Hello,

I'm working on an application using a UR5 robot where I basically need to repeatedly pull tissues from a tissue box. I need the robot to be able to detect when it successfully grasps a tissue, or better yet, when it fails to do so. The problem is that because the tissue is so thin, using the object detection feature does not work. The robot sees its gripper as fully closed regardless of whether it has actually grabbed a tissue or not.

I do have a force torque sensor installed on the arm, and was playing around with trying to capture those measurements to be able to tell the robot that it failed to pull a tissue, but I have had no success.

I'm hoping that someone here might have a simple idea to solve this problem, because my internet searches have so far been fruitless!

Thanks in advance!

Hi,

We have a customer who is doing the the same thing. They are using external machine vision to detect if tissue is ripped or not. I think they are also measuring if the ripped tissue is right lenght. Robot is equipped with normal cheap pneumatic gripper and gripper pads are some soft and high friction material.

I am thinking maybe you could use photoelectric sensor to detect the tissue. If you are using robotiq gripper, you could use accessory holes to attach the sensor onto gripper. Sensors are way more more cheaper than machine vision and all in all you are just detecting the object presense.

Just one note for you. Remember to design the machine layout the way the robot can do nice and smooth movements. This application is quite rough for robot joints if the robot rips hundreds of thousands tissues repeatedly.

Thank you both for the replies! We were considering utilizing sensors, but I had not thought of mounting the sensor directly onto the gripper. So, that is something to consider!

And thanks for the link to New Scale, Matthew. Those grippers might be just what I need!