robot localization 2d model gets worse after adding imu yaw rate

edit

Hi all,

I have been struggling with improving the localization of my mecanum-driven AGV with wheel encoder and imu using robot_localization package. The background is that I have already been aware that the wheel encoder is quite accurate at least for a short period of time and I want to make it more reliable by adding the imu yaw rate data. The problem is when only fusing wheel encoder odometry gets reasonable result, but if I add the imu yaw rate into the filter, the topic /odometry/filtered/position/y starts to drift very quickly which is totally unacceptable.

I have done some experiments which makes AGV start from a specific location and reset the encoder in order to align the odom frame and base_link frame, and then move according to a pre-planed route with the aid of AMCL. In this way, I can start the filter with zero initial state, at least I think so.

In addition, I did some IMU calibration by rotating the IMU through N full rotations in y axis, manually integrate the velocities, and calculate the final error, and then back out the values from that, like Tom proposed

BTW, the imu frame has the y-axis transformed into z-axis on base_link frame, here is the transformation

At time 0.000
- Translation: [0.524, 0.012, 0.374]
- Rotation: in Quaternion [-0.500, 0.500, -0.500, 0.500]
        in RPY (radian) [-1.571, -0.000, -1.571]
        in RPY (degree) [-90.000, -0.000, -90.000]

The data from the calibrated IMU shows here: https://drive.google.com/file/d/1tIsG...

If there are any suggestions for the improvement of my task, I would be very appreciative. Thanks in advance.

Here is the result of fusing only encoder https://drive.google.com/file/d/1jXOp...

Here is the result of adding imu yaw rate https://drive.google.com/file/d/1ka3N...

Here is my ekf_template.yaml

frequency: 40
silent_tf_failure: false
silent_tf_failure: false
sensor_timeout: 0.1
two_d_mode: true
transform_time_offset: 0.0
transform_timeout: 0.0
print_diagnostics: true
debug: false
publish_tf: false
publish_acceleration: false
permit_corrected_publication: false


map_frame: map              
odom_frame: odom             
base_link_frame: base_link  
world_frame: odom #_with_noise           

odom0: /odom
odom0_config: [false,false,false,
              false,false,true,
              true,true,false,
              false,false,false,
              false,false,false]
odom0_queue_size: 10
odom0_nodelay: false
odom0_differential: false
odom0_relative: false

imu0: /d435_front/imu_cov
imu0_config: [false, false, false,
              false,false,false,
              false,false,false,
              false,true,false,
              false,false,false]      
imu0_nodelay: false
imu0_differential: false
imu0_relative: false
imu0_queue_size: 2
imu0_remove_gravitational_acceleration: true

use_control: false
stamped_control: false
control_timeout: 0.2
control_config: [true, false, false, false, false, true]
acceleration_limits: [1.3, 0.0, 0.0, 0.0, 0.0, 3.4]
deceleration_limits: [1.3, 0.0, 0.0, 0.0, 0.0, 4.5]
acceleration_gains: [0.8, 0.0, 0.0, 0.0, 0.0, 0.9]
deceleration_gains: [1.0, 0.0, 0.0, 0.0, 0.0, 1.0]


initial_state: [0.0, 0.0, 0.0,
                0.0, 0.0, 0.0,
                0.0, 0.0, 0.0,
                0.0, 0 ...