A rapidly converging algorithm for estimating respiratory mechanical parameters in a five-element model

A rapidly converging algorithm for computing values for respiratory mechanical parameters from forced random noise independance data was developed and verified. The algorithm, which was based on a five-element Mead-type model, minimized the sum of squared differences between the model's response and experimental data, while imposing a nonnegativity constraint on the parameter values. It yielded parameter values that showed excellent agreement with values obtained previously using standard nonlinear regression analysis, but required much less computer time, 10 s versus 1 h. When this algorithm is coupled with the forced random impedance data collection techniques, it provides a rapid noninvasive method for estimating respiratory inertance, central resistance, peripheral resistance, and airway compliance. The problem of estimating peripheral compliance was not solved by this algorithm.