You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi! I have successed to let the nxtpx4 board work. But there is a little confusion about ekf2 estimator. In px4 user guider, EKF2 seems to require minimal sensors : a imu, a yaw estimator source(mag / vision),a height resource(baro). In nxtpx4, it only has imu and baro, no yaw source. But ekf2 can run well in this situation. It makes me really confusion. Would you like to tell me why it's ok? Thank you for your such brilian project!
The text was updated successfully, but these errors were encountered:
I am not very familiar with EKF2 esimator in PX4, however, from my own experience and knowledge about EKF: without magnetor should be ok for picth and roll estimation, but diable the udpate stage of EKF in yaw estimate, so we just use gyro to update instead. If gyro data is good enough, yaw angle estimation should be fine in short-time, but you will see a drift if you put the board over 10-20 min.
Hi! I have successed to let the nxtpx4 board work. But there is a little confusion about ekf2 estimator. In px4 user guider, EKF2 seems to require minimal sensors : a imu, a yaw estimator source(mag / vision),a height resource(baro). In nxtpx4, it only has imu and baro, no yaw source. But ekf2 can run well in this situation. It makes me really confusion. Would you like to tell me why it's ok? Thank you for your such brilian project!
The text was updated successfully, but these errors were encountered: