Abstract: This talk will present ongoing research in the Collective Dynamics and Control Lab at the University of Maryland on the topic of motion coordination in a dynamic flowfield with applications pertaining to environmental sampling in the ocean and atmosphere. In order to gain new insights for multi-robot applications, we have constructed a multi-target tracking framework using tools from computer vision and estimation theory to automatically process video footage of two types of animal groups. First, we study non-verbal information transmission in schooling fish immersed in a flow tank in order to test the prediction that school polarization increases sensitivity to external threats. Second, we seek to understand swarming and mating behavior in wild malarial mosquitoes by reconstructing and analyzing the three-dimensional flight trajectories of individual mosquitoes in large swarms. Our work in robot coordination is characterized by the Lyapunov-based design of decentralized multi-vehicle control algorithms for idealized models of vehicle motion, including self-propelled particle and (planar) rigidbody models. We have constructed a multi-vehicle control framework that enables motion coordination in spatiotemporal flowfields — in two or three dimensions — even when the flow speed exceeds the platform speed relative to the flow. Although the control framework uses inter-vehicle relative positions and relative velocities, it can estimate relative velocity when only relative-position measurements are available. Motivated by the need to conduct operations in an unknown flowfield, we have designed an observer-based controller that reconstructs a spatially varying flowfield by assimilating noisy position measurements into a distributed information-consensus filter. Time permitting, the talk will describe the application of the estimation and control framework to the problem of using unmanned aircraft to improve forecasts of hurricane intensity, which is the topic of an ongoing collaborative research project.