g-force simulation for VR driving concept

Post by **MSat** » Wed Sep 05, 2012 4:55 pm

From what I can tell, there is no practical way to simulate all the forces that would be exerted on your body from driving, but I think there might be a relatively simple to get fairly close. It involves using an HMD with head tracking, and a cockpit that pitches forward/backward & rolls side to side. The idea is that as long as the HMD orientation remains level, pitching the cockpit back for instance would simulate the sensation of acceleration, or deceleration if pitched forward. Likewise, rolling the cockpit left/right could simulate the forces of turning. Given that the image to the HMD flat in relation to the virtual world, I think a person would be fooled into believing that they're feeling the correct sensations rather than being pitched or rolled relative to gravity. Any thoughts on this?

Post by **brantlew** » Wed Sep 05, 2012 5:09 pm

There are a number of companies and DIY garage projects that make motorized cockpit platforms just as you described. Luckily, that's one thing that's already solved and available (if you've got a few thousand bucks).

Post by **cybereality** » Wed Sep 05, 2012 5:28 pm

Well if you wanted to simulate g-forces you would need a really large space and something like the guys at NASA train with (human centrifuge):

That is probably out of the question for a DIYer. But they do have motion simulation seats you can buy for around $10-15K. Probably a lot less if you build yourself.

http://us.simroom.com/us-store/us-sim-racing-products

Post by **android78** » Wed Sep 05, 2012 5:42 pm

They seem to be going out of fashion now, but there used to be a lot of simulators in arcades that would do what you're talking about. Basically, they have a vector for the direction of acceleration, then tilt you such that gravity is pulling in the opposite direction. It's not accurate g-forces (since the maximum acceleration you ever feel is 1G), but it does a reasonable job of fooling you into thinking that you are moving at least. Most of these had a screen in front of you.
I'm not sure where all of these arcade machines ended up, but if you could get one, they would be a great platform for what you are talking about.

Post by **Chriky** » Wed Sep 05, 2012 5:49 pm

Some versions of Outrun 2 SP had that, it was pretty good.

Does anyone have any experience with hydraulics or however those things work?

Post by **android78** » Wed Sep 05, 2012 5:55 pm

How about this:
http://familyentertainment.biz/product. ... 251&page=4

Post by **android78** » Wed Sep 05, 2012 5:58 pm

... another cool hit here:
http://www.x-simulator.de/forum/2dof-wi ... t2405.html

Post by **MSat** » Wed Sep 05, 2012 5:59 pm

I don't think a centrifuge would work all that much better (if at all) than a simpler motion platform - You might be able to achieve greater than 1g, but you have an issue with accelerating/decelerating such a large object. And as far as space training centrifuges go, I think those were all just 1D to simulate rocket acceleration.

I've played older arcade games built on motion platforms, but they were all gimmicky. If the concept is legitimate, it might be something I'd like to pursue.

Post by **cybereality** » Wed Sep 05, 2012 6:05 pm

I've been on some pretty cool motion simulators, like the stuff they have at Universal Studios in Florida. I remember one, I think the Star Wars one (Star Tours) was very good. Even better was the Back to the Future ride. Totally amazing, it felt like you were there and it wasn't even 3D. It really did feel like you were flying around, it works.

[youtube]http://www.youtube.com/watch?v=MJGZogaGCLw[/youtube]
Skip to 1:05 for the good stuff.

Post by **MSat** » Wed Sep 05, 2012 8:19 pm

I've never been on those rides, but I have heard from numerous people that they are indeed awesome. The only thing that I see an issue with is that the visual "ride" is created with full consideration of available motions. So if the platform has extended its motion in one direction, they have to create the video so that the next motion moves the platform in the available direction. This is fine for a scripted scene, but it wouldn't work anywhere as good for interactive content - the motions would have to be way more limited. Granted, this is less likely of an issue with a driving simulator.

@android78

I've played the afterburner game - totally gimmicky - and that's when I was young and naive

It didn't do variable motion - just bounced around from extreme to extreme. A good motion platform needs a large degree of travel. In order to get a full g, then it has to move 180 degrees in each axis, though that might make it a bit bulky.

Post by **alloy** » Wed Sep 05, 2012 9:52 pm

Funny how i was looking into these things last night.. There are a lot of different designs all expensive as heck yes.. but if i could choose... this is the one i want..

This one D:
[youtube]http://www.youtube.com/watch?v=Arz_BIPC1ek[/youtube]

And english
[youtube]http://www.youtube.com/watch?v=TwoapW20NHI[/youtube]

Post by **brantlew** » Wed Sep 05, 2012 10:12 pm

Wow! I hadn't seen that before. Amazing piece of equipment.

