Title: Insect-inspired Visually-guided Decentralized Swarming

Abstract: This paper addresses the need for fast, lightweight, vision-guided swarming under limited computation and no explicit communication network or position source. The study develops a multi-agent optic flow sensing framework, then integrates perfect information distributed feedback with optic flow sensing to create an analogous visually-guided feedback path for idealized inter-agent velocity and distance structures. The Cucker-Smale flocking example is used to develop vision-guided swarming with rigorous asymptotic convergence guarantees, including under ignorance of agent size.