Collaborative locomotion of artificial microswimmers through reinforcement learning

Artificial microswimmers offer exciting opportunities for biomedical applications, such as microsurgery and targeted drug delivery. Designing artificial microswimmers that can navigate in complicated biological micro-environments such as blood vessels and tumor environments has been of great research interest over the past decades. Collaborative navigation of multiple artificial microswimmers is particularly challenging due to complex hydrodynamic interactions between swimmers. In this work, we utilize deep reinforcement learning to obtain the effective locomotory policy of two collaborative colinear microswimmers. Our learning results demonstrate that phase shift and relative distance are the key factors that influence the net propulsion of the two collaborative swimmers. This reinforcement learning approach opens an alternative avenue to investigate the collaborative navigation of multiple artificial microswimmers.