I did a re-implementation of the Kalman filter object that was floating around the Propeller forum, entirely in PASM. I have to admit, I'm not a mathematician, so it might not have been a good algorithim to begin with. But, I was able to fit it all into 2 cogs. I may be able to help you implement DCM in PASM, if you're interested. I'd need some guidance with the math. This is something I've been meaning to do for a while now anyway.

One question, is it absolutely necessary to use FP? I guess for getting something working quickly it would be good. But if it is possible to use fixed-point I could probably fit it into one cog including the sensor readings. Especially since that would allow the use of CORDIC for a lot of the trig.

I won't have a huge amount of time to work on this, but I'd like to help if I can, since it would be mutually beneficial. I don't see an obvious way to direct email on here, so send me an email at my full name at gmail. I've been working on an unique Propeller-based UAV board, which you might find interesting.

I realize I'm kinda late to the party, but I've posted a first draft of DCM code on the Prop, coded mostly for readability.  I'm giving some semi-serious thought to implementing it in PASM instead of Spin, but I'm going to time it first to see if it's really necessary.  At present it runs at 200hz without problems.  I've posted code for it over at the Parallax forums, along with a C# visualization app that includes an implementation on the PC side to make it easier to debug.  Thought it might be useful for you guys if you're not already light years ahead of me.  :)

 

Thanks, Mathias - The code would be great.  I haven't decided exactly what I'm going to do yet, though I'm shooting for "augmented flight", not autonomous.  The previous version was very stable, but couldn't self-level, which I think is all I'm going for at the moment.  I want to be able to park it nice & high and be confident that if I lose orientation that I can just let go of the controls and have it level itself.

The current code is really simple, but I was hoping to produce something clean enough for others to use as a starting point.

Something else in that code that might be of use to you is the reader for the ITG-3200 and ADXL345 - It's all PASM on a single cog, does temperature drift compensation for the gyro, and there's a decent amount of room to spare.  I was thinking of trying to put the entire DCM estimator in the same cog, though I'm not sure if it's going to be small enough.

 

I'll have to have a look at FlightGear - I never really considered trying the algorithm inside a sim.  So far it's been more "fire it up and hope it works" after a decent amount of testing on the ground.  :)