Good inputs, thanks.
Yes, the Kalman filter is not magic, just math, but definitely a brilliant insight and breakthrough. Here is a great info source on Kalman filters, a bit of history and background and then some of the math:
https://en.wikipedia.org/wiki/Kalman_filter
Note that (according to this page) one of the earliest practical applications of the Kalman filter was the Apollo navigation computer. Very cool.
A "very old technique"? Hey, I resemble that remark! Kalman first published on this in 1960, and I was born in 1955, you trying to make me feel "very old"? Its working.... Hahahaha!
And speaking of "very old", YES my 67 year old eyes could use a bigger arrow than on the Air Avionics display! Or any other wind arrow I've seen in a glider cockpit so far! Hahahaha!
Seriously, what I really want is an understandable-at-a-glance dedicated display (large format preferred, remember 67 year old eyes....) showing average and instantaneous winds simultaneously. The Anemoi display seems to fit the bill for me - but I wonder if the Anemoi winds calculation is any good, hence this post. Right now, to get this instantaneous vs average comparison, I need to bounce between the wind arrow on my Air Avionics vario and the wind arrow on my OpenVario (XCSoar). A pain.
Regarding magnetic compass as a dimension in the Kalman filter, I fully concur that mag compass input to the Kalman filter improves the calculation, as long as its errors are kept low - yes you must have a proper installation. My Air Avionics sensor box installation (under the glareshield in my ASH-26E) is pretty much at the minimum recommended distances from possible corrupting magnetic field sources, and it seems to perform quite well in both wind and vario performance. In over 300 hours of flying with it, I have had a few (less than half a dozen) strange instances of completely wrong readings, but they cleared up quickly (usually within seconds). I have no idea what caused these glitches.
So there are reports that Anemoi (and HAWK?) have problems with unreliable data in straight flight mode? The Anemoi manual does acknowledge that particular characteristic, and they recommend executing three to four >10 deg heading changes after a long straight smooth glide before winds can be considered reliable. I would like to hear from an experienced Anemoi user if this technique truly does work, and if it is or is not onerous in the applicable situations.... And also how well Anemoi winds work in general. Any problems, obviously and persistently wrong wind estimates, or flaky indications? Chronic or occasional? Possible causes? This is the sort of information I am seeking.
I just saw your new post noting "Seriously think that someone looking for better wind measuring instrument would make a mistake by taking their working Air system off." I couldn't agree more! I love my Air Avionics vario (winds and vario), and have no intention of removing it. In fact, I have acquired a spare, in case the one in my '26 breaks. I just want an understandable-at-a-glance dedicated display of both instantaneous and average winds. Jon has suggested iGlide (which I believe links into the Air Avionics instantaneous wind calculation via Bluetooth?), I will be looking into that.... Or maybe I should just saddle up and make my own dedicated display to show Air Avionics instantaneous winds and OpenVario (XCSoar) average winds together.
Hoping to hear from some Anemoi users soon. If you would rather go private, contact me at julietsix at gmail dot com
Cheers,
Jim J6