A Semilinear Parameter-Varying Observer Method for Fabric-Reinforced Soft Robots.

Area of Science:

• Robotics
• Control Systems
• Soft Materials

Background:

• Inflatable soft robots offer unique compliance and adaptability.
• Accurate state estimation (configuration, velocity, forces) is crucial for their control.
• Modeling hysteresis in soft robots presents a significant challenge.

Purpose of the Study:

• To develop an observer architecture for estimating configuration space variables, their rates of change, and contact forces of fabric-reinforced inflatable soft robots.
• To address the challenge of hysteresis in soft robot dynamics.
• To create a robust model suitable for observer and controller design.

Main Methods:

• Discretized the continuum robot into a series of discs connected by inextensible threads.
• Developed a linear parameter-varying (LPV) model with subsystems for different chamber pressure ranges.
• Transformed the hysteresis model into a semilinear form, creating a semilinear parameter-varying (SPV) model.
• Designed a SPV observer architecture with sub-observers for each pressure range.

Main Results:

• Simulations demonstrated the observer's ability to estimate configuration and rates of change with no steady-state error.
• Experimental results showed fast convergence of generalized contact force estimates.
• The observer achieved good tracking of the robot's configuration compared to motion capture data.

Conclusions:

• The proposed observer architecture effectively estimates key states of fabric-reinforced inflatable soft robots.
• The SPV modeling approach successfully incorporates hysteresis, enabling robust estimation.
• The validated observer provides a foundation for advanced control strategies in soft robotics.