@bobv...there's your college project.

Post by **cybereality** » Thu Sep 06, 2012 9:29 pm

That's pretty sick! Looks really responsive too.

Post by **FingerFlinger** » Thu Sep 06, 2012 10:18 pm

I want to play TIE Fighter with that SO bad...

Post by **junglej1m** » Fri Sep 07, 2012 8:03 pm

I like this idea.

There are consumer/diy motion platforms with widescreen TV's attached and the problem with these are that you can still see the room around you making it hard to create the impression of lateral g loads due to the fact that your peripheral vision will see that you're rolling. Using a HMD makes it much easier to trick the bodies vestibular system.
As someone who has to sit in a full motion sim every six months I've experienced the limitations of simulating g's in this manner... however for me it feels real enough!

Another benefit of using a HMD is that they are lighter so the motion platform hardware (hydraulics/motors) can be of lighter construction.

Post by **PalmerTech** » Fri Sep 07, 2012 8:22 pm

If you guys really wanna nerd trip, check out the R360:

They pop up from time to time for just a few thousand dollars, usually with severe mechanical problems.

There are a couple guys who refurb them to brand new condition if you Google around, though. There was an R360 on my local Craigslist about a year ago, but I could not justify $2,900 for the condition it was in.

Motion platforms are another technology that COULD get you a good experience for very cheaply, if the right person came along willing to sink the time and money into making a reproducible product. A pretty decent sim could cost as little as $600 if 1000 people bought it... Kickstarter, anyone? I would back it!

Post by **junglej1m** » Fri Sep 07, 2012 8:49 pm

VR platform or child torture device!?

[youtube]http://www.youtube.com/watch?v=eSJG5e0s43c[/youtube]

Post by **cybereality** » Fri Sep 07, 2012 8:55 pm

Wow! That thing does look pretty cool.

Post by **MSat** » Fri Sep 07, 2012 10:18 pm

Alloy, that thing is legit! All that's missing from that setup is an HMD.

Check out that CRT in the R360! LOLOL

I agree that a motion platform could probably be produced for around $500. I'm thinking yaw motion is unnecessary for flight/racing simulation, does anyone concur?

Post by **junglej1m** » Fri Sep 07, 2012 10:49 pm

I think yaw would be great for racing sims. To get the sensation of oversteer would be a huge plus to the immersion for me.

Flight sims - no need for yaw unless you're going to use it for engine failure practice or want physical feedback of uncoordinated flight.

Post by **FingerFlinger** » Sat Sep 08, 2012 12:16 am

But surely not 360 yaw, right?

Post by **MSat** » Sat Sep 08, 2012 11:42 am

I'm trying to imagine if yaw is even important for simulating oversteer. I think you're really only sensing the lateral acceleration forces, and the visual cues to indicate that you're oversteering. I suppose there might be a slight difference in g-load on your body than perhaps on your legs, but this would be so minimal, and probably irrelevant.

Post by **pierreye** » Sun Sep 09, 2012 2:37 am

Check this out. 15 ultimate gamer chair. The one I'm interested in Force Dynamic 401 which cost USD 60k! With HMD and headtracking, this would be the ultimate for driving and space simulation

Post by **MSat** » Sun Sep 09, 2012 3:20 pm

Hmm.. According to Force Dynamics, they claim that yaw motion is important for giving the sensation of oversteer. I just don't imagine there being any substantial disparity between the inner ears under such events. I have my doubts that it would be worthwhile to implement, at least not if price and simplicity is a concern.

Post by **calebkraft** » Tue Sep 11, 2012 6:23 pm

There are passive systems too. usually used for flight sim since the control method is a single stick. You could, however replace it with a wheel and some pulleys. Unfortunately, there's literally no connection to the game itself so it is possibly not that immersive.

[youtube]http://www.youtube.com/watch?v=2t61nFAuF7c[/youtube]

Post by **MSat** » Tue Sep 11, 2012 7:51 pm

The cool thing is that even a simple setup can yield a great sense of immersion if the game is designed to avoid the moving platform's weaknesses. Can these things be the basis for the revival and evolution of the arcade? The VRcade? To once again play amazing games you can't play at home.

