Paper Contents
Abstract
In the context of a global health emergency, the ability to remotely and autonomously monitor and regulate patient vitals has become increasingly crucial. This study presents an adaptive closed-loop control system designed to manage a patients mean arterial pressure (MAP) through the regulated infusion of sodium nitroprusside (SNP). The proposed system employs Active Disturbance Rejection Control (ADRC) to track the target MAP, while optimizing controller parameters using a Continuous Action Policy Gradient (CAPG) algorithm, a type of deep reinforcement learning (DRL). In this framework, the actor network formulates control policies and the critic evaluates their performance based on MAP error, with both networks trained via gradient descent. Comparative simulations indicate that the developed approach outperforms traditional methods by offering enhanced robustness and stability under diverse operational scenarios, fluctuations, and uncertainties, while precisely maintaining target MAP levels and optimal drug dosage.
Copyright
Copyright © 2025 Shreelekha S N. This is an open access article distributed under the Creative Commons Attribution License.
