Investigating complex collision behavior of inertial active particles using statistical methods and machine vision

We explore a macroscopic active matter system with centimeter scale robotic crawlers. In this system, our self-propelled particles are subject to both active noise and inertia. This is distinct from microscopic systems where inertial effects are often ignored or much larger systems where the role of noise is minimal. Our crawlers (i.e. Hexbugs) use an internally rotating motor to drive motion across a dry surface to exhibit Brownian-like motion with inertial persistence. The crawlers are housed within a lightweight circular container to create a hexbug-cup composite particle that is self-propelled and is geometrically isotropic. The resulting collisions between the active particles and their confining environment are surprisingly complex and exhibit rich behavior beyond that expected for elastic or inelastic interactions.. We use particle tracking and image processing techniques to track the motion and infer fluctuating forces of our self- propelled particles on flat and curved surfaces. Using statistical methods and machine learning we investigate collision behavior for this active system.