How will the Revolution sense the controller?

[ninzel]

Just wondering what your theories are on how it will sense not only it's position but it's orientaion in space.
I figure to sense it's overall position it will use some form of echo location. For it's oriention it could be either echo location but more likely an internal sensor will send back data as to it's oriention. So you will have both types of sensing going on at once.
Theories??

 

[nintenho]

I think it uses infrared. Some people say that it has gyroscopes (little spinning discs that sense any slight movement) but I'm not so sure how cheap that is and I heard that mouses that use that are too inaccurate. The revolution has a special bar that makes infrared signals or something that the controller senses. The controller doesn't look at the screen, so that's why you can change the sensitivity for things like fps games (20 degree movement=60 degree movement).

 

[ninzel]

I think it uses infrared. Some people say that it has gyroscopes (little spinning discs that sense any slight movement) but I'm not so sure how cheap that is and I heard that mouses that use that are too inaccurate. The revolution has a special bar that makes infrared signals or something that the controller senses. The controller doesn't look at the screen, so that's why you can change the sensitivity for things like fps games (20 degree movement=60 degree movement).

I was thinking it would be RF,but since this is the kind of sensing that has to be continually updated I guess infared might be better.
So you think it's the bar that is sending out the signal and not the the other way around? That makes sesne for battery life. If the position is constantly being udated it would make sense for the bar to be constantly putting out the signal and be tracking the position of the wand.
So the bar could put out some type of pulse, and it would bounce off something in the wand?

Gyros make sense for it's orientation.

 

[nintenho]

There was this thread that you posted about an IGN article about the sensor bar:
http://revolution.ign.com/articles/694/694785p1.html

 

[ninzel]

YA but that doesn't really get into the gut's of the system.

 

[nintenho]

oops, forgot about this thread. Maybe Nintendo wants to keep that a secret so that Microsoft or Sony don't copy it. Mebbe?

 

[Shifty Geezer]

Single chip Gyros and accelerometers bought in bulk would cost in the region of $10-20 I expect, looking at publicly available prices of components. The Revmote will have to broadcast to send information on button presses and gyro states. It won't be passive (how many TV remotes do you know that don't need batteries...), The bar receiver will pick up the transmissions and the difference between time taken to reach to left and right receivers gives a triangulation signal from which position can be worked out. So you have triangulation of a signal, gyro/accelerometer state data being broadcast, and the screen-reading (lightgun) method. There could be two transmitters from the revmote, one at the front and one at the back, to determine rotation in the Y axis.

 

[ninzel]

Single chip Gyros and accelerometers bought in bulk would cost in the region of $10-20 I expect, looking at publicly available prices of components. The Revmote will have to broadcast to send information on button presses and gyro states. It won't be passive (how many TV remotes do you know that don't need batteries...), The bar receiver will pick up the transmissions and the difference between time taken to reach to left and right receivers gives a triangulation signal from which position can be worked out. So you have triangulation of a signal, gyro/accelerometer state data being broadcast, and the screen-reading (lightgun) method. There could be two transmitters from the revmote, one at the front and one at the back, to determine rotation in the Y axis.

Yes I figured for the orientation and button presses the Revmote would have to broadcast a signal. But just to sense it position in it's 3d box(if that's the right way to look at it), does it have to broadcast or can it be located by pulses from the bar bouncing off something in the Revmote.
Oh and I didn't realize it would act like a light gun. I thought aiming would be based on it's position. But that makes sense,I was wondering what would happen for aming if the Revmote went out of it's 3d box in space. No need to worry about that if the aiming is done like a light gun.
Thanks for the info.

 

[Rodéric]

Yes I figured for the orientation and button presses the Revmote would have to broadcast a signal. But just to sense it position in it's 3d box(if that's the right way to look at it), does it have to broadcast or can it be located by pulses from the bar bouncing off something in the Revmote.
Oh and I didn't realize it would act like a light gun. I thought aiming would be based on it's position. But that makes sense,I was wondering what would happen for aming if the Revmote went out of it's 3d box in space. No need to worry about that if the aiming is done like a light gun.
Thanks for the info.

I'm also curious, I would prefer to have a passive mechanism, like a radar, to save batteries.
The Revolution's pad is NOT a lightgun, it relies solely on the sensorbar.

 

[pipo]

Oh and I didn't realize it would act like a light gun.

It doesn't as far as I know. I'm not sure Shifty says it does either...

Edit - Ingenu beat me to it.

 

[Squeak]

My conjecture is that the revmote contains an IR camera (a bit like the one in an optical mouse), a gyro and a bluetooth transmitter.

The camera sees the two IR diodes on the "sensorbar" and is then able to able to tell x, y and z by the shifting and explosion/implosion of the two points.

The gyro reads the tilting of the controller, and all the data incl. button presses is sent to the console with bluetooth.

