Thread Tools
May 25, 2006, 08:32 PM
Littlebirdz
toofastdave's Avatar
Thread OP
Discussion

GPS Controlled Camera Gimbal


Anyone out there working on a GPS guided camera mount? I came across a thread where a guy claims to have a Black Widow Camera on a gimbal mount and is using GPS to control servos that aim the camera toward a fixed object on the ground. I am assuming he is powering the GPS up and letting it use the initial $GPRMC string to determine what the target is but he has not responded to posts in over a month. I am very interested in something like this because I have been invited to fly at Langley Airforce Base at two change of command ceremonies for a Predator Squadron to take aerial photos during the event. I also do laser work and kit balsa models and have been developing an AP model over the last year that would be suited for this type of application. I have the basic PIC programming skills, GPS eval kit and PIC programming hardware necessary to do this sort of thing but I'm a little fuzzy on how to get the PIC to take the GPS data and tell the servo to move to the point where the initial GPS $GPRMC string was read. I can control servos from the PIC and have even done a project using an accelerometer to crudely control flight attitude but that's as far as my time allows me to go.

His setup uses GPS altitude information as well because it compensates for altitude when the plane is climbing or decending.

Anyone interested in working on something like this for small scale hobby applications? I will fund the project for parts, cameras, etc... but need some help on the programming side.

If you are interested and have the background to help complete a project like this please contact me at [email protected]

I am looking for serious and qualified interests only or someone who is already worked on or completed a similar project.

Thanks,
Dave
http:/www.littlebirdz.com
Sign up now
to remove ads between posts
May 26, 2006, 08:56 AM
York Electronics
Gary Warner's Avatar
There's no way I could offer to take on such a project, but wow - what an idea. I'm kicking around this idea in my coffee starved head (bit of a handicap to say the least ) and I'm wondering if it could be done well only using GPS data. Well, let me rephrase that - I'm wondering if "I" could do it. Still, it's good brain food.

To do this project you would have to resolve a 3D triangulation. The GPS dose this I guess by making the target a 3D waypoint and the software would resolve the track-to coordinates. I can see this working with a gazillion lines of code. But now this track-to data would have to drive the servos. One for tilt and another for pan. Pan could be done 'easy' with a 2D track-to data, though calibration and slip-skid platform errors are no small hurdle to overcome. The tilt control would need the altitude data from the GPS but would have to have feedback and correction for pitch and roll of the platform.

My head hurts.

I think it's doable but I don't see how it could work without platform attitude data, and very accurate data at that.

What ideas do you all have about this? What would be your approach?

Gary
--
Last edited by Gary Warner; May 26, 2006 at 11:46 AM. Reason: spelling
May 26, 2006, 11:07 AM
Littlebirdz
toofastdave's Avatar
Thread OP
Quote:
Originally Posted by Gary Warner
I think it's doable but I don't see how it could work without platform attitude data, and very accurate data at that.

What ideas do you all have about this? What would be you approach?

Gary
--
Gary, Check out this thread...

https://www.rcgroups.com/forums/showthread.php?t=507764

He has to be using a digital compass in addition to the GPS data and I would assume that he has some sort of Accelerometer and maybe a gyro or two... I would say it would prbably be close to a complete IMU setup but he is boasting 65 grams WITH camera and servos. The video is pretty steady and doesn't look jumpy.

Basically you would be using the GPS data to acquire the initial home position and then the whole time you were flying you would tell the servos to point the camera toward home. I don't think it would be that difficult especially since you can make a boat or robot return home if they lose RF link to a base station. Also with RCAPS and a few other "Open Source" locations you can put together enough information on a PIC 16F877 to read the GPS data and make a plane turn by using rudder control...

I just need to adapt code from a few sources I have found to point the camera toward home instead of having the servos move to make the plane to go home. I figure if I can get the azimuth information to work then it would be a matter of playing with the altitude information to get the elevation going.

