Dataset for system identification of an automated self-inflating bag resuscitator under varying test-lung mechanics

The data presented in this article describe dynamic input-output responses of an automated self-inflating bag resuscitator operating under varying simulated respiratory mechanics. The dataset was acquired from a custom-built motor-driven resuscitator platform developed at the Key Laboratory of Digital Control and System Engineering (DCSELab), Ho Chi Minh City University of Technology, Vietnam. Excitation signals were applied to a direct-current (DC) motor to mechanically compress a standard adult self-inflating bag, while airway pressure, airflow rate, and tidal volume were measured using integrated pressure and flow sensors at a sampling frequency of 1000 Hz. Respiratory mechanics were emulated using a dual adult test lung with adjustable compliance and airway resistance, representing four clinically relevant scenarios: normal lung, stiff lung, obstructed airway, and an extreme combined condition. The repository contains 12 CSV files organized into 4 folders, each corresponding to a different lung configuration. Each folder includes three datasets generated under distinct geometric reference trajectories (cosine-step, parabolic-step, and ramp-step) tracked via a closed-loop control loop. This closed-loop excitation strategy ensures that the resulting commanded pulse-width modulation (PWM) signal possesses sufficient spectral richness and strictly satisfies the persistency of excitation criteria for robust system identification. Each file provides synchronized time-series measurements of gripper position, PWM duty cycle, airway pressure, airflow rate, and derived tidal volume. All data are stored in raw format with clearly labelled columns and physical units to facilitate reproducibility. This dataset provides empirical dynamic data for modelling the nonlinear and time-varying behaviour of automated self-inflating bag resuscitators. It can be reused for developing and benchmarking system identification algorithms, constructing digital twins of low-cost automated self-inflating bag resuscitators, and designing advanced control strategies under diverse simulated respiratory conditions.