Non intrusive physical parameter inference for a hyperelastic soft robot segment
Please login to view abstract download link
The predictive modelling of soft continuum arm behaviour under quasi-static/dynamic conditions is often based on merging finite element models coming from the physics-based principles with experimental data. In this paper we pose the parameter estimation problem in a probabilistic setting seen from a Bayesian point of view. Parameters and observations are estimated online without the need for restarting the FEM analysis using a hierarchical general Kalman filtering approach.
