Abstract

This paper presents a robust control approach for simulating respiratory rhythms using a physical breathing simulator that integrates both respiratory mechanics and blower dynamics. The system models a respiratory rhythm generator driven by a brushless DC motor-based blower, combined with a lumped-parameter artificial lung. To address parameter uncertainties such as time-varying airway resistance and lung compliance arising from physiological variability and modeling inaccuracies, a nominal model sliding mode controller is developed within a repetitive control framework. The proposed scheme consists of two loops: a nominal controller ensuring ideal model tracking, and a plant controller enforcing convergence of the actual output to the nominal response, regardless of parameter mismatches or disturbances. The control algorithm is derived analytically using Lyapunov stability theory and designed to minimize the tracking error of the periodic respiratory volume waveform. To evaluate control system performance, a nonlinear simulation model was constructed incorporating electromechanical dynamics of the blower and physiological variability in lung parameters under time-varying operating condition. Numerical simulations are conducted under three reference inputs with varying tidal volumes and respiratory rates, covering a representative range of breathing conditions. As a result, it demonstrates that the proposed controller reserves high precision across all cases, with normalized root mean square error (NRMSE) consistently exceeding 99.85% and a correlation coefficient of 1.0. The phase trajectory analysis further confirms rapid convergence of tracking error dynamics and bounded steady-state behavior, while the use of a boundary layer saturation function effectively reduces control chattering without degrading tracking performance. This study establishes a high-fidelity simulation platform suitable for ventilator algorithm development, performance testing, and control design under realistic uncertainties. From there, it enables improved robustness evaluation and advanced controller validation in diverse physiological breathing scenarios.