Ardupilot 2.7 Beta

I was going to call it 2.6.3, but so many changes have been made, updates added, I think it requires a full point. Also, the header file has been streamlined a bit and some defines have changed. The PID gains have not, so if your plane flew well with 2.6 it will fly well, or even better with 2.7. BTW, this is the first release that has only been tested via HIL. I've never flown it for real because of the strong winds in the bay area right now. We'll see, but I think this will be the most stable release yet.

What's changed:

I've included a folder for connecting your ardupilot hardware to X-Plane. There are PNG files showing the configurations to make in the menus.

X-Plane supplies IMU data and airspeed data to Ardupilot so you must select the GPS_IMU option to connect.

New altitude estimation - Thanks to Ryan Beall for this one. This smooths out noisy altitude readings from GPS.

Takeoff and landing are now full features. Their use has been simplified and are assigned full flight modes which can be seen with a ground station.

New Fly by wire throttle command by Christof Schimd.

Updates altitude hold for those with airspeed sensors. Airspeed is now recorded in m/s instead of air pressure value. Since calibration for your particular pitot tube setup could be needed there is now a calibration value in the header file.

Improved stick mixing response for all modes.

Waypoints will be skipped if you miss them after one loop around. The number of degrees can be changed in the code so let me know how many is good.

Loiter hold has been improved.

Cross track gain now on by default.

Numerous bugs are gone.
Old, unused code and vars have been removed. Code has been reformatted and commented.
If you have any issues, please comment here and I'll get them fixed in the final release.


E-mail me when people leave their comments –

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

Join diydrones


  • Hi Jason,

    I noticed that ground_course went from an int to an unsigned int when I moved from 2.6.2 to 2.6.3 on Friday. (I noticed when my X-Plane simulator was reporting courses >360 degrees.

    Won't this cause a problem when working with ArduIMU 1.7 which uses a signed integer to pass ground course? Or has this been dealt with elsewhere?


  • Nice work. In-silico testing seems like the way to go. For HIL testing, do you add noise/inaccuracy to the GPS and IMO data from X-plane? I'm guessing that the values from X-plane are "actual" values, but the values from the Ardupilot IMU/GPS will have noise/inaccuracy. Do you need to consider this for HIL testing?

    Apologies if this is a silly question, I'm just getting started
  • Developer
    Hi Jason,
    Thank you so much for your great work in Ardupilot´s code. Thanks to you and the entire community for improve the software.

    Currently I'm flying this plane with version 2_6_1:

    AutoLaunch mode works perfectly so I can launch my plane with the help of this launcher made by me:

    So it's time to prove Autoland :) I have noticed that in this new version (2_7) Autoland mode is better defined. My Cularis lands very well with the help of Stabilize mode and my question is ... could be that Autoland mode continues this scheme? For those who have airplanes with propellers at the front would be the best solution.

    The idea is that Autoland maneuver was defined by three points (waypoints) and that on reaching the last waypoint the Ardupilot change from Auto mode to Fly By Wire mode A or Stabilize mode (because in these modes the throttle may be 0).

    I am currently on vacations and I'm writing from my girlfriend´s Aspire One. So I had to do the drawing with the Microsoft paint:)
    In August I'll be in my house and I could test the new code. If yoy want I could put in this post my findings as a tester of this new code version.

    Best regards from Spain
  • Looks good on bench
  • Developer
    New update is posted on the SVN.
  • Looking forward to it.
  • 3D Robotics
    Got it. I'm running off the SVN and that's clearly a work in progress. I'll wait for the update.
  • Developer
    Hi Chris,
    I'm reworking some things right now. You can try and fly the zipped version.
    I'll have an update in an hour.
  • 3D Robotics
    With Airspeed Sensor = 1, I get a compile error:

    'alt_energy_error' was not declared in this scope In function 'void read_control_switch()'
  • Thanks for the SVN

This reply was deleted.