Artificial potential field-enhanced optoelectronic tweezer technology for path planning and intelligent sorting of particles

Efficient extraction of specific particles from heterogeneous samples is critical for both fundamental research and clinical diagnostics. Optoelectronic tweezer (OET) technology, characterized by its non-contact and label-free manipulation capabilities, has emerged as a promising tool for such applications. However, conventional OET systems often rely on predefined optical patterns and offline trajectories, limiting their adaptability in complex or dynamic sample environments. To overcome these limitations, we propose an automated particle sorting strategy for OET systems based on an artificial potential field framework. This approach integrates real-time multi-object detection and tracking algorithms to enable accurate classification of particle types, while the artificial potential field dynamically optimizes the motion paths, effectively minimizing the risk of inter-particle collisions. Experimental results demonstrate that the developed system supports parallel, automated manipulation of multiple particles, achieving label-free, non-contact, real-time, and high-precision sorting. In validation experiments using mixed polystyrene particles, the method achieved a sorting purity of up to 95.11%. This study proposes an AI-enhanced optoelectronic-tweezers particle-sorting scheme that holds considerable application potential in biomedical and related fields.