ISSN print edition: 0366-6352 ISSN electronic edition: 1336-9075 Registr. No.: MK SR 9/7 Published monthly

A recent survey of artificial intelligence-based modelling and control of droplet dynamics in microfluidics

Zarfishan Kanwal, Shazia Bashir, Asifullah Khan, and Sikander M. Mirza

Department of Physics and Applied Mathematics, Pakistan Institute of Engineering and Applied Sciences (PIEAS), Nilore, Islamabad, Pakistan

E-mail: shazia@pieas.edu.pk

Received: 3 November 2025  Accepted: 9 February 2026

Abstract:

Over the past decade, the integration of artificial intelligence (AI) into droplet microfluidics has emerged as a transformative approach, accelerating innovation in design, automation, and data-driven analysis. In this review, we explore advances in AI-enhanced droplet microfluidic systems from 2017 to 2025, emphasizing how machine learning, deep learning, and reinforcement learning models optimize droplet generation, manipulation, sorting, and monitoring. We examine how researchers have used AI to improve key parameters such as flow stability, droplet monodispersity, and real-time control in both passive and active microfluidic platforms. We also highlight significant applications across biomedical diagnostics, single-cell analysis, real-time cell sorting, and material synthesis, particularly for polymeric and metallic nanoparticles. By systematically mapping how AI tools are reshaping experimental workflows, we reveal the growing synergy between computational intelligence and microfluidic engineering. Despite these advances, the field still faces critical challenges, including limited high-quality datasets, poor model transferability across platforms, and difficulties integrating AI algorithms with physical systems. We conclude by identifying emerging research directions that could bridge current gaps and unlock new capabilities. This review provides a timely, structured resource for scientists seeking to explore and contribute to this rapidly evolving interdisciplinary domain.

Keywords: Microfluidics; AI algorithms; Active method; Passive method; Deep learning

Full paper is available at www.springerlink.com.

DOI: 10.1007/s11696-026-04735-3

Chemical Papers 80 (5) 5677–5696 (2026)