Three-dimensional Piecewise Constant Curvature dynamics for a soft robot in hardware
OA Version
Citation
Abstract
The Piecewise Constant Curvature (PCC) assumption is a well-proven technique to model a continuum soft robot whose dynamics is “infinite-dimensional” in reality, but the commonly employed angle-based state parametrization falls short of expanding the current paradigm of soft robotics controls beyond planar motion and into true three-dimensional motion. However, a novel, alternate, arc length-based parametrization has recently been developed which this thesis aims to evaluate and validate, by performing a system identification using the model on a real pneumatically-actuated soft robot. In order to move beyond the simulative realm, a transformation mapping the pressure levels of the pneumatic actuators to the control input as defined by the model must be formulated. This mapping must not nullify the assumptions that the PCC approximation and this parametrization hinge on but rather build upon it. This thesis shall approach this challenge by building on motivating assumptions and first principles to propose an expression for this mapping, before then applying it in the system identification process involving real datasets, and finally evaluating the accuracy of the dynamic model against the dataset when using both the parametrization and the proposed mapping.
Description
2026
License
Attribution-NonCommercial-ShareAlike 4.0 International