hbrs-sdp / ss22-pour-feedback Goto Github PK

The following information are expected, please add anything else that you think would be relevant

• robot configuration and pose
• containers location (in-hand and on table)
• container material
• (simulated) "liquid" properties

This is expected for the minimum viable product (MVP) to be demonstrated at midterm. It may look something like this:

def ctrl(target, feedback, gains, **kwargs):
    while condition:
        actuation = compute_actuation(target, feedback, gains)
        ctrl_robot(actuation)

Please think about the different variational points that can influence the design of your software. Some of these are:

• Timing requirements if the control algorithm involve differentiation and integration of the error signal (just sleep() won't work)
• What if you want to use a different control algorithm, how to reuse the code when you have different compute_actuation() implementations and different gains
• What are your assumptions about the properties of the feedback and actuation signals? Do you have to change the implementation significantly if these assumptions change?