Autonomous Racing

"Jesus Arias" (99pontiacfirebird)
12/11/2019 at 15:34 • Filed to: None

Hello Oppo community! I am currently doing an assignment for my tech class when I stumbled upon Roborace . I want to know what are your opinions on autonomous racing? Is it better to have the car drive itself or should someone be behind the steering wheel?

DISCUSSION (20)

Textured Soy Protein > Jesus Arias
12/11/2019 at 15:43

Remote control drone car racing, sure maybe.

Fully autonomous car racing where the fans are supposed to root for one AI to beat the other, ehhh no.

farscythe - makin da cawfee! > Jesus Arias
12/11/2019 at 15:45

im fine with autonomous racing

as there is no risk of driver injury they should only race on the most dangerous of tracks tho.... maybe add a couple loops....and rocketlaunchers and shit

RallyDarkstrike - Fan of 2-cyl FIATs, Eastern Bloc & Kei cars > Jesus Arias
12/11/2019 at 15:47

Remote control racing where the drivers drive them by telepresence? I’d be OK with that as it’s still down to the driver’s skill.

Just watching a bunch of cars drive around that are controlled by AI? No thanks.

Jesus Arias > farscythe - makin da cawfee!
12/11/2019 at 15:47

lmao if they add loops and rocketlaunchers then I am all in .

farscythe - makin da cawfee! > Jesus Arias
12/11/2019 at 15:49

right?

now that would be fun to watch

Maxima Speed > Jesus Arias
12/11/2019 at 15:50

Boooring * yawn*

Jesus Arias > Jesus Arias
12/11/2019 at 15:51

I feel like it would be super boring because it’s basically just showing that one team is better than another team at coding.

Danimalk - Drives a Slow Car Fast > Jesus Arias
12/11/2019 at 16:10

I am all for autonomous racing. There is definitely room for both driver and driverless racing series, and the more racing the better in my book.

I am intrigued at the idea that a driverless car could be lighter or more aerodynamic than a drivered car, not to mention less safe.

The cars look bad to the bone as well

MasterMario - Keeper of the V8s > Jesus Arias
12/11/2019 at 16:12

I think it could be interesting, cars on the same team could even be given different driving style, with one being coded more aggressively. It also would allow much higher speed and dangerous racing since no person is in danger.

Snuze: Needs another Swede > Jesus Arias
12/11/2019 at 16:13

It’s an interesting concept but I don’t see a mainstream appeal. Reaching way back into history, what is the basis of sport? It’s combat. The original O lympic games were based on tests of physical and martial skills - the original events were running, jumping, throwing discus, throwing javelin, wrestling, pan kr ation (sounds similar to MMA, basically a mix of bare knuckle boxing and wrestling ), horseback riding, and chariot racing. These are essentially combat skills.

If you fast-forward to today, the most popular sports - football (soccer), as well as football (American), basketball, baseball, and hockey - still focus on these core concepts of athletic prowess and combat skills (kicking, throwing, hitting, etc.). All endure a certain amount of violence (wether it’s slide tackles in soccer or full on fist fights in hockey).

Racing, on it’s surface, seems to buck the trend because there is less of an emphasis on physical skill (though people in the know realize that racing IS physically demanding), and it’s reliance on mechanical devices, but make no mistake that it’s ess entially based on equestrian events and chariot racing, other important combat skills.

It sounds like I’m harping on the combat aspect, but it’s important because even in todays sports, with all our understanding and advancements in safety, there is still a huge element of mortal danger. As much as we decry it (i.e. NFL concussion scandal), we still watch and still support men and women breaking their bodies in pursuit of peak human physical capability.

There’s also the personality aspect to it - we all have a favorite personality in sports. Modern sports amp up the team dynamic, and I’d argue that’s a product of our evolution and understanding as a species, but there’s still always a few key personalities in any sports team that stand out. That allows us to make a human connection to the players in the sport.

