A Modular Closed-Loop Wearable Platform for Sweat Extraction, Analysis, and Drug Delivery

Abstract

Precision medicine requires wearable systems capable of transforming real-time physiological sensing into adaptive therapeutic intervention. However, existing sweat-based platforms remain fragmented, typically focusing on sensing alone and relying on empirical threshold-triggered control strategies. Here we report a modular, model-driven wearable architecture that integrates active sweat extraction, multi-channel electrochemical sensing, and ion-electroosmotic transdermal drug delivery within a unified closed-loop framework. Active iontophoretic sweat extraction enables stable sampling under non-exertional conditions, while standardized multi-channel electrochemical interfaces support flexible biomarker configuration on a shared hardware backbone. Beyond hardware integration, we introduce a dual-layer control architecture combining data-driven physiological prediction with a physics-informed iontophoretic transport model, establishing a quantitative and interpretable mapping from sensed biomarkers to calibrated dosing decisions. Experimental results demonstrate accurate multi-biomarker sensing performance, including glucose sensing linearity (R² = 0.9913) and ascorbic acid sensing linearity (R² = 0.9965), together with near-linear controllability of transdermal delivery (R² > 0.99), consistent with the proposed transport model. By introducing a model-based quantitative regulation framework beyond conventional threshold-triggered actuation, this work provides a generalized system architecture and methodological foundation for precision and personalized wearable therapeutics.



Received: 12 March 2026 | Revised: 25 May 2026 | Accepted: 16 June 2026



Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.



Data Availability Statement

The source code underpinning the embedded control firmware, host-side application, backend server, and hardware design files is openly accessible in an open-source repository on GitHub (https://github.com/BITMaJinhang/EStat). The repository encompasses firmware source code, PCB schematics, backend implementation, and application interfaces utilized in the present study. All other relevant source data are available from the corresponding authors upon reasonable request.



Author Contribution Statement

Jinhang Ma: Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization. Shuo Yang: Data curation. Sijie Yin: Conceptualization, Writing – review & editing, Supervision, Project administration, Funding acquisition. Xiaohan Wang: Conceptualization, Methodology, Writing – review & editing. Ningning Song: Conceptualization, Writing – review & editing, Supervision, Project administration, Funding acquisition.