Dear forum members,
We are using RSI for streaming trajectory velocity commands to our robotic arm using the Cartesian control block in RSI.
When the robot is close to it's maximal reach (axis 2 and 3 fully opened) we sometimes suffer from Max command gear torque error msg.
The trajectory is smooth and the weight of tool is defined properly, however, probably due to gravity the robot is limited to lower velocity and acceleration commands when close to max reach.
We tried setting up a prediction model but it does not seem to fully correlate to the calculated torques from RSI (in all three possible options on RSI torque block) - some non linear behavior. We had to use ML to fine-tune the model to properly predict and saturate the torques as the same output from the RSI.
Was any one here able to successfully predict the torques from the robot and use it to avoid max command gear torque?
BTW, the RSI had around 5 samples delay from sending to actual execution. Not sure if it is related.
Best regards,
Avi