Robot pick fails

[jrgnicho]

When running the main application (see Demo.md), as a part comes close to the robot, the robot arm moves in position to pick the object, then it lowers its tool towards the conveyor just in time to touch the target object. Once the tool and object make contact the gripper appears to fail at making a successful grasp. After re-attempting a grasp for approximately 10 seconds the robot gives up and retreats.

[alex07zzz]

Yes, I also found this problem. And sometimes the object is attached to the side of the gripper, but the gripper attached state will still be false. The vacuum gripper seems to have a very strict z-direction height and orientation requirement. I noticed that we have a very complex-looking vacuum gripper. Could it be the reason that the collision size of our gripper's tip is too small? I didn't notice the same issue for ARIAC gripper which is big and flat. 在 2018年2月1日，下午3:46，Jorge Nicho <[email protected]<mailto:[email protected]>> 写道： When running the main application (see Demo.md<https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fswri-robotics%2Fgilbreth%2Fblob%2Fkinetic-devel%2FDEMO.md&data=02%7C01%7C%7C051c030921ea409673f108d569ce1348%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636531256157508403&sdata=bmEA97xHxLWTZy4VK3kSaY6iLWfq2Wg%2B7xBwYx4Hix4%3D&reserved=0>), as a part comes close to the robot, the robot arm moves in position to pick the object, then it lowers its tool towards the conveyor just in time to touch the target object. Once the tool and object make contact the gripper appears to fail at making a successful grasp. After re-attempting a grasp for approximately 10 seconds the robot gives up and retreats. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fswri-robotics%2Fgilbreth%2Fissues%2F35&data=02%7C01%7C%7C051c030921ea409673f108d569ce1348%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636531256157508403&sdata=ckhUYoQLWP8uAOAMYRZTJpoMFYB8i1v4%2F53vYcvd9vo%3D&reserved=0>, or mute the thread<https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAPybXDwjEx0XIUMS57q0sj850CAccxCdks5tQkzsgaJpZM4R2d1m&data=02%7C01%7C%7C051c030921ea409673f108d569ce1348%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636531256157508403&sdata=KwijyXUaCd9h0Reeynz3rpcdySy4VWES3bX7YN2voBc%3D&reserved=0>.

[jrgnicho]

After observing the robot more carefully, it looks like the robot tool does not make contact with the top surface of the object as it is going by, it instead arrives at the conveyor a bit too early and blocks the object from moving forward. That pick strategy needs to be modified. I'd suggest moving the gripper into position just a couple of cm above the expected pick point and then wait until the object passes right underneath it; at this point the arm will move the tool downwards towards the pick point and turn on the gripper. This will require implementing a countdown logic in the robot execution node that will execute that move at the right time. Also, the rail or the arm should be moved one at a time and never simultaneously.
The tool size may be an issue too but our grip strategy will continue to be an problem even if we modify the tool.

[alex07zzz]

I agree with you. I will change the strategy. 在 2018年2月2日，上午8:32，Jorge Nicho <[email protected]<mailto:[email protected]>> 写道： After observing the robot more carefully, it looks like the robot tool does not make contact with the top surface of the object as it is going by, it instead arrives at the conveyor a bit too early and blocks the object from moving forward. That pick strategy needs to be modified. I'd suggest moving the gripper into position just a couple of cm above the expected pick point and then wait until the object passes right underneath it; at this point the arm will move the tool downwards towards the pick point and turn on the gripper. This will require implementing a countdown logic in the robot execution node that will execute that move at the right time. Also, the rail or the arm should be moved one at a time and never simultaneously. The tool size may be an issue too but our grip strategy will continue to be an issue even if we modify the tool. — You are receiving this because you commented. Reply to this email directly, view it on GitHub<https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fswri-robotics%2Fgilbreth%2Fissues%2F35%23issuecomment-362628346&data=02%7C01%7C%7Cd0e002ebba994a059ee308d56a5a9438%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636531859616268512&sdata=VIQDIdGz9V04JNmZIlhNszbguGa2ZavlSqqkGc9EXCU%3D&reserved=0>, or mute the thread<https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAPybXBtygH3kcHQDmCoAnqX2jAZkE0bfks5tQzLPgaJpZM4R2d1m&data=02%7C01%7C%7Cd0e002ebba994a059ee308d56a5a9438%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636531859616268512&sdata=4lZ%2BzvnBSweJDr8CPTBDbKeuwl5o%2BH73CSaJlwnMMOQ%3D&reserved=0>.