[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Paparazzi-devel] VFF_FLOAT - question about the Kalman filter

From: Gautier Hattenberger
Subject: Re: [Paparazzi-devel] VFF_FLOAT - question about the Kalman filter
Date: Sat, 08 Dec 2012 23:11:42 +0100
User-agent: Mozilla/5.0 (X11; Linux i686; rv:17.0) Gecko/17.0 Thunderbird/17.0


The vertical filter used for rotorcraft is based on a kinematic model not a dynamic model. It means that the input of the filter's propagate function is actually the accelerometer measurement, instead of the acceleration estimated by the a dynamic model. The reason is that the dynamic model is usually unknown, and it makes the filter independent of it. If you want to tune the noise associated with the accelerometer, you have to look at the system noise.


Le 08/12/2012 01:21, Michal Podhradsky a écrit :
Hello guys,

I am currently adjusting the VFF filter (z_dd, z_d, z estimation) to our rotorcraft airframe (i.e. changing the process and measurement noise covariance). As the filter is currently implemented, in the propagate phase (vff_propagate) the measured acceleration is propagated to all three states with zero noise (line Is there an assumption that the IMU accelerometer is always correct?

If I understand it correctly, the barometer (z measurement) and accelerometer(z_dd measurement), and optionally other sensors (sonar etc.) should be updated at different frequencies. In the current master branch ( ) are already separate updates for baro (offset) and gps/sonar/etc (no offset) sensors, but the acceleration is still propagated directly.

Do you know why the filter was implemented in such a way?

Thanks a lot!

Paparazzi-devel mailing list

reply via email to

[Prev in Thread] Current Thread [Next in Thread]