Emil Fresk

Yes, in betaflight I believe it is only in the low end, but the same idea works fine on the high end. :) Usually the issue is that throttle saturates...

Hi, I am seeing the same issue @superjax , also on a NAZE32 (AfroFlight). By simply hovering, after a while the estimation starts to drift away until it has a...

Hi, I have been testing all day and it seems to be a vibration issue. I had a cable that was touching the clean frame (making the vibration dampening moot)...

Hmm, it could be, and I will make a dataset to test on which exhibits the problem. I was thinking the same about the notch filter, though we could sadly...

I tried to get it to happen again and was successful, though not at extreme as before, but it shows the effect. I have created a rosbag here https://drive.google.com/open?id=1acsgVGwNRCfrIj9HF0cErhoZFMuiIZxl with:...

Thanks! I'm quite interested as the estimator you are using is theoretically solid, so perhaps there is some other effect at play. As far as I tested they went the...

Thank you for the initial analysis! I checked the data and I was running my node to convert everything into NWU frame, `/pixyrf/imu` is in the NWU frame, while `/pixyrf/rosflight/imu`...

Perhaps something like this `W = 1 / ( 1 + 10000 * (x - 9.8)^4 )` 

Awesome that you gave it a test! I have been abit swamped at our lab for now, but it seems to indeed be helping. I designed this cutoff gain to...

Indeed! I can do a small write up on this as well when I make the one for ESC telemetry :)