Convergence properties of a new technique for estimating parameters in a nonlinear viscoelastic lung model in newborns

The aim of this study was to develop a new technique to estimate parameters (airway resistance, inertance, tissue damping, and elastance; RIGH) of a viscoelastic lung model. The nonlinear RIGH-model was linearized by re-parametrization (model linearization, ML), and the parameters were calculated by one-dimensional line search of least-squares estimations. The convergence properties, the number of iterations, and computing time were compared between different search algorithms using the frequency responses of small animals and infants without and with added noise. While all of the algorithms converged in case of undisturbed frequency responses, only two algorithms converged in case of noise. ML provided always the lowest number of iterations and the shortest computing times. ML allows for reliable and accurate parameter estimation of the RIGH model.