Proxy Test Structures for Improved Hyperelastic Material Parameter Estimation for Soft Robotic Components

Soft robotics relies on compliant materials like silicones that enable gentle and adaptive interaction with the environment. For effective sim-to-real transfer, accurate mechanical simulations are essential. Soft materials are typically modelled using hyperelastic formulations such as Ogden. Material parameters are obtained from standardized test structures via least squares fitting, a process that can be ill-posed. Here, we extend the standard approach by introducing design-specific 2D axisymmetric proxy test structures (2DAxTS) that better capture the behaviour of complex geometries, such as artificial silicone fingertips. Global optimization on these proxies improved agreement between experimental and simulated force-displacement (FD) curves, reducing the normalized root mean squared error by 39.1% and 17.8% for two different designs. This highlights the importance of geometry-aware material parameterization for enhancing simulation fidelity in soft robotics.