SITL Simulation of single physic steps

[j0hn (Johannes)]

Hello everyone,

I´m doing SITL Simulation of a quadplane. I use a Matlab physics model and Cygwin/MAVProxy to connect and run with the firmware. The way it is described in the Ardupilot SITL Matlab example.

I want to verify my SITL environment. I have flight data of an aircraft which is controlled by a flightcontroller on whose the ardupilot firmware runs. For the verification i’d use a Matlab Script which sends the RCIN values, of the flight data, to the firmware with the needed information of the JSON format. To then compare the PWM values that are send back by the firmware with the PWM values of the flight data.

My question is if someone knows how to send these RCIN commands to the firmware automatically every time the JSON format is send? I’ve read that it’s possible to extend the JSON format, is this the way to go?

Thanks for your answers in advance!

[TunaLobster (Charlie)]

I went a slightly different route for validation of SITL. It still involves creating an error for the control loops to correct for, but it changes the angle rather than the RC input.

SITL vs physical servo output differences

I’m trying to create a process to validate that the system dynamics of the attitude controller is the same in SITL and physical hardware. For the SITL side I am using the MATLAB/JSON interface to simulate the body rotating. I noticed some unwanted randomness on the angle reported in the GCS so I dropped the RND parameters for attitude sensors to zero. That model is available here. I’m still working on it, but this version helps illustrates the issue I am having so far. For the physical side I …

[j0hn (Johannes)]

Thank you very much for your answer!