twinturbostang

Yes, you're right. I was thinking dual rates when I typed that (not end points). Adjusting the rates in the transmitter should have the same effect as a gain adjustment with a potentiometer.

mklarich

Yea, the end points will need adjusted with pots or push buttons and then its up to the user to turn their head approximately the same amount when they calibrate the unit. Sound good or do I need to come up with something else?

Servo reversing is a must as well.

Matt

Skycam41

So if I get the picture right, we will be able to have an on board mounted camera in our planes move in relation to our head movement...up/down, left/right? This is tied into our existing radio TX I take it.....wow, this could be very very useful for AP work, in fact we would have the ability for pin point accuracy for lining up a shot.

twinturbostang

That's the idea. I've built a prototype pan/tilt camera mount that's large enough to put a wireless video camera, and a small digital still camera on (mounting face is 4" x 4" ). So the point is, whatever I can "see", so does the digital camera. I'm just now getting into aerial photography (took my first aerial pictures yesterday!), but I can see there's a lot of trial and error. And a bit of luck in framing a nice shot of a specific target. With this technology, you can be assured of getting a great shot the first time around.

Skycam41

Thanks Turbo.

So Matt, do you need someone to test? when you are done?

mklarich

You know it Parts should be in early next for a benchtop proto-type.

TwinTurbo, you could always use one of my mini 2 camera video switches that I am finishing right now for that pan/tilt unit. I have 100 more circuit board on the way, eta early next week. Board is 1" x 1". Funny story, I actually lost the first 25 boards that I ordered, they are somewhere in my house, car, or at my day job. There were 25 of them, in a small bundle which made a 1" square cube and I have NO clue where they ended up. Don't take that as a sign of my organizational skills though, but as a sign of how small they are

Should have some done next week as well. They are setup with power/servo signal coming in on soldered leads upper left corner, output is upper right corner on deans/soldered connections. Inputs are bottom, deans or soldered as well. Spot for a button on bottom left if you want the version where you can change the switch point. Using the same board I will make a version without a button that just switches as 1.5ms for a couple bucks less. The button isn't too expensive but there is a chunk of software development that will go into it. The difference in price will be like $2 by the time I'm done.



Matt Klarich

twinturbostang

Is that the board for the headtracking unit? I'm a bit lost. I didn't think there were any servo connections on the pilot side. Or are you showing something different now?

BTW, I'll be a Beta tester too!!

mklarich

Come on, keep up with things That picture is a 2 camera switch, 1" x 1". I'm getting development stuff for headtracker monday, figure out how its going to work and get some boards made.

Matt

kd7ost

Nice Video switch Matt. I keep waiting for the good stuff but you guys are goofing around with head trackers.
Are you planning on writing a little hysteresis in that switch cutover point?

Dan

mklarich

Define good stuff


Matt

twinturbostang

Uhh, hello. Thread title is Commercial Headtracker!

kd7ost

Actually this is good stuff. This would be perfect for telepresence.
I would like to put it on a system like this.

Dan

mklarich

Well yea, thats what the thread is about, but you were talking about using 2 cameras on the same platform, it was the perfect place to plug my video switch as well Parts are coming in via UPS monday, I'll see what I can throw together. Still to make some kind of pan/tilt just for testing, nothing airworthy.

Matt

lvspark

How about an option for when things go bad, you bend over and put your head between your legs but instead of kissing your blank goodbye, it activates a 3rd output and pops a chute!

twinturbostang

Sort of... One video camera for navigation. And one digital still camera for taking pictures. The video camera doubles as a targeting camera for the digital still camera since they will be both mounted on the same pan/tilt mechanism. Unless the digital still camera has some sort of video output, I don't know how the video switcher would be usefull here.

mklarich

Some digital cameras have an NTSC video output, some don't. Depends on the camera your using.

Matt

whogdriver

Twinturbostang,
Can you post a picture or two of your pan and tilt mechanism? I'm assuming this is going on the canopy mount for the Twinstar II that you should pictures of on one of my posts.
(Sorry this is off-thread guys)

twinturbostang

Yes I can. Although I had to steal one of the servos I had mounted on it for my TwinStar.

Actually, the plan was to use the wing from the TwinStar but build a custom fuselage that will allow me to mount this directly to the nose. I'm not sure if I can mount this to the TwinStar fuse. I may have another servo laying around somewhere, so give me a day or two to see if I can put it back together, and I'll snap a couple of pics. If mklarich objects to it being here, I can start a new thread.

Rj-TailSpin

Hey Mark, I was wondering how the head tracker and the physical mechanism (XYZ pan/tilt) stay in relative calibration? Lets say the camera is pointing straight forward and the person's head is also. The flight begins and for whatever reason the pan/tilt doesn't follow the head movement for a few degrees and then it locks back on and it starts following moving again. Would the camera position and the head position be out of relative calibration? I would think over time, the error would multiple. How does this work? Would there be a "Home" position or zero command to re-sync or is that not necessary?

I still believe this would be a great project, but I fear that most people would lack the coordination to use it. Besides, it would look really funny as an RC pilot is flying while moving their head around like Steevie Wonder I would probably fall down myself.

I have a pan/tilt from RobotZone all ready to go. If you need a tester... let me know.

Rob Lehman

mklarich

I definitly planned on there being a giant re-centering button somewhere, just in case anything gets outta whack or the pilot wants to lock the camera back at its home position. Using accelerometers I shouldn't have a huge issue with that though since its all relative to earth.

Matt

twinturbostang

Rob: Definitely an issue in my oppinion. I can't speak for mklarich's equipment. But from the research I've done so far, this is a concern. Because it is detecting relative motions there could possibly be linearity issues, noise, drift, hysteresis, etc. etc. All of this can add up to angular errors and differences from what the pilot is supposed to be "looking at". There are a couple of ways to get around this. But as mk has mentioned, the easiest is a reset button. Filtering of the signals could also be used to smooth out noise and drift.

muc

Matt,

How are you going to get pan (side-to-side) control from an accelerometer? I worked on a project about a year ago where I used a 2-axis accelerometer to control a couple of servos. I can see how tilt would be easy. Pan would require an “ear to shoulder” kind of motion if the accelerometer were mounted flat on top of your head. At least this is true for the way I had things setup. I’m looking forward to seeing the final product.

lvspark

What about a magnetic sensor??
http://www.ssec.honeywell.com/magnet...s.html#MODULES

mklarich

First thing that jumps out at me is the cost comparison I just did, the magnetic sensor has a lot more discrete components associated with it. It had a lot higher refresh frequency then I expected. I will definitly take a look at it. Their eval boards are a little pricey from digikey, need to find a rep close by who will give me one of their demo boards for free

Matt

twinturbostang

How do those magnetometers work? I'd seen those before but didn't consider them in my research because I thought that they relied on an external magnetic source. If that's the case, then they would not work for this application. However, if they actually sense the earth's magnetic field, they could possibly work.