Self-Sensing Control for Soft-Material Actuators Based on Dielectric Elastomers.

Area of Science:

• Robotics
• Materials Science
• Control Systems

Background:

• Dielectric elastomer (DE) transducers offer muscle-like properties, making them ideal for soft robotics.
• Their integrated actuator and sensor capabilities eliminate the need for external sensors, simplifying designs.

Purpose of the Study:

• To present a novel self-sensing control strategy for DE stack-transducers.
• To enable flexible control of various DE transducer quantities (voltage, force, deformation) using a single controller.
• To facilitate the development of human-machine interfaces with soft-bodied robots.

Main Methods:

• Development of a self-sensing state and disturbance estimator using an extended Kalman filter.
• Design of a sliding mode energy controller tailored for a bidirectional flyback converter.
• Implementation of feed-forward controls for precise regulation of voltage, force, or deformation.

Main Results:

• The novel estimator does not require superimposed excitation, allowing compatibility with economical power electronics like flyback converters.
• The self-sensing control strategy achieves results comparable to traditional sensor-based approaches.
• Validation confirms the effectiveness of the integrated self-sensing estimator and control system.

Conclusions:

• The developed self-sensing control for DE transducers offers a flexible and efficient solution for soft robotics.
• This approach enables advanced functionalities in human-machine interfaces by integrating sensing and actuation.
• The system's compatibility with cost-effective power electronics broadens its applicability in soft robotic systems.