Inverse Dynamics Algorithm Parameter Tuning

[Emmanuel]

Hello,

 

I have been attempting to combine the Sawyer's trajectory tracking and interaction control capabilities, into a unified algorithm, although it appears that the activation of the 'joint_trajectory_action_server' immediately precludes the usage of any interaction control features, such as impedance and/or force control.

As a preliminary test, it might be interesting to prevent the inverse dynamics algorithm''s broadcasting of computed feedforward torque values, to specific joints. Is the latter possible?

Alternatively, would it be feasible to modulate the PD gain values that are being executed on specific joints, in either a real-time or static fashion? My previous assertion is based on the premise that your custom inverse dynamics algorithm comprises both PD and feedforward model-based terms, which may not necessarily be the case.

 

Best regards,

Emmanuel