Design, fabrication and modelling of soft robots with novel actuation and materials

Soft robots are gaining attention due to their potential for safe human-robot interactions and delicate operations while maintaining low commercial costs. However, challenges remain in promoting the use of electroactive polymers (EAPs), particularly in improving product quality (particularly in beha...

Full description

Bibliographic Details
Main Author: Chang, Jung-Che
Format: Thesis (University of Nottingham only)
Language:English
Published: 2025
Subjects:
Online Access:https://eprints.nottingham.ac.uk/80512/
Description
Summary:Soft robots are gaining attention due to their potential for safe human-robot interactions and delicate operations while maintaining low commercial costs. However, challenges remain in promoting the use of electroactive polymers (EAPs), particularly in improving product quality (particularly in behaviour consistency and product lifetime) and developing dynamic models that enable coordination between multiple actuators. To address these issues, we designed a robot incorporating an angular stroke amplification mechanism to optimise the balance between actuation force and stroke. The frame’s weight was significantly reduced by employing 3D puzzle-like strip structures. A novel method for constructing a stable conductive paint on a silicone-based dielectric film was developed, using laser engraving to pattern the film and applying a sandwich-type electrode made from conductive particles (carbon black) and photosensitive resin (Formlab elastic 30A). Compared to conventional carbon grease electrodes, this approach fully cures the resin and prevents the dehydration-caused electro resistance and stiffness increment, demonstrating greater consistency in dynamic performance, especially after extended periods of use. A key contribution of this work is the development of an dynamic model that incorporates an expanded Bergstrom-Boyce model to capture the constitutive behaviour of the dielectric film, addressing the effects of hyper-elasticity, hysteresis, and viscoelasticity. Experimental validation showed that the model accurately modeled the robot’s output force with a root mean square error (RMSE) of 12.4% in stationary phases and achieved path tracking with an RMSE of less than 2.5%. Furthermore, the hysteresis-considered model was applied to a planar, ultrathin robot with high viscoelasticity for forward control, demonstrating the versatility of the model across different applications.