Synthetic Machines from Feedback-controlled Active Matter
Biological cells exhibit remarkable functionalities, such as motility, division, and self-healing. Reproducing these life-like behaviors in synthetic materials would both revolutionize engineering and advance fundamental science. Active fluids, which are composed of motile energy-consuming microscopic units, are a promising platform for achieving these ambitious goals. In contrast to widely studied conventional passive materials, active fluids generate internal forces that drive persistent autonomous motion, an alluring life-like feature. On their own, however, bulk active fluids exhibit chaotic flows. Thus, they are unable to perform useful functions such as generating work or driving net material transport. By seamlessly merging experiments, theory and machine learning methods, this project aims to harness the chaotic dynamics of active fluids to achieve functional behaviors. In particular, the project will measure the instantaneous configuration of a light-responsive active fluid and use model-dependent theory and/or model-independent machine-learning methods to forecast its evolving dynamics. This information will impose theory-guided external signals that steer the system toward a targeted state such as a persistent rotation of an inclusion or cell-like persistent crawling of a deformable droplet that encapsulates an active fluid. The project will also pursue several tightly integrated education and outreach activities focused on (1) providing rigorous training and mentoring in interdisciplinary sciences to graduate and undergraduate students, (2) encouraging underrepresented groups to pursue work in STEM-related fields, (3) and raising general awareness of the importance of scientific research to broader communities.