Thesis Dissertation Defense

Abstract: In movement on solid terrains, animals, vehicles, and robots can make use of well-established contact dynamics for planning movement and locomotion gaits. However, when the terrain can be deformed significantly, effective traversal can be inhibited by terrain heterogeneities created before and/or during locomotion. Understanding the physical behavior of such deformations such complex forcings like locomotion can inform robotic navigation strategies and expand our physical intuition of soft matter substrates. We examine various rate-dependent phenomena for a specific class of flowable substrates abundant in the natural world: granular media, which exhibit multiphase and hysteretic properties as a collective of many small rigid bodies. The physics of granular substrates is dominated by a network of frictional contacts between simple particles, which nevertheless display a wealth of unexpected multiphase phenomena depending on their stresses and packings. We present a series of experimental studies on such media. An anthropogenic mode of terrain traversal, wheeled locomotion, can locomote via the reaction force generated from actively shearing a granular substrate. We experimentally show how such locomotion can induce rate-dependent weakening via the centripetal acceleration of the media under shear and present a fundamental physics-based cause for why vehicular slippage occurs at high wheel rotation rates. In another robophysical study, we demonstrate how a small rover robot can effectively remodel steep granular slopes via strategic open-loop gait selection. By selectively avalanching frictional media towards itself, the robot could traverse loosely consolidated granular hills that otherwise would not be possible for it to climb. Finally, we investigate how directional fluidization of granular media during intrusion could modulate the resistive forces to allow a body to move efficiently within a substrate that constantly attempts frictionally hinder it. This dissertation showcases not only the strength of experimental investigation in physics to discover new phenomena, but also how simple reduced-order models can be adapted to explain them and synthesize our observations for applications in robotics, terradynamics, and more.