Modeling, Control, and Numerical Simulations of a Novel Binary-Controlled Variable Stiffness Actuator (BcVSA).

Area of Science:

• Robotics
• Control Systems Engineering
• Mechanical Engineering

Background:

• Safe human-robotic interaction necessitates compliant robotic actuators.
• Existing actuators often lack adaptable stiffness control.
• Variable stiffness actuators are crucial for advanced robotic applications.

Purpose of the Study:

• To present the design, modeling, and control of a novel Binary-Controlled Variable Stiffness Actuator (BcVSA).
• To evaluate the effectiveness of Multiple Model Predictive Control (MPC) strategies for BcVSA control.
• To compare the performance of different MPC variants and traditional control methods.

Main Methods:

• Developed a BcVSA utilizing a stiffness varying mechanism with clutches and torsional springs.
• Implemented three Multiple Model Predictive Control (MPC) strategies: Multiple MPC, Multiple Explicit MPC, and Approximated Multiple Explicit MPC.
• Performed numerical simulations to assess controller robustness and performance against Computed Torque Control (CTC) and Linear Quadratic Regulator (LQR).

Main Results:

• Multiple MPC and Multiple Explicit MPC demonstrated comparable robustness.
• Approximated Multiple Explicit MPC offered offline computation of sub-optimal results but with lower robustness.
• MPC-based controllers were compared with CTC and LQR, providing insights into their relative effectiveness.

Conclusions:

• The BcVSA concept is viable for compliant robotic manipulators.
• Multiple MPC and Multiple Explicit MPC are promising control strategies for variable stiffness actuators.
• Further experimental validation on a hardware prototype is planned to confirm simulation findings.