Inside the Revolution controller (maybe)

Here’s one MIT nerd’s theory on how the Revolution controller may work.

Here’s what my MIT nerd connection has to say about how the Revolution controller might possibly work. There are two sensors mounted on the TV (we know this from 1up.com’s report), presumably on top of the TV on the far right and left sides. The sensors connect to the Revolution console itself, and can receive both IR and ultrasound signals.

The remote control thingie (which I’ll call the “wand,” I guess) sends out simultaneous IR and ultrasound blasts. The Revolution detects the time delay between when the ultrasound hits the right sensor and when it hits the left sensor. This gives the position of the wand relative to the vertical centerline of the TV screen. (Imagine a plane slicing the TV in half vertically and extending out perpendicular to the surface of the screen. The time delay gives the distance of the wand to the left or right of that plane.)

The ultrasound signal travels only about 300 m/s, but the IR signal travels much faster—so much so that the delay between when it hits one sensor and when it hits the other is negligible. Since the IR and the ultrasound blasts are emitted simultaneously, the time difference between when the IR hits the sensors and when the ultrasound hits can be used to measure the distance of the wand from the TV screen.

Finally, the wand itself has something in it that can detect the angle of the wand relative to the plane of the floor. (Well, relative to the direction of gravity really, but presumably the floor is nice and level.) This could be something like the sudden motion sensor chip in PowerBooks and other laptops, for example. But some of those chips seem to "re-center" themselves if you hold the laptop very still, so maybe the technology is slightly different.

Either way, this system gives the Revolution enough information to do something sensible. The distance from the screen alone gives one degree of control. The left/right offset gives another degree of control. The distance from the screen combined with the left/right offset gives the angle (along a single axis, anyway) of a line drawn from the entire wand as a unit to the center of the TV screen. And the wand itself can sense its “tilt” along all two axes.

There could be one more sensor or system at work to perhaps give an “up/down” position relative to a plane that slices the TV in half horizontally, but that may not be necessary. The IR/ultrasound signaling plus the chip in the wand itself is enough to explain the controller demos given to the press. It seems plausible to me.

This article originally appeared at Ars Technica. It is reproduced here with permission.