hey I'm not sure why that's not being updated but another way of tackling the problem is to reduce the max_velocity of those joints in ar_moveit_config/config/joint_limits.yaml

see: https://github.com/marcoesposito1988/easy_handeye2?tab=readme-ov-file#tips-for-accuracy for calibration accuracy tips

cool things look mostly good now. One thing I did notice is that you move the marker to different spots on the screen, but actually it's recommended by easy_handeye2 that the marker should stay at the same spot, and the only change you make is rotate it in all different axis. This is something to try

Also I see that the aruco marker is attached to the gripper, but it should actually be attached to the gripper mount. This will significantly affect the results. I printed a spare J6 gripper mount and glued my marker to it just for calibration purposes.

Thanks for the info. So it looks like that camera_color_frame is in the wrong spot, which means the calibration program (easy_handeye2) failed to compute the camera transform. One thing to check is the physical side length of your aruco marker. The default that I have in the system is 0.07m. You can also try to visualize camera_color_frame and /marker_poses (probably using a PoseArray visualization) while running the calibration, to verify that markers are in the right location relative to the camera. I will replicate your setup on the weekend if I have time.

Cool, I guess there is some bugs within the marker following program, I haven't tested other camera placements, and I mostly tested with the marker facing up. Using a custom J6 mount with the aruco glued to it is totally fine, you'd just have to create a new link in the URDF of the robot that has the translation between any robot link and the center of the marker, and then set robot_effector_frame to that link in calibrate.launch.py

This means that the robot was commanded to a position it cannot reach. You can use the "Pose" and "Axes" visualizations to debug. Create a Axes visualization and set the Reference Frame to camera_color_frame to visualize where the camera is. Create a Pose visualization and set the Reference Frame to /flipped_cal_marker_pose to visualize the pose of the aruco marker, which is where you're trying to follow. Btw, which version is your AR4?

Update: AR4 servo gripper control has been added!

Thank you very much! Great to hear this result! I've updated the repo to say that MK2 is supported.

@Jacob it would be great if you can test it!

I directly wired the encoder cables to the teensy board, bypassing the mess of wires and connectors in between for the standard wiring. Once I verified that that wasn't working, I can only conclude that the encoder is broken. If you email Stepper Motor Online, they can mail you just the encoder without the stepper motor.

Yes it should work on the raspberry pi, you'll have to install Ubuntu for Raspberry Pi. I'm not supporting the nano/gripper yet but will be working on that in the medium future

the encoder was broken and after getting a replacement, things are working again. I think I applied too much pressure to it while assembling the robot

I got both parts from the kit. I was able to remove the bearing from the post by adding WD-40 between the two parts and then rotating while pulling the bearing. I then filed down the bearing post a tiny bit using a machine file. Now the two parts are rotating freely.