HILDEBRAND: Should autonomous cars race?

HILDEBRAND: Should autonomous cars race?

Formula E

HILDEBRAND: Should autonomous cars race?

By

To me, it’s a foregone conclusion that we’ll see fully electric vehicles racing at the Indianapolis 500 at some point in the next 20 years. There seems to be a clear relationship between the development of technologies that are relevant to the automotive industry, and not mixing the message in terms of what we think is great about racing.

But automated systems … that’s much more complicated. We look at those two things as being in parallel in the automotive space, but how they might impact racing could be very different. The core of why we like watching racing is the human element, and that doesn’t suddenly go away if we’re driving electric vehicles instead of internal combustion engine vehicles.

There’s a larger question of whether we’re doing a good enough job as it is of illuminating the human element of what we’re doing at the track. But nevertheless, if we introduced automated systems in that same way, all of a sudden you’re removing that human element, and that’s a significant change.

Traction control and ABS could be looked at as automated systems that have existed for decades now, and we purposely don’t have those in racing even though they would make the cars a little bit faster. There would be an improvement in overall performance to have them, but we’ve realized that they take away from the human element of what’s going on.

I think it’s a matter of where automation can be designed into the format of what we do. There is no doubt in my mind that integrating automated systems into a motorsports environment will increase the rate at which those systems are developed. So from an industry perspective, it makes a lot of sense to be trying to find ways to integrate the development of automated systems – and maybe eventually fully-autonomous systems – into what we do at the track.

That, at its core, is what racing – from an industry perspective – is best used for, and how we create maximum relevance. But that has to be done in a way where it is truly designed to exist in places where it is not countering the reason why people want to watch a race.

As an example, everything that goes on in pitlane could be fully automated. Watch an F1 race: are you ever impressed by what the driver is doing in pitlane? No, other than some reaction time element to what is going on. Everybody is within one percent of doing it at the same speed.

And pitlane is a complicated scenario. You’ve got equipment to avoid, you’ve got people to avoid. So that could be an example where we could look at creating pitlanes so that the infrastructure supports automated systems being able to understand where they are, and what the car is doing, and how it gets in and out of its box.

Caution periods are maybe kind of similar; managing cars under caution, managing speed over an entire field … that’s kind of a safety-focused thing. As a driver you have some strategy to what you’re doing under caution, but it’s not something that I think even the die-hard fan is really that aware of. So if that were something that ended up being done in a fully-automated fashion, would that be taking something away? Not really. Is it adding an area for development? Possibly.

So I think there are a lot of interesting integration strategies that could exist, and it’s important that we really think hard about that.

There is an interesting alternative that doesn’t really utilize that, which is more of a project-oriented version of using the motorsport environment to generate a platform for innovation. It would be a little bit more of what I would call the ‘Red Bull model’: could we create a set of challenges that happen at the Indianapolis Motor Speedway using IndyCars?

Personally, I’m not that interested in a bunch of autonomous race cars racing each other around a track. That would be sort of interesting, but I’d have to see some example of that to make me feel like it’s going to pique my interest. But I would be super-interested in showing up at the race track and seeing a fully-automated IndyCar rip off a 220mph lap. That would be awesome. And I don’t know the answer to this, but I would question whether or not you might actually achieve a more focused goal in developing technology in that space if there was a more focused and detailed challenge.


As part of the REVS Program at Stanford, there’s an autonomous Audi TTS (ABOVE) that a bunch of graduate students work on. The sole focus is to get the car to build and execute its own lap as close to the pace of a high-performing human driver, with the intention of exceeding the pace of a high-performing human driver. But that’s it – it’s just a one-lap thing. They’re not trying to get it so that it can race against other cars, they’re not trying to get it so that it is better over 10 laps than a high-performing human, or whatever. It’s just focusing on the single-lap time over a road course.

And the thing that I’ve been really impressed by is just how complicated that is; how many overlapping systems of math have to exist in order to allow the car to achieve its limits, and how it prioritizes what it is paying attention to corner-to-corner, and all this kind of stuff. It makes you realize quickly how complex what we humans do in that environment really is, without taking into consideration there being other cars around you that you have to account for.

So in some ways it would be interesting to compare and contrast the development of applied tech in something like Roborace from that perspective, where it’s having to figure out how to do a lot of things pretty well, compared against a more focused environment, whilst all of these programs are still very much in their infancy, to understand where the true value proposition really lies. Or, if there’s a reason to even have all those things.

There’s no doubt that Roborace has real value in terms of developing systems. There are a lot of things that happen on a racetrack that, if you can develop technology such that they work proficiently in a competitive racing environment, will have some serious value.

What remains to be seen is whether or not those vehicles designed to have to do so many things at once actually create an interesting show. To me, from a fan perspective, we don’t have that human element with Roborace. There’s no version of man versus machine aside from the engineers programming the cars, which is interesting, but not front and center in terms of what’s going on. They’re not playing an entirely active role in the on-track product.

Yes, they’re developing relevant, interesting technologies, and I think the cars look really rad – having [automotive futurist] Daniel Simon do that design was an awesome idea. And I’m really excited to see them in motion. But it’s possible them having to perform the actions that they have to perform while they’re on track together, and having to be aware of each other on the track, and all that kind of stuff, could make it so that it’s kind of like watching an amateur race. That’s where I wonder where the real shock and awe factor is going to come from.

And honestly, I think that’s missing from a lot of racing; that’s not saying that Roborace is going to be worse than a lot of other stuff out there. But that’s really why I want to watch a race, that’s what’s going to draw me to watch something new – that ‘holy s***, this is really impressive’.

So I’m really interested to see how it works. I recognize from the outside just how complicated the whole thing really is. So it will be really interesting to see how it develops, how it’s received, who it is received by, and – whether it is the platform of Roborace or a tweak of the Roborace concept – what it can teach us about automated systems and how they can be integrated into a performance environment.

A lot of the pieces of the puzzle are all there. But I’d hate to be the one jumping to a conclusion about how it’s going to work, and how it’s going to be interesting to people, and what about it is interesting to people.

In some ways it’s not entirely different to Formula E. You have the fact that it is clearly very relevant to today’s modern mobility environment in terms of where the development is happening, but it is done so in a way where there I don’t think there is a tremendous amount of tension in what’s happening. Formula E is not racing against internal combustion vehicles, and that would add an element of tension that would draw attraction to it.

I understand that the whole premise is zero emission events, so that wouldn’t work given the model, but just in terms of the on-track product and grabbing the attention of the causal motorsports fan, or somebody who is really interested in the development of these technologies, it’s hard to see the development of the technology when there’s nothing that it’s being compared against.

So if it’s not this sensory, visually spectacular thing that’s happening, and there’s not the tension of seeing it compared against the alternative, you might still have great success in that it’s a great platform for development from an industry perspective. We may end up seeing with Formula E and Roborace that that alone can sustain and be a major growth pillar for it within motorsports.

But there’s another part of it that’s like, if it can’t be interesting enough for people to want to engage with it and check it out and see it in person and watch it online and do all of that stuff, that would seem a definite negative.

It’s not unlike Formula E and its development – I’m interested to see how both of those continue to go and how they establish a model that either is replicated elsewhere, or is tweaked and built upon elsewhere to see how these things continue to come down the pipe. I have no doubt that motorsports needs more of that. I think it’s just a matter of how it is accomplished.

JR Hildebrand is an adjunct lecturer with Stanford’s vehicle dynamics lab. He will race the No.21 Ed Carpenter Racing Chevrolet in the 2017 Verizon IndyCar Series.

More RACER
Home