I could not be happier for PolyMOVE, for TUM, for all of the universities, University of Hawaii, UC San Diego, University of Virginia, just all of them, such a great set of students.

To me, it's just amazing. They're doing this with open source. What Apex and our friends did was actually develop ROS 2 improvements and ROS middleware improvements that were specific to the performance requirements of the Indy Autonomous Challenge.

What you don't see in this video is that a few weeks before, PolyMOVE took their car out to Apple's five-mile high-speed oval in Yucca Flats south of Kingman, Arizona, and they hit a 176 miles an hour out there on Apple's Oval using 100% free open-source (software).

At Apex, we take ROS 2 and make it real-time, deterministic, and safety-certifiable. We couldn't be happier and we're going to be doing a lot with these universities and with some of their spinoffs, like from TUM that is doing some really exciting things and some of the PolyMOVE people as well.

So yeah, it's been neat.

Wong: Excellent. What's next for the competition?

Speed: More racing. We're trying to get all the cars on the grid and on the track at the same time, racing head to head. So we look forward to that this fall. Some details will be announced hopefully very soon and I couldn't be happier.

I am rather excited about that. We're one of the sponsors. Our logo is on every car and code that we wrote and contributed is in every car. We helped architect ROS 2. We helped develop ROS 2. We helped develop the ROS middleware that every Indy Autonomous Challenge University team has decided to use.

What we're trying to do is give them this very good base vehicle software. It becomes the job of each university to develop their own algorithms, like "What's my race strategy?" "What's my path planning?" "What do I do in a passing situation?" We try to have them not worry about things like "Why is that camera slow?" "Why isn't the radar working?" "Why do I have too much jitter?"

We try to sort all of those things so they don't even have to think about it.

Wong: Excellent. Well, I'm looking forward to seeing the next race and building into platform.

Speed: Thanks. Absolutely. Look forward to seeing you there, Bill.

Check Out the Entire CES IAC Rade

Sponsored Recommendations


To join the conversation, and become an exclusive member of Electronic Design, create an account today!