Work

Synchronising coronary artery system via physics-informed neural feedback control

Year: 2025

Type: article

Abstract: Synchronisation of nonlinear biomedical systems remains a critical challenge in cardiovascular modelling. Traditional control strategies often lack robustness when applied to chaotic and parameter-sensitive coronary artery models. The present work introduces a physics-informed neural network based feedback control framework. It directly embeds the governing differential equations into the learning process. This approach allows accurate modelling and control of highly nonlinear dynamics. The developed method ensures rapid and stable synchronisation of diseased and healthy coronary artery system states. It achieves a reduction in mean absolute error by over 80% and maintains a root mean square error below 0.05. These results are obtained in comparison with baseline linear state-feedback controllers. High Pearson correlation coefficients ( ≥ 0.90) confirm accurate trajectory tracking under diverse initial conditions. Overall, the findings emphasise the advantages of integrating physics-informed learning with control design. This integration offers a scalable approach to biomedical system synchronisation. Furthermore, it lays a strong foundation for future studies on uncertainty handling and efficient control implementation.

Source: Biomedical Signal Processing and Control

DOI: 10.1016/j.bspc.2025.109311

Authors: K. Sanjay, P. Balasubramaniam

Institution: Gandhigram Rural Institute

Language: English

Topic: Model Reduction and Neural Networks

Subfield: Statistical and Nonlinear Physics

Field: Physics and Astronomy

Domain: Physical Sciences

SDG: Good health and well-being

Open Access status: hybrid

Estimated APC paid: $2,590