3D Robotics
I spent a few hours yesterday trying to perfect the gyro-stabilized camera (shown) in preparation for a test flight today. But even when I tweaked the settings it wouldn't take long before the gyro lost track of where "down" was and it ended up with the camera assembly at one side or another when it should have been level. It turns out that the drift cancellation wasn't perfect, which isn't too surprising. Unfortunately I really did need to it be perfect to avoid the little errors adding up over time and rendering the whole thing useless.

And then it struck me. I'm an idiot. The PLANE knows where down is! In many of our UAVs we're using IR stabilization to keep the wings level, and the way that works is that a FMA "Co-Pilot" sensor measures the infrared gradient between sky and earth on both sides and front and back, and uses that to establish a vertical axis. Then it just moves the ailerons and elevator to keep the plane flying perpendicular to that axis.

All I needed to do was to let that same FMA Co-Pilot drive the camera stabilization, too. Once I'd slapped my head and realized that the solution was right in front of me, it was a simple matter of removing the gyro, attaching the camera tilt servo to the aileron output of the Co-Pilot via a Y-harness (it's still driving the ailerons with same channel) and putting on a longer arm on the tilt servo to compensate for the lower throw distance of the Co-Pilot's signals. (All the other components and build instructions are as described here)

Today we tested it, and it work brilliantly. It's SO much better than the gyro-driven model. Here's a video of it in action:

The advantages include:
  • Doesn't need special calibration and doesn't drift. "Down" is alway down.
  • Much cheaper. Without the gyro, the cost drops from $100 to $25 (two servos and some aluminum)
  • Doesn't take up a separate channel. The camera stabilization automatically comes on when I turn on the plane stabilization.
  • Saves power because the tilt servo isn't always jittering with every gyro twitch.

But what about our UAVs that use gyro-based autopilots, rather than IR, for stabilization? There's no good way to have those autopilots drive the camera assembly, too. The answer is to bolt on a cheap ($49) and simple Futaba "pilot assist" sensor and controller, which uses visible light to do what our FMA units do with IR. You can just put it on the camera mount where the gyro was and it will keep the camera pointed down. It's not quite as neat as the ones that use the same stabilization system as the entire plane, but it's equally effective.

E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones

Comments

  • Has anyone tried the Flymentor 3D? I have one NIB and wonder if that could be used to stabilize a camera mount.

  • Hey mate,

    Can you tell me a little about the Co-pilot system? I'm looking to stabilise a camera on a motorcycle so gyro drift will be a definite problem... What would be the way to go given that I'm only looking to stabilise in roll?

    Do I just give the co-pilot a neutral signal and hook the output to a servo?

  • Moderator
    For those who want a head start using their APM/IMU, see http://diydrones.com/profiles/blogs/announcing-arducopter-the?id=70... . Max presented a couple of lines for stabilization that I have used to reproduce the effect successfully on my quad.
  • I also want to build one. can you tell me more how should I start? thx.
  • I think in some cases you may be able to simplify it even further by losing the servo for the shutter release. You are using a Canon camera, which is most likely to be compatible with CHDK, an open source firmware. One of the features of CHDK is to allow for using a USB cable as a remote shutter release. This can trigger a script in the camera to take one photo, or start a timed sequence etc.
    As another thought the KAP community have been playing with stabilised camera rigs for a while , there may be some learning there.
    Jim
  • Noth666: thaks for that. I just read it again alongwith your message - I now get it. I actually didn't read this properly in the main text: 'and putting on a longer arm on the tilt servo to compensate for the lower throw distance of the Co-Pilot's signals.' Good stuff - will do some tests and see how it compares with a Co-Pilot mounted on the camera itself - unless some has already done it?

    Thanks,
    Mike
  • Mike: When the plane banks, the copilot will compensate this with the aileron servo, and at the same time using a Y cable it will point the camera to compensate for the bank angle of the fuselage.
    The same applies for pitch, since the co-pilot strives to compensate and keep the plane level it acts on the gyros always when the plane is not fully level, and therefore also at the same time correcting the camera angle until the plane is level again - when the servos will return to the level flight and also point the camera straight down or ahead depending on the set up.
    And no it will not drift ;-) that is the whole reason for using the co-pilot instead of gyros.
  • Chris, if the CoPilot is attached to the fuselage how does it know the tilt of the camera which is moving independently of the sensor?

    Mike
  • Hi Chris, I've been inspired by your effort using the co-pilot system...I've order the kit just today. How have you been coming along? Is drift not an issue at all? Thanks!

    AJ
  • does anyone here have a simple(cheap) solutions on how to film video from a moving car?(even maybe a mount or a clamp that can fit on an elbowed arm-fit that can move in an out of a carwindow?)
This reply was deleted.