Data-Driven Adaptive Control of Multiple Active Micromotors With Optoelectronic Guidance for Independent Transport and Cargo Delivery

Active micromotors, which self-propels with optoelectronic guidance, hold great promise in biomedical applications such as immune-sensing and antibacterial, where motion accuracy and task efficiency are highly desirable. However, independent transport and cargo delivery for multiple micromotors is great challenging due to the time-varying self-propulsion velocity and nonlinear interactions with cargos. Here, we propose a novel data-driven adaptive control method for multiple micromotors, which enables independent control of micromotor motion and cargo delivery behaviors on demand. The motion and interaction states can be guided by programmable light patterns in optoelectronic systems. By dynamically adjusting parameters of light patterns, the state of multiple micromotors were decoupled, enabling individual trajectory and velocity for each micromotor. A data-driven motion controller was developed with consideration of system nonlinearity and random interference. The controller enabled online updating to learn micromotor differences, and providing individual features for optimal prediction to enhance the cargo delivery performance. The data-driven controller was validated by accurately controlling several micromotors with diverse reference velocities in a confined space with 10 collision-risk regions. The time-average velocity error was consistently within 1/2 of body length per second, and the success rate of cargo delivery was improved from 42% to 90% with optimal prediction.