Study problem related to how joint trajectory is defined when calibrating with calibration type 10

[MSECode]

As shown in this 2 comments:

• Fix AMO calibration 10 #529 (comment)
• Fix AMO calibration 10 #529 (comment)

we need to understand the "possible problem" related to how the joint trajectory is defined when using calibration type 10 for calibrating the joint at the hard stop.
One thing to mention is that, from the data we have dumped we have identified this behaviour as a possible problem, but we first need to better analyze that to be sure that it is actually a problem and why it is happening.

cc: @valegagge

[MSECode]

Analyzing the behavior of the whole calibration procedure on the bench setup with the AMO encoder we can conclude that the spikes that we have seen is due to the fact that, considering how calibration type 10 has been developed, the joint is forced to reach the position in degrees defined in the parameter calibrationZero in the configuration file after the calibration. This is done because of we are using the encoder as an incremental one and thus we cannot know its starting position and how much space it is spanning while calibrating. Thus, with the command of the trajectory for moving to the zero position, we also force the joint to measure the value in degrees that we can configured in the calibrator as shown by the image below.

[valegagge]

Checking the code I noticed in the function JointSet_do_wait_calibration , the embeddeed motion control performs a JointSet_do_odometry and then set the joint in position control mode when the hardstop is reached. I guess these two operations are the cause of the change of the joint position and of the trajectory.

@MSECode Can you check and confirm, please?

[MSECode]

Checking the code I noticed in the function JointSet_do_wait_calibration , the embeddeed motion control performs a JointSet_do_odometry and then set the joint in position control mode when the hardstop is reached. I guess these two operations are the cause of the change of the joint position and of the trajectory.

@MSECode Can you check and confirm, please?

Yes exactly.

[valegagge]

ok so we can close this issue.