Ditch the sim rig and use your car instead (2019) | Hacker News Reader

Ditch the sim rig and use your car instead (2019)

outlandnish.com

111 pointsalavry20 comments

Blog post author here - Didn't expect a hack I did 5 years ago to show up here!

This project was a weekend project that I took on. I used to professionally reverse engineer cars for building self-driving research vehicles for companies.

Here's a demo of it using a VR headset that I made shortly after the blog post: https://youtu.be/jfWcgWdSK28

To answer the other questions, I made it work directly with PCs as a USB accessory. It doesn't require the Xbox gamepad in that variant.

I never bothered building out the FFB for the steering wheel - but it would be a nice addition. That said, slightly jacking the nose of the car up gave enough resistance on the wheels but it was still easy enough to turn.

It's still been a great tool for me to get seat time. Recently, I just taught at Spa-Francorchamps in Belgium and the Nürburgring in Germany. I used this sim rig setup to get properly acquainted with the tracks before teaching in person.

Last, along with a production company in NY, we used the same technology in a big rig truck to make it into a super realistic truck simulator for Pilot Logistics. I think there's media on it floating around the internet somewhere.

Cheers and happy to answer any questions!

This looks like a fun project, but that's about where I'd draw the line. A few issues right off the bat:

- In all such comparisons, people conveniently leave out build and research time/costs. I suspect that at least a few hundred hours went into this project, which would significantly increase the cost, if we account for hourly rates. Whenever I've run the numbers on similar custom hardware projects, I've come to the conclusion that if I'd simply worked the hours as overtime, I would be getting a much better product even if I bought high-end equipment, and quite a bit cheaper. So, these are strictly for people who enjoy the journey, and as such I find such comparisons insanely misleading.

- What exactly are the wheels doing, when you're turning the steering wheel? Is the car stationary in the garage, or is it lifted? I probably wouldn't want to spend dozens of hours turning my car's wheels in a stationary position, it puts strain on things.

- While on face value it seems like this should be about as real as you can get with sim equipment, I'd argue that it's actually fairly low- to mid-tier. Neither the steering nor the brake pedal will have any adaptive feedback, which is fairly easily obtainable with direct drive wheels and in some high-end pedals.

So all in all, quite a cool project, but not particularly useful or reasonable, IMHO.

>I probably wouldn't want to spend dozens of hours turning my car's wheels in a stationary position, it puts strain on things.

He blew his engine, at a raceway. Something tells me that gently steering back and forth on his garage's polished concrete floor isn't something he is concerned about.

Learning and building are very rewarding which tips the cost-benefit balance here into profit – for me personally!

Re. driving feedback from the controls, near the end of the post:

> Next Steps:

> Force feedback support using the electric power steering and the ABS module.

> Neither the steering nor the brake pedal will have any adaptive feedback

I think this is probably the biggest disadvantage compared to using an actual sim wheel. In the linked video demo it looks like the driver is struggling to apply the right amount of steering and correcting after the corners. At times the game footage looks more like someone playing with a gamepad than a wheel.

I don’t like that the Xbox adaptive controller is necessary for this. Microsoft gets praise for how it enables alternate controllers - and it’s great for DIY - but that is only necessary because they use a cryptographic protocol to disallow allow unlicensed controllers in the first place. There’s no reason home brew or third party USB game pads should not work on the Xbox in the first place. Sony is the same. Nintendo at least, while they don’t ‘support’ third party pads, don’t use cryptography to disallow them.

I don't own a console, but I always imagined the motivation was partially to level the playing field for competitive multiplayer games.

I recall playing COD4 online (on PC) and someone had hooked up a controller with rapid-fire and was pwning the noobs, quite literally.

I strongly suspect that the reason is more to do with producing artificial scarcity of hardware so that they can continue their loss-leader pricing strategy. Typical Xbox controllers go for £50 new; that's a quarter of the price of the Xbox itself, yet all one gets is a piece of moulded plastic with a few buttons on it!

PS. COD4 is apparently set in 2011 - do players not have automatic weapons anyway?

> that's a quarter of the price of the Xbox itself, yet all one gets is a piece of moulded plastic with a few buttons on it

While I agree that they probably are having a decent profit on accessories like controllers, electro-mechanical parts are surprisingly expensive in my experience. A good potentiometer or joystick can cost way more than the rest of the circuit that uses them.

> COD4 is apparently set in 2011 - do players not have automatic weapons anyway?

Sure, but with games being games, to make single-shot weapons interesting game designers usually make single-shot weapons more accurate and do more damage as a trade-off for their lack of automated fire. Of course with a rapid-fire controller you can get best of both worlds in that regard.

I think it's a design mistake to allow single-shot weapons to fire as quickly as your input device allows.

That is actually how Cassidy works in Overwatch - the tradeoff is the exceptionally small hit radius of the projectiles. But there are other games where your ability to hit buttons quickly matters. Fighting games, for example. The recent kerfuffle around Razer's snaptap keyboard (and the software version with "Null binds") have well demonstrated that technological advantages in the controllers bring advantages into the games as well.

On the flip side, I've played games which ignored the "fire" input for N milliseconds after triggering. Works well to prevent the issues mentioned, but was highly annoying as I would frequently click just a millisecond or so too early, and miss the shot.

There's been a recent article [1] about a very similar students' project at Darmstadt University. Interesting to see that the post referenced here precedes this by about five years.

[1] https://www.heise.de/en/news/Car-hacking-real-cars-turned-in...

I’ve been told that a local wealthy fellow turned his preferred racing car into a sim rig when he got too old/rich to risk racing it for real. It’s in his garage and is up off the ground. The car has been completely redone to make it a realistic sim machine—screens in the windshields, vibration motors, pedal sensors, the works. The person that told me that didn’t know how much they’d spent but it was in the six figures.

Reminded me of the Ridge Racer Full Scale recovery/restoration story: https://arcadeblogger.com/2022/11/20/the-last-ridge-racer/

In which a very limited edition version of Ridge Racer in arcades used a full-size Mazda MX-5 (technically the Eunos Roadster) shell and panoramic projection.

That's one way to get a workout. Without the engine, there's no power steering, right?

BRZ has electric power steering assist. If he puts the wheels on alignment pads he should be able to move the wheel easily. I’d still be worried about wear and tear on the components (you really don’t want your tie rods to give out at speed) but it’s a fun weekend project.

He says ge jacked up the car to give less friction. The wear should be mininimal as no actual cornering forces are involved, which are suprisingly large, ofren exceeding the weight of the car in (momentary) lateral force. While this will still cause some wear, it is negligible. I would imagine the ideal would involve a force feedback module and completely lifting the wheels from contact. Perhaps even utilizing the cars own electric power steering.

What about the simulated acceleration on the body by tipping the sim rig? That's going to be a tall ask in a real car sitting still.

interesting project indeed, but lack of steering feedback might already be a deal breaker for most. in terms of emulating the car you'd practice virtually for real-world driving, there is probably a nice market for easy-to-convert platform for getting input from cars.

Reply on news.ycombinator.com