Discussion

Left ArrowBack to discussions page
Samuel_BouchardSamuel_Bouchard Posts: 135 Handy
Hi @Ryan_Weaver , how did you do this master-slave with 2 URs? Was it for a specific application?


CEO & Co-Founder @ Robotiq

samuel@robotiq.com

Best Answer

Comments

  • Ryan_WeaverRyan_Weaver Posts: 41Founding Pro, Partner, Beta Tester VIsion 1.1 Program Handy
    Hi @Samuel_Bouchard - I'm actually working on a white paper on the topic, so stay tuned!  The inspriation came from a customer who wanted to get two robots working together, so I figured this was the epitomy of that.  I used the Modbus Server on the UR5 to read the standard Modbus register outputs from the UR10 for joint positions.  There are some tricks to it, but I'll detail it all in the white paper and post the link here when it's complete. 

    Ryan Weaver   |   Automation Engineer   |   Axis New England
    rweaver@axisne.com
    https://www.youtube.com/user/AxisNewEngland
    https://twitter.com/axis_newengland

  • EnricEnric Posts: 74Founding Pro, Partner, Beta Tester VIsion 1.1 Program, Beta Tester Multi Gripper, Beta Tester Camera URCap 1.2.0-beta Handy
    Well @Ryan_Weaver
    I hope your white paper to try it!

     ;) 
    Enric Vila Avila
    Vicosystems S.L.L.
    Industrial Collaborative Systems

    evila@vicosystems.com

    twitter  linkedIn  YouTube  Google
  • EnricEnric Posts: 74Founding Pro, Partner, Beta Tester VIsion 1.1 Program, Beta Tester Multi Gripper, Beta Tester Camera URCap 1.2.0-beta Handy
    Ohh yes!!
    Enric Vila Avila
    Vicosystems S.L.L.
    Industrial Collaborative Systems

    evila@vicosystems.com

    twitter  linkedIn  YouTube  Google
  • matthewd92matthewd92 Posts: 460Founding Pro, Tactile Sensor Beta Testers Handy
    edited March 13
    @Ryan_Weaver That is really cool.  Great example of using the real time feedback of the UR for control.  Thanks for the white paper as well, really concise.  We did a proof of concept last year for a customer for a dual armed UR3 that had a small portion of the program that had coordinated motion.  The rest of the time the robots were doing different tasks.  When they needed to coordinate we used the modbus register to let them communicate with each other but this would have been so much better!  Basically we let the master tell the slave it was time to move, they would both move and then the master would wait for the slave to say that motion was complete and then it started again until the motions were totally complete.

    Did you see any issues with differences in joint kinematics between the two causing the actual positions to not be the exact same?
  • Ryan_WeaverRyan_Weaver Posts: 41Founding Pro, Partner, Beta Tester VIsion 1.1 Program Handy
    @matthewd92  Since we were moving based on joint angles and not TCP location, the "poses" of the robots should have been identical, but the TCP locations would have been some scaled ratio of each other's.  It would have been interesting to give this a try using the TCP location as the driver instead, having the UR10 follow the UR5.  

    Thanks for the feedback!

    Ryan Weaver   |   Automation Engineer   |   Axis New England
    rweaver@axisne.com
    https://www.youtube.com/user/AxisNewEngland
    https://twitter.com/axis_newengland

  • David_LevasseurDavid_Levasseur Posts: 25 Handy
    Very clever @Ryan_Weaver, thanks for sharing.
Sign In or Register to comment.
Left ArrowBack to discussions page