Adaptive controller for a low-cost mechanical ventilator

Abstract

Mechanical ventilation is a key tool in treating patients with respiratory failure, where precise control of variables such as airway pressure and flow is essential for safe and effective support. This paper proposes an adaptive cascade control strategy for airway pressure regulation in a low-cost mechanical ventilator (VMAR+), addressing challenges such as patient variability, model uncertainties, actuator nonlinearities, and communication delays. The control scheme combines cascade control with feedforward action and a Smith predictor to compensate for dominant time delays. Lung resistance and compliance are estimated online using a Recursive Least Squares algorithm, enabling continuous updates of the model and controller parameters. The approach is validated experimentally on an artificial lung under varying respiratory conditions, including changes in compliance. Despite hardware limitations, the system achieves stable and responsive pressure control, highlighting the potential of the adaptive control strategy proposed for real-world ventilator applications.