state vector in robot_localization -- accelerometer bias

asked 2020-08-18 05:11:27 -0600

Tomas Novak gravatar image

I have a question related to the state vector of robot_localization package.

I am trying to get a localization system working based on IMU and an absolute position sensor in 3D. It turns out that without any velocity measuremetns, I end up with non-zero velocity while having stabilized position (which is not the correct one in the end). This is likely caused by the bias of accelerometer, which is around 20 mg.

I think, that it should help to have "accelerometer bias" in the state vector of robot_localization and let EKF estimate it as well.

I found two questions related to this:

  1. https://answers.ros.org/question/2715... , where Tom Moore writes: > I have yet to add acceleration bias correction for the IMU data
  2. https://answers.ros.org/question/2843... , where he states that: > The filter does not have active bias correction, so unless you have a velocity reference (e.g., wheel encoder or visual odometry), the integration of acceleration data will cause your state estimate to start running away rapidly.

I am about to start implementing the bias estimation into EKF. I wanted to ask -- what is the reason for not having accelerometer bias in the state vector of robot_localization package?

edit retag flag offensive close merge delete