I assume that the compass would have to be mounted on the camera mount (rotational) and you would take the difference between the GPS heading and the compass heading and get your vector for target on the camera. It's simple in my head but not so simple in MicroCode Studio...

You could probably use a heading lock gyro, a PIC and modified servo to do the same thing and have manual control of the camera + heading lock capability and forget the GPS completely but I wanted to use GPS...

Anyway, food for thought..

Cheers,
Dave
May 26, 2006, 11:45 AM
York Electronics
Gary Warner's Avatar
Dave,
I see. I can't get the videos to play right, but it looks good from what I can get to play. It would be nice if he'd share a bit more.

He's obviously done his homework. It's an extreme PIC project to say the least. Since I can't get the videos to play without freezing or skipping large chunks of time, I was wondering if he has 360 degree pan capabilities or is he limited to a "range" that the servo can handle. And how does it handle the >360 degree pan?

What do you think would be a reasonable price for such a device?

Gary
--
May 26, 2006, 01:58 PM
Littlebirdz
toofastdave's Avatar
Thread OP
No idea on price but if you watch the videos... or right click on them and save them to your hard drive you can see that they go about 180 degrees in either direction. Watch the one ending with "enter orbit.mpg" and you'll see the camera spin around when the plane gets to a point that the camera can't spin to. I don't think it is a full 360 degree system. The files ARE pretty large though at between 45 and 50Mb...

I am not even sure whether or not one PIC could handle the whole deal by itself, maybe ATMEL or some other real time processor but it would be nice to get a smaller SM16F877 or faster PIC to do everything. I don't really want to have to learn a new programming language...I use PIC Basic Pro as it is...

Anyway, the >360 degree pan is handled by a PID (part of the code) in the chip that tells the chip if the value of the servo signal is over a certain amount then it equals a smaller value of the servo signal on the opposite end of the servo's travel, pretty basic... it's called a PID loop for motor control and error correction but in this case it's not a true PID per se'... Just an error catching loop...

Looking at the video again I am changing my opinion on the elevation compensation. I think he just has the camera pointed downward and the camera is spinning. I didn't see any movement of the camera in up/down range... So I may only have to worry about azimuth.. Looks good though. I may be able to do that with just a digital compass, heading lock gyro and a 360 degree servo... Dunno.. But GPS would be totally kewl..

Dave
May 26, 2006, 08:16 PM
"Simplify, then add lightness"
Don't know how tough the project would be, but using a better processor would probably make it easier. You should look at the AT91SAM7 series from Atmel, or the LPC21xx parts from Philips. Both are families of 32/16 bit ARM7 processors with ram, eeprom, and I/O ports, but are cheaper than some of the high end PIC's. They are supported by the open source WinARM C/C++ compilers, and would be orders of magnitude better at number crunching than PIC's.
May 26, 2006, 08:43 PM
Littlebirdz
toofastdave's Avatar
Thread OP
Quote:
Originally Posted by jeffs555
Don't know how tough the project would be, but using a better processor would probably make it easier. You should look at the AT91SAM7 series from Atmel, or the LPC21xx parts from Philips. Both are families of 32/16 bit ARM7 processors with ram, eeprom, and I/O ports, but are cheaper than some of the high end PIC's. They are supported by the open source WinARM C/C++ compilers, and would be orders of magnitude better at number crunching than PIC's.

I realize the need for a better processor but that would involve buying new programmers, learning a new language, adaptors for SO packages... etc.. I don't want to sound like I am making excuses but I have EVERYTHING to program PICs and have spent a BUNCH of money on that stuff. I'd like to see how the PIC's do no matter what the initial cost of the processor is. In the end if I went commercial with the project I could pay someone to port the code to something else and it would be even better. Are you a software guy?

Dave
May 27, 2006, 11:09 AM
"Simplify, then add lightness"
Dave,
Im a hardware guy, but do mostly software by necessity. I haven't looked at your GPS gimble project, so don't know how complicated the number crunching will be. I just mentioned the ARM7 because you were questioning if a PIC would handle it.

Jeff
May 28, 2006, 11:42 PM
Littlebirdz
toofastdave's Avatar
Thread OP
No biggie, just wondering how hard it would be...
Jun 01, 2006, 08:11 PM
Registered User
Steve McBride's Avatar
I'm in waaaay over my head here, but couldn't you do it with GPS alone? Perahps you'd need a more accurate altitude measurement than GPS, but if you know the speed of the aircraft, its heading and its altitude, isn't it some math and trig to figure out the correct angle (and rotation) to point the gimbal and how fast to move it? I think the key is knowing these variable quite accurately though. You'd also have to know the altitude of the target as well as it's lat. and long.

I'm not saying that I could ever do it or that it's easy though, but it seems do-able with GPS to me. How repeatable would be another story (and perhaps why we haved not hear more from the original poster in the original thread linked above).
Jun 02, 2006, 06:19 AM
York Electronics
Gary Warner's Avatar
Steve, how would you correct for changes in the attitude of the platform? The GPS would have no idea about immediate pitch, roll and yaw changes, no?

Gary
--
Jun 02, 2006, 06:38 AM
Registered User
vintage1's Avatar
I suppose you use a gyro stabiised platform...with GPS that shouldn't be too hard to translate all that to controlling the camera on such a stabilised platform...assumoing you had acurate angular feedback on pan and tilt.

It is, after all, just trigonometry.
Jun 02, 2006, 07:52 AM
Registered User
Steve McBride's Avatar
Good points Gary and Vintage. Will an electronic compass detect changes in pitch or would it be best to use a heading hold gyro or some form of accellerometer (or all of the above)?

Like I said - I'm way over my head here - I'll leave it up to the real smart people to figure out It's great to have these 'what is' conversations though.

Steve
Jun 02, 2006, 08:24 AM
York Electronics
Gary Warner's Avatar
Quote:
Originally Posted by Steve McBride
Good points Gary and Vintage. Will an electronic compass detect changes in pitch or would it be best to use a heading hold gyro or some form of accelerometer (or all of the above)?

Like I said - I'm way over my head here - I'll leave it up to the real smart people to figure out It's great to have these 'what is' conversations though.

Steve
You might have to use a compass, but I'm thinking it would not be required since the GPS provides the heading information. An electronic compass usually does not output pitch information and is only usable in a limited pitch/roll range. I think most here agree that the gyro stabilized camera mount would be needed to best performance. Still, depending on the requirements a non-stabilized mount might be well enough.
Jun 02, 2006, 09:06 AM
Littlebirdz
toofastdave's Avatar
Thread OP
Electronic compass modules are very sensitive to pitch and roll and the heading error increases pretty dramatically with any changes in either attitude. You have to use an accelerometer or something like that with routines to compensate for the degree of pitch or roll. There are charts on some of the web pages (compass) that tell you what degreee of error to expect when coding for the compensation.

I don't think his setup uses tilt compensation because if you look at the distance between the top of the camera picture and the bottom of the aircraft it never changes. I just think he has the camera pointed downward at a predetermined angle and is using pan only. The aircraft he is flying is pretty stable in the horizontal (roll) axis...(or seems to be)...

Dave


Quick Reply
Message:

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion RotorKraft G7 camera gimbal for sale Greg McNair Aerial Photography 2 Nov 25, 2008 08:35 PM
Idea Gravity controlled camera mount - any ideas? Duck! Aerial Photography 12 May 04, 2006 03:07 PM
Servo controlled camera mount nealbloom Aerial Photography 26 Mar 26, 2006 10:51 AM
servo controlled camera mount nealbloom UAV - Unmanned Aerial Vehicles 5 Dec 03, 2005 11:34 AM
The ultimate RC-controlled camera-convertion pullin-gs DIY Electronics 0 May 26, 2004 10:07 PM