Results of first IMU flight tests

Today i made my first tests with a Gyro/accelerometer stabilized airframe. The system is based on an Ardupilot board and the Ardupilot expansion board with an additional gyro. The software is my own, but the Kalman filter is based on code published at the Sparkfun forum.The airframe is this 150 cm span flying wing:

In stabilization mode the ardupilot reads elevator and aileron input from the transmitter and uses this input as target roll and pitch angles. If the stick is centered the plane is supposed to fly straight ahead.The first flight tests were not a complete success. The Kalman filter seems to do a good job on the ground, but when flying i seem to get unreliable readings when turning. If i enable the stabilization et level flight it continues to fly level, but when i enable it during a turn it continues to turn rather than leveling out.My guess is that this is because the accelerators have to much weight in the kalman filter so when i make my perfect coordinated turns :) the filter is unable to keep its angles.I will keep you updated in this blog, but if anyone can give me some advice on how to tune the Kalman filter it would be appreciated./Magnus
E-mail me when people leave their comments –

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

Join diydrones

Comments

  • Mook,
    The platform is custom made. Its a foam wing with paper covering and a balsa fuselage. Span is 160 cm with a root coord of 40 cm and a tip coord of 17 cm. It has a 10% Zagi profile at the root ant a 12% Zagi profile at the tip.

    With its large winglets it is extremely easy to handle. Very gentle stalls. The only difficult thing is throwing it at launch. After a few failed attempts i now use a short bungee cord to get it in the air.

    Weight is 1.5 kg.

    /Magnus
  • What platform is that, is that a custom made one?
  • All right... I think i will give the ArduIMU code a shot instead. I will have to convert it to avr-gcc but that shouldn't ba a problem (I hope).

    Maybe a new test flight tomorrow...
  • Developer
    It is for sure the problem!
  • Yes Ryan you are right. I am working with a separate filter for each axis. I know this is a simplification, but i think it should work. Do you think this is the cause for my problems?

    /Magnus
  • Developer
    Your Kalman is probably not complete. I would guess that you have two single axis kalman's which is not the same as a single two axis kalman
  • This is a zip file containing my AVR-Studio project (written in avr-gcc). Please keep in mind that this is VERY early code and only intended for my own tests. It is not very well documented.

    IMU stabilization.zip

    /Magnus
  • Good try ! Can you post the code here for reference
This reply was deleted.