7th axis bias and Joint positioning errors in open loop


Ernesto

Recommended Posts

Hi all,

We have recently acquired 2 Sawyer robots and now we are starting in "playing" with them.

We completed the setup using lunar distro and everything seemed to run good.

In order to control the robot and close the loop using force or camera sensors, we wanted to know parameters such as control rate, positioning errors, joint resolution, behavior of these robots in open loop, etc.

So the first thing we did was the calibration of both robots using the example script provided in the SDK

We realized that one of the robots had a bias of the seventh axis of 0.1741 rad (is this normal?)

Afterwards we proceeded to test the joint positioning errors and repetitivity, testing  both set_joint_positions() and move_to_joint_positions() commands and using a force sensor of 6DoF, getting in both cases strange behaviors:

1) Sometimes some axes didn't updated their values whilest others did -> poritioning errors : if we integrate the joint values we get no repetitivity

2) The joint absolute error, computed as e=(cmd_sent-joint_values_cur)/(cmd_sent-joint_values_prev), runing at 100ms overcomes the 80%, starting to have a good behavior (not allways) when the period is 2s up

So, we think there is somthing we are not setting correctly or perhaps the commanding functions used are not correct.

Do you have any idea of what we are doing wrong? Do you have any test program in order to see if the robot behaves as it should?

 

Thanks in advance for your help!!

 

PS: we update the comands each Periode using the sleep function in our control loop. The rest of the setup is similar to the one provided in joint_position_keyboard.py example provided in the SDK

 

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.