AI-assisted optoelectrokinetic control of active self-propelling micromotors for independent navigation with...

Agile and controllable motion is essential for active micromotors to navigate various microscale scenarios and manipulate the microscopic world, where parallel navigation and independent control are highly desirable. However, most micromotor navigation strategies suffer from limited agility and poor predictability, with motion typically confined to single-micromotor self-propulsion along predetermined trajectories. Herein, we propose an artificial intelligence (AI)-assisted optoelectronic control strategy that involves converting stochastic micromotor self-propulsion into controllable omnidirectional motion by synergistically exploiting multiple electrokinetic mechanisms. This strategy enables independent navigation of individual micromotors while simultaneously supporting parallel manipulation. By spatiotemporally configuring two or more optical patterns, agile motion primitives, such as directional propulsion, passive propulsion, in situ U-turns, and motion pause and restarting, were developed. To improve navigation robustness under coupled electrokinetic effects, a spatial-temporal AI model was developed for accurately predicting micromotor motion to facilitate the optimization of dynamic guidance schemes. These motion primitives are sequentially integrated and automatedly switched along long-term, reconfigurable trajectories, thereby enabling continuous navigation guided by discrete optical patterns. Independent control of active micromotors was demonstrated through the parallel manipulation of multiple Janus micromotors that were navigating intricate networks, in which each micromotor followed individual trajectories and adapted in real time to local terrain variations. This work showcased an agile and predictable navigation strategy for active micromotors that facilitates independent and massively parallel manipulation in intricate terrains, thus opening further possibilities for advanced applications.