Summary：Researchers have developed an AI-assisted optoelectronic micromotor control technology that enables precise, predictable, and independent navigation of multiple micromotors. By combining programmable light fields, electrokinetic propulsion, and deep-learning prediction models, the platform could advance targeted drug delivery, microassembly, lab-on-a-chip systems, and next-generation microrobotics.
A research team has introduced an advanced AI-Assisted Optoelectronic Control Strategy that transforms the traditionally random motion of active micromotors into agile, programmable, and highly predictable navigation. The technology combines artificial intelligence, machine vision, electrokinetic propulsion, and dynamically reconfigurable optical patterns, enabling unprecedented control of microscopic robotic systems.
At the heart of the platform is a new navigation architecture that leverages the combined effects of Induced-Charge Electrophoresis (ICEP), Dielectrophoresis (DEP), and Alternating-Current Electroosmosis (ACEO). Rather than relying on magnetic fields or fixed microstructures, the system uses programmable light patterns to create dynamic electric-field landscapes that guide micromotors in real time. This approach enables omnidirectional movement, including forward propulsion, reverse motion, in-place rotation, U-turns, pause-and-restart operations, and adaptive path switching.
Development of multiple motion primitives induced by two or more optical patterns.
A major breakthrough is the integration of a CNN-Transformer spatiotemporal prediction model, designed specifically to understand how optical-pattern geometry influences micromotor behavior. The AI model analyzes pattern width, curvature, relative position, and motion history to accurately predict both instantaneous velocity and long-term trajectories. This predictive capability allows researchers to optimize navigation strategies before deployment, significantly improving motion reliability in complex environments.
Spatial-temporal prediction based on AI model.
The research team also demonstrated independent control of multiple micromotors operating simultaneously within maze-like optoelectronic networks. Individual micromotors successfully navigated different routes, avoided collisions, adapted to changing pathways, and responded autonomously to dynamically reconfigured optical guidance. Such capabilities represent a major step toward scalable multi-agent microrobotic systems.
AI-assisted reconfiguration of optical patterns for long-term navigation of multiple micromotors.
The work builds upon earlier advances in programmable optoelectronic micromotor steering developed by researchers from Harbin Institute of Technology (Shenzhen), the University of Toronto, Beijing Institute of Technology, and collaborating institutions, where light-controlled electric fields were first used to guide self-propelled Janus micromotors through complex microenvironments.
Beyond navigation, the team demonstrated microscale transportation networks featuring intersections, channel splitters, storage zones, and programmable routing systems. These structures function similarly to miniature traffic networks, allowing micromotors to be transported, sorted, stored, released, and redirected on demand.
As AI-driven microrobotics continues to evolve, technologies such as the AI-Assisted Optoelectronic Control Strategy could enable future applications in targeted drug delivery, precision medicine, microfabrication, autonomous laboratory systems, environmental monitoring, and intelligent lab-on-a-chip platforms, bringing autonomous robotics into the microscopic world.