I think roboracing fails to satisfy most of these key aspects of sports: there’s no physical, martial demand places on the participants, there’s no air of possible mortal danger, however real or remote, and there’s a lack of personality for people to relate to. I think there’ s a flimsy argument to that last one that you could highlight the programming team, or team managers & owners, but there’s something that feels hollow and unrelatable to that, unless perhaps you’re a programmer yourself. And that’s the falling off point - I think there is a portion of the population that would be interested, fascinated even with the dynamics of AI racing, but for probably 90% of people there’s a huge technical barrier to even understanding the challenges of the sport, that it’s just not going to catch on.

MasterMario - Keeper of the V8s > Jesus Arias
12/11/2019 at 16:13

I mean, isn’t that basically the difference now? Humans are organic machines and our minds, while complex, are still operating on a sort of coding. So different driver s are just differently coded machines in essence.

Snuze: Needs another Swede > farscythe - makin da cawfee!
12/11/2019 at 16:13

I think that would be fun to watch once or twice, but it doesn’t seem sustainable as a long term event.

farscythe - makin da cawfee! > Snuze: Needs another Swede
12/11/2019 at 16:14

maybe..... i vote for lets find out

Snuze: Needs another Swede > farscythe - makin da cawfee!
12/11/2019 at 16:21

Agreed. I’d be totally in for watching at least the first race!

Demon-Xanth knows how to operate a street. > Jesus Arias
12/11/2019 at 16:28

Once you remove the person from the car, it really stops being so re late able. I understand it, but it isn’t something you can get emotionally attached to nearly as readily. If there was even a person driving remotely, you can cheer on a driver. But with an entirely autonomous, it’s a company. At the amateur level cheering on your college team I could see, because it’s a group of people you can relate to. But at the top level, it is like cheering on Mercedes, without Hamilton or Bottas.

I get it. But at the same time, I just don’t see myself getting INTO it.

Pickup_man > Jesus Arias
12/11/2019 at 16:39

I’m actually good with autonomous racing. There’s usually two lines of thought when it comes to racing, driver focused, and car focused. While I can’t ignore the immense talent, skill, and importance of drivers in a race, as an engineer I’ve always been more interested in the car side of things, mainly who can build the best car. Autonomous racing, so long as it isn’t too restrictive should be the perfect showcase for this. Which manufactures can build the best cars? W hich ones develop the best software? Autonomous racing should show who does this best, and should help immensely in the development of this technology.

There’s also the potential that these cars could start operating beyond the capabilities of even the best drivers which would be interesting to see.

Pickup_man > Snuze: Needs another Swede
12/11/2019 at 16:54

There is another side of racing that I think you’re over looking, the design and engineering side of things. While I can’t ignore the romanticism of drivers battling it out on track trying to determine who’s the best driver, there are also the iconic and legendary cars as well. The Chaparral cars, the Daytona/Superbird, the GT40, the Quattro, the RS200, etc. etc. etc. T hose cars didn’t become legendary because of the people who drove them, they did it by being better, or that much different, than everything else on the track. This is where autonomous racing needs to be focused, on the tech. Basic dimension and weight rules and nothing else, let the builders/programmers run free to see who can develop the best tech. That’s the race I want to see. If the autonomous series becomes as regulated as any current form of racing where any advantage is immediately banned it will be a dead sport before it even begins.

Jesus Arias > MasterMario - Keeper of the V8s
12/11/2019 at 16:57

I guess that’s one way to look at it. Bu t even so human can panic or act irrationally while driving. A machine cant. Sure a flaw in the AI’s code can cause it to veer off the road but it’s just not the same.

Snuze: Needs another Swede > Pickup_man
12/11/2019 at 22:53

Counterpoint: Tho se cars were iconic because they were advancements in tech that allowed their drivers to win races, but the drivers were excellent drivers in their own right, and many drivers of those iconic cars were celebrities in their own right . No Joe Schmoe came off the street, hopped in a Lotus 49, and won the Indy 500. These iconic cars were part of a man-machine team, similar to a person riding a horse that form a team. There have been plenty of “advanced” cars that flopped due to unreliability or otherwise an inability to win races because their advancements didn’t pan out in practical application.

