Ask Your Question

Revision history [back]

Can you clarify what you mean by "incorrect" yaw? What exactly is incorrect about it? I'm assuming you realise that when you fuse only yaw velocity data, the estimate will start to drift over time. Moving while stationary is a different issue, though.

I agree with the sentiment that you should perhaps turn off linear acceleration, though it's less of a problem when you have a velocity reference, as you do.

How is your IMU mounted? You posted sample IMU data (many thanks), but you didn't show me what the _base_link_ -> _new_imu_ transform is. I'm assuming it's mounted in a neutral position, given the linear acceleration values that I see, but I want to be sure.

Finally, at the point that you notice false motion in the robot when it is stationary, have you tried actually plotting the raw IMU values for Z angular velocity (or whatever axis makes sense, given the IMU's mounting)? Can you post values from all the sensors, as well as the EKF output, _while you see the robot showing false rotational motion_? Thanks!

Can you clarify what you mean by "incorrect" yaw? What exactly is incorrect about it? I'm assuming you realise that when you fuse only yaw velocity data, the estimate will start to drift over time. Moving while stationary is a different issue, though.

I agree with the sentiment that you should perhaps turn off linear acceleration, though it's less of a problem when you have a velocity reference, as you do.

How is your IMU mounted? You posted sample IMU data (many thanks), but you didn't show me what the _base_link_ base_link -> _new_imu_ new_imu transform is. I'm assuming it's mounted in a neutral position, given the linear acceleration values that I see, but I want to be sure.

Finally, at the point that you notice false motion in the robot when it is stationary, have you tried actually plotting the raw IMU values for Z angular velocity (or whatever axis makes sense, given the IMU's mounting)? Can you post values from all the sensors, as well as the EKF output, _while you see the robot showing false rotational motion_? Thanks!

Can you clarify what you mean by "incorrect" yaw? What exactly is incorrect about it? I'm assuming you realise that when you fuse only yaw velocity data, the estimate will start to drift over time. Moving while stationary is a different issue, though.

I agree with the sentiment that you should perhaps turn off linear acceleration, though it's less of a problem when you have a velocity reference, as you do.

How is your IMU mounted? You posted sample IMU data (many thanks), but you didn't show me what the base_link -> new_imu transform is. I'm assuming it's mounted in a neutral position, given the linear acceleration values that I see, but I want to be sure.

Finally, at the point that you notice false motion in the robot when it is stationary, have you tried actually plotting the raw IMU values for Z angular velocity (or whatever axis makes sense, given the IMU's mounting)? Can you post values from all the sensors, as well as the EKF output, _while while you see the robot showing false rotational motion_? motion? Thanks!