Matrix Kalman

airspeed
posted 2013 https://groups.google.com/group/uavdevboard/browse_frm/thread/98486af69cf21ec7?hl=de Regarding the integral terms in the control loops, what I generally found is that they tended to generate low frequency,high amplitude oscillations. I never got around to tracking down why that was, I was satisfied with KP and KD instead. I had a couple of speculations about what was going on: 1 There might be "parametric amplification" through attitude errors in the direction cosine matrix which fed back into the matrix.My visual observation was that there were low frequency variations in pitch. 2 The errors that the integrators were trying to correct were not physically in the frame that the correction was being computed. The nice thing about KP and KD is that you can show that it does not matter what frame (earth or body) the errors are in, it all comes out in the wash. The problem with KI is that it very much matters what frame the error is in. Problem is, some errors are in earth frame, some are in body frame. In that case, since you cannot separate out body frame error from earth frame error, you are likely to get parametric amplification. --- airspeed control by energy management journal paper by Mahoney. Bill, I think it might be worse than that. Aircraft have a natural tendency toward airspeed-altitude oscillation, known as a phugoid. The frequency is dependent on airspeed. For a paper plane the oscillation period is in seconds. For a jumbo jet the period can be in minutes. The dynamic range of oscillation means that at some point, the airspeed integrator is reducing the control loop phase margin. Thanks to the idea (was that you bill?) for controlling with feedforward plus height energy, there is very little oscillation. We get into trouble when the pitch damping is too much so pitch response becomes a problem. This is acting a little like integrator wind up so we get some airspeed overshoot response on a step change. There is an easy fix for this. --Matt https://docs.google.com/drawings/d/19ULkFRjNJVHtQhK-K78KJUtmxZoOits3oUGSV_HAN_0/edit pitch control detail -

autonomous flight
https://groups.google.com/group/uavdevboard/browse_frm/thread/f86567b2eb090bb7?hl=de

Maya airframe
Provides PID settings for Maya airframe. Each airframe will have its own PID settings. http://diydrones.com/profiles/blogs/successful-flight-of-the-maja?xg_source=activity&id=705844%3ABlogPost%3A373776&page=2#comments

http://diydrones.com/profiles/blogs/successful-35-km-full

http://vimeo.com/24236706

http://code.google.com/p/ardupilotdev/source/browse/#svn%2Ftrunk%2FAPM_MAJA

links
Gentlenav Main link for matrixpilot or gentlenav