Topology-based coordination control for multi-droplet tasks in autonomous digital microfluidics

Digital microfluidics (DMF) is a versatile technique for parallel and field-programmable control of individual droplets. The challenge of large-scale parallel droplet manipulation in DMF is essentially a cross-scale complex system control problem that combines multi-droplet coordination optimization and feedback control. Here, we develop an unmanned topology-based digital microfluidics control (TDMC) system that integrates adaptive path planning with semantic segmentation feedback for autonomous multi-droplet operations. The core innovation lies in a dynamic droplet-electrode topological graph that both unifies the representation of droplets with arbitrary sizes and morphologies and resolves inter-droplet conflicts. Building upon this representation, the adaptive-topology path planning algorithm implements a leading-vertex guidance mechanism to efficiently coordinate the movements of droplets covering multiple electrodes while preserving morphological integrity. By fusing an encoder-decoder semantic segmentation model with event-driven feedback control, the system achieves closed-loop autonomy for dynamic path reconfiguration and real-time task adaptation. Experimental validation demonstrates that the TDMC system successfully handles complex multi-droplet scenarios including morphological adaptations, obstacle avoidance, and dynamic droplet operations, achieving complete on-chip automation of biological assay workflows. Thus, this unmanned TDMC system provides an adaptive, flexible, and robust microfluidic manipulation for point-of-care testing, high-throughput drug screening, and synthetic chemistry.