Post by **brantlew** » Tue Sep 11, 2012 8:23 pm

That would be awesome if VR arcades became popular. The only problem I see is that VR naturally just requires more space. Arcades could stay afloat because they were able to jam 40 or 50 machines into a smallish room. But if you can only fit 10 machines then the economics is tougher. Maybe Dave & Buster's could run with it though since their business model revolves a lot more around food and beverages.

Post by **MSat** » Tue Sep 11, 2012 8:55 pm

I hate to say it, but the premium that would be charged would probably offset the lower number of machines. Also, the machines don't necessarily have to be very large - some could even be smaller than those motorcycle games, because you don't need that giant box sitting in front of it. It's also possible for some of the machines to have different games making use of the same motion platform.

If it's awesome, there has to be a market.

Post by **PalmerTech** » Wed Sep 12, 2012 4:42 pm

Also, laser tag, paintball, and airsoft arenas all stay open with low prices and high square footage.

Post by **cadcoke5** » Wed Sep 12, 2012 8:02 pm

I have long had an idea in my head for a low-cost motion system. It would not be unlike the systems posted earlier here, but the idea is that the chair is hung from a frame, rather than in a gimbal mount.

The main benefit in my mind is that the rotation point is not in the center of the user. This would allow the sensation of acceleration, with greatly reduced sensation of rolling over. And while roll may be beneficial in some situations, it is certainly not the main force you would experience in a typical simulation. Also, the very act of shifting the person forward to get an acceleration, would generate an acceleration itself.

I was picturing a main frame sort of like this Yoga swing (why would people want to hang from rubber bands?). It suspect it would be too light-weight of a frame for real motion simulation, but it is a good example of what I am thinking.
http://www.yogaswings.com/2010/10/28/omni-gym/

Then, add a chair like what is in the middle of these gimbal type simulators,
http://mydreamflyer.com/# (this is a non-actuated chair)

For the actuators in my concept, there would be 3 pull-only actuators (which don't have to be as strong as a system needing to push). Twist would be more complex, but probably doable with only one actuator. I can also imagine the benefit of a 4th actuator that gives some vertical pull. This would provide the sensation of going over bumps. I wonder if one or two feet of vertical would be beneficial for the really rough rides?

A HMD would be the final part... well actually the software to make it all work together is probably the final and hardest part of any motion platform. Even if you get one game to work, you probably wouldn't want to have to do that for every game.

Joe Dunfee

Post by **MSat** » Wed Sep 12, 2012 8:38 pm

There's definitely no generic motion platform that's suitable for all games.That's why the best results will be had when you design the game around the platform. In a mech game for instance, you might only need a platform you stand up in that bobs up and down, yaws, and maybe moves back and forth a little. Of course, that would be useless for a flight or driving sim. But since it's ideal for a mech game, the results could be amazing. Though this isn't feasible for the average home gamer, it's certainly possible for a VRcade machine developer.

It's funny that Palmer mentions places that require large indoor spaces. I had been thinking about all the various empty strip malls around me, and imagining how awesome it would be if some of them were VRcades.

Post by **Namielus** » Wed Sep 19, 2012 7:48 am

The g-force simulation and the motion simulation works alittle different. Lets say you do a barrel roll with an airplane, and the cockpit rolls the same way. Simple motion simulation.

However while accelerating or slowing down, the vehicle is changing velocity but your body wants to stay at its own velocity.

In a car, when breaking, you are are pulled forward toward the steering wheel, while accelerating you are pushed back into the seat.
Those g-forces can be imitated with gravity.

But what about the accelerometer in the HMD, all that motion can be interpreted as head movement.
I think in a g-force simulator all the head motion has to be tracked relative to the seat.

The actual motion in the cage, is the opposite of the effect you feel and direction you move in the game.

I mean when your head bobs around from all the gravity/g-force, its actually okay that your view wobbles around because that is the intended effect. But as the cage rolls back to simulat acceleration it might potentially think you are looking up etc.

Post by **MSat** » Wed Sep 19, 2012 1:20 pm

^^ That's what I had in mind. Since the system knows there's going to be a change in pitch or roll or whatever, you can negate the associated sensor readings from the hmd, or as you suggested, track head movements relative to the platform, rather than gravity. Taking that design further, if you roll the platform to the right, you can simulate centrifugal forces when turning left, and vice versa.

One thing I was thinking that may be necessary is to have the platform rotate along the axis of your inner ear, rather than somewhere along or outside of your body. In your example, say you simulate braking - what's going to happen during the transition period is that your upper body will rotate forward, causing acceleration in the wrong direction - which you will sense in your upper body, and inner ear. If you pivot it around your inner ear instead, the transition period will simulate the appropriate forces on your body and inner ear at all time, making the rotation less perceptible.

