Under the Hood at Indy Autonomous Challenge – Video

Under the Hood at Indy Autonomous Challenge – Video

Speaker 1: Around the Go

Speaker 2: Tomb spin.

Speaker 1: We’re here at the Las Vegas Motor Speedway for a very special kind of race. Looking at these cars, you’ll notice that there is no human driver behind the wheel because everything is being done with ai. Teams from around the world are going head to head For the chance to win a $1 million cash prize [00:00:30] may the best AI win.

Speaker 3: The autonomous challenge is a competition among top, uh, research universities from around the world who program software to become basically an AI driver or robot driver that’s capable of piloting fully autonomous race cars at speeds in excess of 170 miles an hour. [00:01:00] Our goal is not to replace human race car drivers. In fact, what we want to do with this technology is partner with top racing series, whether it’s IndyCar or Formula One NASCAR to bring some of these technologies into human driven race cars to allow them to go faster and safer. So we have teams from, from all around the world. Uh, we certainly have several from the United States. We also have teams from Germany, from Italy, from Korea, uh, from the Middle East, uh, from Canada. [00:01:30] Some teams are just one university, one cars. Others are a partnership with university.

Speaker 1: Every team has the same vehicle. This is the same kind of car that human race car drivers train on. The AI are training on the exact same vehicles here. They’ve just unplugged the human driver and put in a computing stack that is the same across all the cars. To keep things fair, each team has the same exact hardware [00:02:00] donated by some of the biggest companies in the AI and automotive space, including Bridgestone, aws, continental, Luminar, D space, and more. These companies get to test their gear at high speeds and in extreme racing conditions, which is hard to achieve anywhere other than a proper racetrack.

Speaker 4: What we have here today is at Delara AB 21, which is a special purpose, uh, full scale race car that’s been modified to be able to support our autonomous racing series. We put in the front of the vehicle [00:02:30] a set of lidars. In addition, Dario cameras at the front fish eyed cameras on the, on both sides, and that gives us the ability to have a 360 degree picture of what’s going on around the vehicle. And then the last part of that equation is the radar units, which is bouncing off of everything around us.

Speaker 1: Leading up to race day. The teams get time [00:03:00] on the track for testing and qualifying rounds.

Speaker 5: Our vehicle’s currently on the track. Normally they give us 15, 20 minutes to test whatever one session is enough to fill one or two terabytes of

Speaker 2: Data.

Speaker 1: During these test runs, teams are allowed to use a joystick controller to make slight adjustments that might avoid costly crashes that could knock them out of the competition. But even with this added precaution, accidents still happen.

Speaker 5: [00:03:30] We had an unfortunate crash at tm. You know, their vehicle crashed into us as they were attempting to overtake our vehicle at our own, um, 60 miles per hour. It was just an unfortunate incident, which, you know, for us it’s, it’s racing. It happens.

Speaker 1: Cindy’s team had to repair the significant damage to their vehicle in roughly 48 hours in order to qualify to race on the big day.

Speaker 5: We’re happy to say that, you know, after the immense amount of repairs that were required at the ve uh, we were able to continue, [00:04:00] you know, just porting over all of our software. In

Speaker 4: Between each of these sessions, refine, improve, and work on this software because if you can make fast, accurate, and correct decisions at 200 miles an hour, it’s a little easier to imagine doing it at 60 miles an hour

Speaker 1: With race day. Finally here, excitement is in the air. There’s no joysticks or controllers allowed today. Once those cars leave the starting line, AI [00:04:30] is taking the wheel and one will be winning its creators 1 million.

Speaker 3: Once the cars leave the pits, the team has zero communication with the car. They really just have to sit back and watch and hope that their driver does the right thing. Each AI driver has a different personality. That personality is really the reflection of the team that that brought it to life. So some teams are more risk taking, some are more cautious, some are better at [00:05:00] overtaking, others are better at top speed.

Speaker 4: Part of AI that’s probably most unique about our program is in our perception system. A lot of the teams are using radars and lidars. Uh, we’re using all three radar, lidar and cameras. It’s the ability to track or detect competitor’s cars, maybe just a little bit farther away, maybe a little more accurately by pulling all that together that we think gives us at least a, a bit of an edge.

Speaker 6: AI racing [00:05:30] tech and mid pit r w at this level in an easy pass made on the entrance to the corner, it does need to consolidate the pass by rule.

Speaker 5: A lot of restraints lie on perception stack, and so perception is pretty much how the vehicle sees and observes scene of the world around it. You know, I think that’s one area where we do find ourselves really strong in

Speaker 7: Ai. Racing’s running around 1 35. They fit rws about 10 miles an hour faster. So they’re closing in, Ooh, making the pass on the bottom in the corner. This actually really daring. Looks like they might back off that pass for now and [00:06:00] maybe try again on, on the other, on the back straightaway.

Speaker 6: And again, it’s up to the algorithm to decide if this is too greater a risk to undertake. And it looks like the car is actually off the pace a little bit. It stopped tracking

Speaker 1: AI racing Tech won a battle for third place against MIT Pitt r w, while reigning champions poll move took home the top prize and a dramatic win over tomb, autonomous motorsport. The

Speaker 6: Outside opened the gray stuff and round they go. Tomb spins and poly move

Speaker 1: Back where the winning AI [00:06:30] reached 180 miles per hour, setting a new world record for autonomous driving on a racetrack. So what’s next for all this AI racing tech?

Speaker 3: Getting people comfortable to believe in or trust that AI and autonomous technology could do something to make their lives safer or or more efficient for when they see a car go 150, 170 miles an hour? I think intuitively they’re realizing that AI is capable of doing something that only a small percentage of humans can do. Only the most elite race car drivers can handle those conditions. I [00:07:00] believe that that will encourage people to perhaps say, okay, I can trust this adas or this driver assistance technology when I’m driving on the highway, and maybe someday be willing to get into an autonomous vehicle that’s traveling a hundred miles an hour from one point to another.

Speaker 1: Next up, the India autonomous challenge is headed to Italy where it will host a competition on the famous Manza F1 racing circuit in June, where the AI drivers [00:07:30] will go beyond the oval racetrack facing twists, turns, and obstacles, more in line with what drivers deal with in the real world. As amazing as it is to see these driverless cars going close to 200 miles an hour overtaking each other and doing what professional human drivers do, it’s important to remember the human beings behind the scenes making it all work. Do you wanna see autonomous driving in your car? Let us know down in the comments and subscribe to CNET for the latest and greatest tech news. Thanks so much [00:08:00] for watching.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top