Tilt Radio Control

david43085

Looking at the project of the week with the tilt radio control, I am wondering:· Imagine you were trying to control an airplane with this tilt remote but was also sitting inside the airplane.· The airplane is pitching and rolling at the same time you are trying to tilt control it, so any movements you make with the remote will have to be in condsideration of what the movement of the plane is relative to horizontal.· One solution is to have two of the accelerometers, one that is mounted to the plane, and other that you are holding in your hands.· The Stamp would have to be able to receive data from both accelerometers, calculate that difference in position between the two, and apply the net result as the absolute command for the control.· Is this possible?· Can the stamp recieve two signals, calculate the difference, and apply the command?

Just Curious

Dave

·

SRLM

First off, this sounds incredibly dangerous to try and implement. Get something wrong, and you'll end up with unstable equilibrium (at least until you become permanently stable stuck into the ground).

But, theoretically, yes, the stamp can take two numbers and perform calculations on them. However, you have to keep in mind that the control system in the example provided uses gravity to provide a reference value to calculate tilt. When it's in the moving system, the gravity becomes more of a nuisance and not useful for calculating tilt. Therefore, the system proposed would not work.

ProxyProdigy

In a much safer scenario, I used the accelerometer to build my own controller for a video game I'm designing. It's still stuck in my dev board until I find time to create my own board. But, it works along the same lines. I simply tilt my board, the stamp sends the tilt coordinates to my PC where I have a port listener waiting for commands. The game is written in XNA C#. In the game, you can select from several different vehicles: plane, tank, car, boat, etc... Playing the tank for instance, you can tilt the controller forward to drive forward, left to turn left, etc.... When I get around to creating my own PCB, I'll also add joysticks and triggers for firing the guns.

Anyways, I thought I'd share this since I feel like it's a much safer scenario. Although, instead of a plane, it would be kinda cool to build a terrain vehicle of some sort that was controlled using the method that david43085 discussed.

sylvie369

I can't tell for sure whether or not you're aware of this, but in a properly coordinated turn, everything inside the airplane will "believe" that down is directly towards the floor of the airplane. An accelerometer will not somehow sense what direction the ground is actually in.

www.theatlantic.com/unbound/langew/turn.htm

Your accelerometer mounted on the airplane itself would always believe that it was level, just like the drink in your glass sitting on the tray table doesn't slosh out the side when the plane turns smoothly.

Post Edited (sylvie369) : 1/30/2009 2:49:40 PM GMT

david43085

Please note I am not attempting to build an airplane control system, just an example of using two accelerometers and basing the control off the net difference of the two.

As far as the the article posted by sylvie369, I was not aware of that, that was a very interesting realization. Thanks for the post.

sylvie369

david43085 said...
As far as the the article posted by sylvie369, I was not aware of that, that was a very interesting realization. Thanks for the post.

There's a nice demonstration video here:

www.alexisparkinn.com/photogallery/Videos/707%20Roll.mpg

in which Boeing test pilot Tex Johnson does a complete barrel roll in a prototype 707 airliner. He is interviewed, and you'll hear him say there's no risk at all, and that it is simply "a 1G maneuver". That's the technical way of saying "my drink wouldn't have spilled", despite the fact that the airliner goes completely upside-down.

Again, an accelerometer mounted in that airplane would register essentially no tilt at all. A gyro, on the other hand, would easily be able to detect the maneuver.