Internship: Reliable Force control on the R1 Glove

            Proposing a solution while taking the following
            into account

• Increasing Kp or adding I/D terms can introduce instability due to noise and system dynamics.
• The controller operates at approximately 400 Hz.
• No motor encoder is available; only raw sensor values from the exoskeleton encoders can be used at low latency.
• When the cable goes slack, sensor readings may differ from actual motor movement.
• The controller board cannot run expensive kinematic computations.

         
          Propose a strategy for a better controller

• Develop a control strategy that improves force-feedback stability in the ~400 Hz control loop while dealing with sensor noise and dynamic effects that limit aggressive gain tuning.
• Use only low-latency raw exoskeleton encoder signals, since no motor encoder is available and computationally expensive kinematic processing is not allowed on the controller board.
• If device characterization is part of the strategy, clearly define the experimental tests: specify motor command signals, whether finger movement is allowed during testing, and which signals will be measured to analyze system behavior.