May 02, 2011, 10:32 PM Registered User United States, TX, Austin Joined Aug 2007 375 Posts Discussion Simple HH Gyro for Yaw Sensing I originally posted this in DIY electronic but it applies here too: I'm currently in the process of creating a diy head tracking system mimicking the setup seen here. However I've run into a bit of a roadblock and could use some collective help. In the link above you'll see he's using an accelerometer to calculate X and Y values that get converted and sent to the servos via the rx/tx. What I need is to sense the yaw, rotational movement of my head which the accelerometer doesn't do. I've thought about using a rate gyro but that wouldn't do position sensing and would only dampen the movement, as it should. So my big question is what's a good, simple way to either sense rotational position or convert the rate gyro signal to a crude heading hold? Any help is greatly appreciated. I've combed the web for a simple solution. Last edited by superevan; May 03, 2011 at 07:32 AM.
 May 03, 2011, 03:25 AM Up up and away! Sweden Joined Oct 2010 705 Posts Hi. The link is broken.
May 03, 2011, 03:58 AM
Registered User
Joined Aug 2002
30,732 Posts
Quote:
 Originally Posted by superevan Any help is greatly appreciated. I've combed the web for a simple solution.
There isn't a simple solution, or more specifically, the solution is not as simple
as you think it is. Early gyro based HTs worked pretty much the way you describe,
with a HH gyro on pan and tilt but they also need a mechanism to re-center on the
fly, because they always experience some drift. Some of it due to
temperature/voltage changes, and some of it due to a geometry problem inherent in
then look down, and finally return to center in a straight line (diagonal motion) the gyros
in the pan and tilt axis don't see the same amount of rotation on the return trip,
because the diagonal motion occurred in all 3 axes while the original right angle
motion was in only 2 axes. Any pure gyro based HT will always end up off
center. Always. Even if you use only one gyro on pan, you'll still have
the same problem, because you won't hold your head perfectly level all the time.

Some HTs handle this by building in a slow "drift to center" algo that causes it to
always slowly drift the output signal back to center, to correct for
the geometry drift, but it assumes that you'll spend most of your time looking
forward. This causes problems if you look to the side for a long time,
while it's slowly drifting back to center intentionally. When you do re-center your
head, the output from the HT is now *way* off center.

Other HTs handle this by using an IMU which has both gyros and accelerometers
and a magnetic compass which can be used on yaw axis, or can be used to
at least re-center the yaw axis. Summing the 3 axes of rotation accurately
is a very tricky mathematical problem though.

Others still use pure magnetic sensors for both pan and tilt, but you
must be facing due north or south to use both axes, and they're often
overly sensitive, non-linear or low resolution.

I use a simple mechanical HT that uses only a scavenged set of joystick gymbals
which has none of these problems.

ian
 May 03, 2011, 07:39 AM Registered User United States, TX, Austin Joined Aug 2007 375 Posts Ian, thanks for the info! Yes I've read a little about people using magnetometers or combination IMUs for head tracking. This seems like it would provide a good option for head tracking although programming and code are a bit out of my range of capability at the moment. Learning it would be fun but would take lots of time for something as simple as a pan/tilt head tracker (Example). Last night I found a really simple solution to the yaw problem, (here). It basically places a pot on the back of your head, and the swing of the cable is what reads the yaw position of the head. I think I might give this a try as a quick fix for now. superevan
May 03, 2011, 08:37 AM
Just trying to get a nut.
United States, VA, McLean
Joined Oct 2006
6,608 Posts
Quote:
 Originally Posted by superevan Ian, thanks for the info! Yes I've read a little about people using magnetometers or combination IMUs for head tracking. This seems like it would provide a good option for head tracking although programming and code are a bit out of my range of capability at the moment. Learning it would be fun but would take lots of time for something as simple as a pan/tilt head tracker (Example). Last night I found a really simple solution to the yaw problem, (here). It basically places a pot on the back of your head, and the swing of the cable is what reads the yaw position of the head. I think I might give this a try as a quick fix for now. superevan

Here's how I've been flying...
 Scrtsqrl's Poor Man's Head Tracker (3 min 40 sec)
May 03, 2011, 10:33 PM
Registered User
United States, TX, Austin
Joined Aug 2007
375 Posts
Quote:
 Originally Posted by scrtsqrl Here's how I've been flying... http://www.vimeo.com/925056
Pretty slick, have you thought about incorp. a tilt sensor for dual axis head tracking? Here you'll see a really simple setup for tilt using an accelerometer breakout, OpAmp, pots and some resistors.

That's why I want to find an equally simple setup to measure yaw. I've poured over the internet though.
May 03, 2011, 10:40 PM
Just trying to get a nut.
United States, VA, McLean
Joined Oct 2006
6,608 Posts
Quote:
 Originally Posted by superevan Pretty slick, have you thought about incorp. a tilt sensor for dual axis head tracking? Here you'll see a really simple setup for tilt using an accelerometer breakout, OpAmp, pots and some resistors. That's why I want to find an equally simple setup to measure yaw. I've poured over the internet though.
MelihK, of www.flytron.com, had a hybrid HT kit. It measured yaw with a magnetometer and pitch using simple electronic tilt thingy...

I asked if it was possible to replace the magnetometer with a pot. Apparently it is possible to do so.
May 03, 2011, 11:02 PM
Registered User
Joined Aug 2002
30,732 Posts
Quote:
 Originally Posted by superevan Pretty slick, have you thought about incorp. a tilt sensor for dual axis head tracking? Here you'll see a really simple setup for tilt using an accelerometer breakout, OpAmp, pots and some resistors. That's why I want to find an equally simple setup to measure yaw. I've poured over the internet though.
As I said above, I use a mechanical HT that uses a set of old joystick gimbals.
The gimbals themselves are mounted to my goggles/cap and the stick is
held straight with a simple badge reel retractor. When I move my head
the stick stays point straight ahead and the pots in the gimbals move.