calibratin error range about 2-3 cm

[IanYangChina]

Hi, thanks for the codes! I have carried out the hand-on-base calibration with a kuka and a zivid camera. I tried 3 times and got 3 slightly different results, with an error range about 2-3 cm. I have checked other issues in this repo, and made sure I got the right marker id and size, took samples when the marker frame is steady in the Rviz simulation (it sometimes flickers). I used the default algorithms for both aruco marker tracking and calibration.

Here's the results: parameters: eye_on_hand: false move_group: manipulator move_group_namespace: /iiwa namespace: /my_eye_on_base/ robot_base_frame: iiwa_link_0 robot_effector_frame: iiwa_link_ee tracking_base_frame: zivid_optical_frame tracking_marker_frame: camera_marker transformation: qw: 0.04948431839228293 qx: 0.006091250404755587 qy: -0.9948842771176843 qz: 0.08786053747718335 x: -0.5081866403237757 y: 0.1642988819302053 z: 1.0733152560547734

parameters: eye_on_hand: false move_group: manipulator move_group_namespace: /iiwa namespace: /my_eye_on_base/ robot_base_frame: iiwa_link_0 robot_effector_frame: iiwa_link_ee tracking_base_frame: zivid_optical_frame tracking_marker_frame: camera_marker transformation: qw: 0.02473297586797349 qx: 0.0009234893264673997 qy: -0.995647569407959 qz: 0.0898517924373391 x: -0.5351059892468916 y: 0.15972440464955695 z: 1.0835655779153843

parameters: eye_on_hand: false move_group: manipulator move_group_namespace: /iiwa namespace: /my_eye_on_base/ robot_base_frame: iiwa_link_0 robot_effector_frame: iiwa_link_ee tracking_base_frame: zivid_optical_frame tracking_marker_frame: camera_marker transformation: qw: 0.06580536882194765 qx: 0.009977368156501928 qy: -0.9917885237153115 qz: 0.1092045318907628 x: -0.42471393222760484 y: 0.1898217777090142 z: 1.0544555687445611

So, just wanna be sure such an error is normal or not? Thanks :)

I noticed there are three arguments in the example file iiwa_kinect_xtion_calibration.launch (below). Are they necessary for calibration? `

<arg name="robot_velocity_scaling" value="0.5" />
<arg name="robot_acceleration_scaling" value="0.2" />

`

[IanYangChina]

I figured out that the error came from the aruco marker detection process.

[GuoqiangRen]

I figured out that the error came from the aruco marker detection process.

Hi, recently I met the same problem, and not sure what you mean by aruco marker detection process, can you illustrate more about this?

[IanYangChina]

I figured out that the error came from the aruco marker detection process.

Hi, recently I met the same problem, and not sure what you mean by aruco marker detection process, can you illustrate more about this?

So, the eye-to-hand calibration needs a marker attached on the end-effector. The camera will use the estimated pose of the marker relative to the camera frame. That detection process may have some error. In my case, I use an aruco marker, you can use other markers, such as a checkerboard.

Beside that, another error source is the end-effector link poses read out from the robot, If the robot's own state computation process is not accurate. This may happen due to wear-and-tear of the robot arm.

[lin-fangzhou]

I calibrated the robotic arm, but each time there were huge differences in the results. One noteworthy value is that my camera was always fixed on my hand, and I calibrated it 6 times as follows: result 1 translation: x: 0.17973868323082823 y: 0.006782370501410505 z: 0.07438441881749708 rotation: x: -0.45153815132695097 y: 0.5453857381574472 z: -0.43437538148312277 w: 0.5567636145368047

result 2 translation: x: -0.06653427597392521 y: 0.025643998598674816 z: 0.06696783370561568 rotation: x: -0.5327081554416593 y: 0.5008184543910814 z: -0.46062857442271155 w: 0.5032138842406616

result 3 translation: x: 0.0769910508819065 y: 0.12000227049890966 z: 0.02890224129850164 rotation: x: -0.506780500127259 y: 0.4405819116710794 z: -0.6227926098119726 w: 0.4014853284525028

result 4 translation: x: -0.010360640808195888 y: 0.0058058783075903 z: 0.07165832040066378 rotation: x: -0.4284056687898974 y: 0.5838619571409756 z: -0.382383827061095 w: 0.5738958152437608

result 5 translation: x: 0.05797963988648782 y: 0.06623176322982123 z: 0.021080226239054548 rotation: x: -0.4491078296992648 y: 0.5201030730048655 z: -0.5053875608486993 w: 0.5218988063725695

result 6 translation: x: 0.01822461983207547 y: 0.023970008797324964 z: 0.05262544780092089 rotation: x: -0.5289914513286127 y: 0.5074252851659499 z: -0.4647529069925332 w: 0.4966813463756137