I am trying to use robot_pose ekf to fuse wheel odometry with imu data but I am having trouble getting it to work well. As seen here. I thought I had the tf tree set up correctly but now I am starting to wonder if I have it right.

I am generating the wheel odometry from encoders and publishing a odom message (similar to this).

I would have thought this would be very common place so I was wondering if anyone has a rqt_graph example of how this should look. Either that or a clear description.

So just to clarify I have these frames: map, odom(wheel), baselink, imulink. I know that odomcombined -> basefootprint is provided as well as a robotposeekf/odomcombined topic. How does robotpose_ekf fit into all this? I have tried various combinations with limited success so I was hoping to have a definitive answer to this so that I can either fix it or eliminate it as an issue.