Modelling the ventilator-patient interaction: a pressure-cycled control strategy

In the last few years, mathematical models of lung ventilation have often been used to support the Anesthesiologists and Resuscitators choices in the mechanical ventilator parameters setting. In this context, a real-time control strategy is doubtless crucial to avoid the occurrence of induced pressurederived trauma. In the present work, we develop a first version of a simple but realistic physiological lung ventilation mathematical model. The patient-ventilator complex is taken into account by modeling the pressure wave provided by the mechanical lung ventilator as an external (control) input. With the aim of reaching the correct amplitude for the pressure wave at the mouth provided by the mechanical ventilator, hence limiting the risk of Acute Respiratory Distress Syndrome (ARDS), we consider two different scenarios: the patient who needs to be ventilated as a consequence of having an insufficient respiratory drift (assisted ventilation) and the patient without a spontaneous breathing (controlled ventilation). An output-feedback control law is proposed, based on the flow measurements provided by the ventilator, and not exploiting the full knowledge of the model equations and parameters. The considered approach looks promising, since a preliminary in-silico validation of the resulting patient-ventilator system shows that the target value of the tidal volume is readily tracked in both scenarios, without dangerous oscillations and with limited control effort.