Identification of respiratory parameters in frequency and time domain with Forced Oscillation Technique

摘要

Respiratory impedance was measured in healthy subjects using Forced Oscillation Technique (FOT). Measurements with a FOT system were performed for five healthy male adults. Based on these measurements, we compared parameter estimation approaches in frequency and time domain. In frequency domain, the fast Fourier transform and windowed cosine fitting methods were applied to calculate respiratory complex impedance. Using these resulting impedance, we performed parameter identification with least squares fitting for RLC lung model and some models with more complex structures. In time domain, the extended Kalman filter was applied for online parameter estimation. We introduced the approach Sum of Frequencies for a stable convergence of the algorithm. Finally, a comparison between our results of different methods with some literature references is presented and an interesting conclusion on the frequency dependence of lung compliance is given.