A suite of robust controllers for the manipulation of microscale objects

Area of Science:

• Robotics and Control Systems
• Microelectromechanical Systems (MEMS)

Background:

• Microscale object manipulation in MEMS is challenging due to dominant unknown forces like adhesion and surface tension.
• Actuator constraints and measurement noise further complicate precise control in these systems.

Purpose of the Study:

• To introduce novel robust controllers for reliable microscale object pickup in MEMS.
• To address challenges posed by unknown dynamics, actuator limitations, and measurement noise.

Main Methods:

• Development of a robust controller assuming known upper bounds on dominant forces.
• Implementation of a robust adaptive critic-based neural network (NN) controller for online estimation of unknown forces.
• Design of an output feedback controller using a high-gain observer and a reference system to handle unavailable states and measurement noise.
• Utilization of the Lyapunov approach to prove the uniform ultimate boundedness of the manipulation error.

Main Results:

• The proposed controllers ensure performance despite unknown contact dynamics and actuator constraints.
• The adaptive NN controller effectively estimates unknown forces online.
• The output feedback controller successfully mitigates measurement noise and unavailable states.
• Theoretical conclusions are substantiated by simulation results.

Conclusions:

• The developed robust controllers are effective for microscale object pickup in MEMS.
• The NN-based approach offers adaptive compensation for unknown dynamics.
• The output feedback strategy provides robustness against measurement imperfections, paving the way for practical applications.