Viscoelastic Model Identification of soft tissue Under Compressive Loading

Tissue modeling is a recent and widespread area in medical and engineering sciences. The material properties of various tissues have been measured in-vitro/vivo for decades. In-vitro tests means removing sample from the body and obtaining experimental data, while in-vivo test measures the material properties of tissues in living animals. Viscoelastic models are capable to parameterize time-dependent dynamics and behaviors of biological tissue by relatively small errors. Various forms of viscoelastic models have been developed, where comprehensiveness and simplicity are necessities of modeling of soft tissue mechanical properties. In this paper’s approach ARMAX (Auto-Regressive Moving Average eXogenous input) model is considered. Experimental data are collected from indentation experiment performed on a chicken breast in in-vitro test. A model order selection procedure and Least Squares Error (LSE) identification algorithm have been utilized to estimate model parameters. The effect of measurement noise is considered during identification and eventually it has been shown that a 3rd order ARX model yields the best accuracy accompanied by mathematical simplicity. Accuracy of estimated model is quantitatively validated using FPE (Finite Prediction Error) criterion. Moreover, estimated model output is compared to real experimental data

Keywords:

ARMAX model , viscoelastic model , Least Squares , model validation , FPE criterion