For whatever weird reason, I’ve kind of gotten into watching video game speed runs - people trying to beat video games as fast as possible. There’s different categories for runners to compete in based on how much of a game you have to complete to “win” and how many/what kind of glitches (if any) you’re allowed to use to win. But there’s also a category called “TAS”, Tool Assisted Speedrunning, which means using a computer program that provides synthetic inputs to a game to do things that are impossible for a human to do. For example, the best human time on Castlevania: Aria of Sorrow is around 47 minutes. A TAS can do it in about 16.

The reason I bring this up is that I see a lot of similarity here. TAS speed running is neat, it’s cool to see these smart programmers figuring out how to break games, but at some point it’s just an exhibition. Its completely alien and totally unrelatable. Watching a human speed runner, on the other hand, has a relatable element, I can see them do something in a game and go “That’s cool, next time I play Mario, I’m gonna try that!” Si milarly, I can relate to Mario Andretti, or Dale Earnhardt Junior, or John Force - how many times have you driven in your own car imagining you were leading the grid at Monaco, or tearing down the Mulsane straight at Le Mans? I have, and I don’t imagine myself as a box of circuit cards doing it.

All that said, I think I see what you’re getting at, and that’s the problem of a lack of innovation in motorsport. There hasn’t really been anything fresh or new on the grid in most major racing series in a long time. But I think technology is exacerbating that problem, not fixing it. For any given set of rules, there is most likely a single optimum solution, and in top tier motorsports where millions in sponsor dollars are on the line, there’s too much risk to deviate from known good solutions, due to a need to stay competitive. I think you might see some crazy cars for a couple seasons in an AI series, but once a team hits on a car that’s consistently fast and winning races, you’ll see everyone rush to copy them and make incremental improvements. Just look at how much Grand Prix cars changed from 1965 to 1975 vs 2005 to 2015.

Again, I’m not opposed to AI racing, I think people can do whatever they want, and if someone starts up a series and it draws participants and spectators, I think that’s great. I just feel like it’t not going to be very exciting, especially after a couple years.

Pickup_man > Snuze: Needs another Swede
12/12/2019 at 10:17

The drivers were absolutely vital parts of the stories and wins of those cars, no question, but I don’t think the drivers are what made those cars iconic, at least not to me. Apart from Petty, I don’t know a single driver who drove the Superbirds/Datonas, and before the movie I had no idea who Ken Miles, or any of the other drivers of the GT40 were. I still don’t know who was driving the 3rd place car, and I have a model of that one on a shelf at home. The advancements in tech, like you mentioned are what make those cars iconic, it usually takes a hell of a driver to fully utilize that, but it’s the tech that makes them stand apart.

I’d argue that the reason we don’t see anything fresh and new in motorsport isn’t due to a lack of innovation, it’s due to a stifling of it. Over the years an incredible amount of innovative ideas have come forth, but because racing these days is so driver focused, once that tech is discovered it is quickly banned. Active aero, flexible body work to alter aero, traction control workarounds, “cheater” turbo bypasses* , nose cones and wings taller than the car, the boost bumper, 2" fuel line instead of the standard 1/2" , just to name a few. All (or most) of these ideas were within rule sets when they were developed, but immediately banned when discovered. NASCAR these days is so spec that they have to fine people for body slamming cars because it might give a slight advantage. Now, if you’re the kind of person who watches racing for the drivers, then a level playing field, and lots of regulations are what you need. That’s not what I’m interested in though, I’m here for the wild and out there cars, the Superbird wings, the 427 Cammers, the Quattros, the 2J fan cars, the Tyrrell P34 six wheeled monstrosities and so on. That’s where I think AI racing can shine, if they let it.

In short, autonomous racing can go one of two ways, heavily regulated spec cars where the emphasis is on the programming and nothing else, or loose sets of rules where the cars push the limits beyond what a human could handle and the programmers are challenged to reign that car in. If option one is where we go I’m not going to pay the slightest attention, but option two will hold my interest for far longer than any racing series around today.

*this was clearly a cheat, but it was really damn smart, innovative, and cool.