That way the revmote will only have to send one signal (cos we know it will use bluetooth for the controllers).
And I also think the camera method is more precise than any of the other possibilities.

 

[ninzel]

I'm also curious, I would prefer to have a passive mechanism, like a radar, to save batteries.
The Revolution's pad is NOT a lightgun, it relies solely on the sensorbar.

Yes it could be partly active where it has to be, and passive where it can be. That would make sense.

 

[Shifty Geezer]

You could do like RFID tags and have a transmitter powered from by a transmission signal from the Revolution, but if you've already got Wifi of some sort in there for sending button and gyro data, I think that'd be an unneccesary complication without a great deal of benefit.

And I was wrong about the 'light-gun' feature. I think I got that idea from article(s) talking about it being able to determine where you're pointing at the screen, and that specific reference gave me the wrong impression. Looks like it's all derived from positioning without any reading the TV screen.

 

[Rodéric]

I think it'll be using RF just like the NGC pads are.
I can see a radar + gyro.

I don't see how an IR camera would work at all... when I point left or right or up or down, it wouldn't see the sensorbar, so it wouldn't be effective.

The sensorbar is rather a triangulation system, two sensors here, two more on the pad, and you should get almost everything you need, add gyro for precision and you should get something nice.

Anyway whenever someone's get a Revolution I bet we'll see pics of both the pad & the console open up on the web

 

[Squeak]

I don't see how an IR camera would work at all... when I point left or right or up or down, it wouldn't see the sensorbar, so it wouldn't be effective.

Rubbish. You wouldn't need to point outside the screen anyway.

The sensorbar is rather a triangulation system, two sensors here, two more on the pad, and you should get almost everything you need, add gyro for precision and you should get something nice.

Would you stop it with the triangulation already! It won't work.
You would need a third sensor at a fixed position.

 

[Phil]

What I am thinking is, how easy would it be to throw of the calibration of the controller by simply changing the rev-mote to a different position either by standing up, stepping a step either to the left or right... It may not make a big difference, except for games that require very high precision...

 

[Guden Oden]

and the screen-reading (lightgun) method.

Doesn't work on DLP, LCD or plasma displays, and perhaps not on projection CRTs either...

There could be two transmitters from the revmote, one at the front and one at the back, to determine rotation in the Y axis.

I'd think the internal accelerometer would take care of that. Since the revmote is quite tiny (seems to be roughly 15 cm long), I'd think any doppler-shift detector would have to be incredibly precise to figure out rotation on such a small object... Besides, two transmitters = as much as twice the power draw, twice the possibility of interference... With 4 controllers it would be 8 signals + quite possibly wifi traffic on top of it all. Crowded!

 

[Shifty Geezer]

Rubbish. You wouldn't need to point outside the screen anyway.

What about a baseball came where you're pointing away from the screen ready to bat? I can think of a fair few situation wher you dont want to limit the controller to having to point at the screen.

Would you stop it with the triangulation already! It won't work.
You would need a third sensor at a fixed position.

Wrong. You need two fixed positions and the third point of the triangle is the moving one in the Revmote. This means you aren't constrained to any box playing area or limits of orientation, and would offer the ideal solution to the free-roaming controller. the only thing missing is altitude or Y position of the controller, which I guess would be worked around using controller tilt to determine the input you'd associate with this. eg. Instead of lowering the controller in a snowboarding game to duck under a branch, you'd tilt the controller down.

I'd think the internal accelerometer would take care of that. Since the revmote is quite tiny (seems to be roughly 15 cm long), I'd think any doppler-shift detector would have to be incredibly precise to figure out rotation on such a small object... Besides, two transmitters = as much as twice the power draw, twice the possibility of interference... With 4 controllers it would be 8 signals + quite possibly wifi traffic on top of it all. Crowded!

I can go with that. The downside to this approach is no absolute twist positioning, creating the prospect of accumulated inaccuracies. I'm not sure how you'd work around that.

 

[_xxx_]

It will probably use yaw rate and acceleration sensors. The stuff that ESP systems in cars use, essentially. It determins the angle speed (in 3D space) and the speed (as integral of the accelelration). That's how I'd implement it, I don't know what they'll use.

Some examples:
Automotive sensor example
http://www.spp.co.jp/sssj/silicon-e.html

The position is communicated to the console within some proprietary protocol through whatever wireless interface with enough bandwidth, IR could suffice I believe. Although I think it'll rather be something non-directional.

And:

mechanical gyroscopes

ROTFLMAO!!! Hilarious idea! Though technically, the silicon yaw-rate sensor have little comb-like structures which bend through inertia and change the voltage. That difference is measured and delivers the acceleration accordingly. So it is kinda mechanical as well.

 

[Shifty Geezer]

Aren't large gyros all the rage at the moment for strengthing fore-arms? Could add some broader appeal to the revmote as you play against 3 G gyroscopic forces!