Feeling your neighbors: Lagrangian methods for quantifying collective motion

Douglas H. Kelley, Assistant Professor, Department of Mechanical Engineering

Hopeman 224

From herds to flocks, swarms to epithelial sheets, organisms move collectively at all length scales and for many evolutionary reasons. The tools of statistical mechanics can characterize and quantify animal collective motion, yielding insight into nature and strategies for human-made multi-agent systems. Recent successes in simulation highlight the dearth of experimental data, where Lagrangian tracks of individual motion are badly needed. I will talk about two studies that yield that kind of data. First, by tracking individual cells in an epithelial sheet, we find that non-affine rearrangement of neighbors is more likely where they are crowded, a trend starkly different from inanimate granular materials. Second, by tracking individual midges in a three-dimensional mating swarm, we characterize a global energy potential and a local exclusion zone for each midge. By measuring flight paths between neighbor interactions, we find that discrete Lagrangian models are likely needed for accurate simulation.