Post by **Namielus** » Wed Sep 19, 2012 2:01 pm

hey!

The design involves roll, I just removed it so the pitch would be visible, lol.

This is a model withouth the steering wheel mount.

The whole design was supposed to be driven by stepper motors instead of actuators, but could alternatively be driven by actuators as well.

I would really like to build this, but I would not know how to program/drive the stepper motors.
Just the build part

Post by **MSat** » Wed Sep 19, 2012 3:46 pm

Namielus wrote:hey!

The design involves roll, I just removed it so the pitch would be visible, lol.

This is a model withouth the steering wheel mount.

The whole design was supposed to be driven by stepper motors instead of actuators, but could alternatively be driven by actuators as well.

I would really like to build this, but I would not know how to program/drive the stepper motors.
Just the build part

Oh, cool! Is that 360 degrees in pitch and roll?

How good or poor is your electronics skills? I have worked with motorized feedback loops on a few projects. I'm sure if you looked, you would find tons of schematics and code for all sorts of microcontroller platforms that will do exactly what you need. If you use the Parallax Propeller (http://www.parallax.com/tabid/407/Default.aspx), I can help you out with both.

A big thing to consider when using any kind of axial motor in such a motion platform is how will you keep the load low or nonexistent when it's in a static position? The easiest way would be to rotate around the center of gravity, or use simple counterweights. Having thought some more on what I said about pivoting around the head, I think, it should actually pivot _above_ the head for pitch and roll. The biggest issue with this is that it takes up more room.

Post by **Namielus** » Wed Sep 19, 2012 5:36 pm

I can do soldering and stuff like that, but I have never programmed anything other than simple HTML and PHP.
That has stopped me from some diy-projects before, but then again I have never tried learning it.

Yes its 360, or it could actually keep on rolling forever on both axis.

There is something wrong with the physics simulation at the moment, I will try to get an animation of the rolling/pitch action
so you can at least visually estimate the balance.

The design is made so that the center of gravity is as close as I could get to each pivot point, or so was the idea.
It was just based on visual references.
And each pivot point could rest on heavy duty ball bearings -
Therefore I think other than inertia of the whole thing there is very little weight to drive.
As I have understood stepper motors before, they move the distance they are told in increments then they stop.

It could be fitted with weights to balance it out almost perfectly, except for the shifts of weight of the player.
Then it would take almost no effort to start spinning it.

Post by **MSat** » Wed Sep 19, 2012 5:56 pm

You probably want it to be quick and responsive, so you'll need a decent sized motor to accelerate and decelerate that much inertia. Depending on your gear reduction, you might be able to get away with using a DC motor and PWM driver - car starters or some golf cart motors might be suitable for this. If you want to use steppers, which is probably most ideal, an automotive alternator can be easily modified to work as a cheap and powerful 3 phase stepper.

Post by **Namielus** » Wed Sep 19, 2012 6:04 pm

Easier to see pivot points, keep in mind the side view is missing the rig holding the wheel and also some other things, so the balance seem more off than it really is on the pitch axis.

also, alot of the support tubes has been removed to allow a better view.

Post by **android78** » Wed Sep 19, 2012 6:08 pm

I think that MSat is right with regards to where the pivot point should be. If the pivots line up with the middle of your head, the actual rotation will be less noticeable. The problem I see with the latest design you posted is that, when you accelerate, the cage will pitch up - while the cage is rotating, the acceleration of your head will be backwards, not forward.
For this, you will need a much larger frame (to allow the user to sit well below the pivot) and counterweights so as not to strain the motors/actuators.
Looks like a great design though... will probably be pretty expensive to build though.

Post by **Namielus** » Wed Sep 19, 2012 6:17 pm

I think pivoting a person at the head, with racing seat + structure holding up the steering wheel would be very heavy indeed.

If you know how to bend pipes and weld, and you have a friend in an industry that use lots of metal or aluminum this wont be that that expensive at all to build. I have no idea about the electronics parts, as I dont know what strength it needs to be.

I can check the prices of the materials, but its per kilogram, so I have to decide for dimensions/thickness and calculate the kg per meter and so on.

I would not guestimate it to be over 700 dollars in raw material minus welding rods.

Possible materials:

Stainless steel= heavy , cheap, easy to work with
aluminum=light weight, cheap, nightmare to work with (both bending and welding)
Carbon fiber=extreme lightweight, crazy expensive, easy to work with, and would require some changes to the design