Modern Physics of Autonomous Agility: From Wave-Particle Dualism to Spacetime Geometry
Abstract
In natural environments, robots lack the autonomous agility and manipulation capability of living systems. But as devices from life-sized humanoids to nano-swimmers enter the real world, they will encounter unstudied and a priori unknown environmental interactions. Recent approaches to locomotion control have relied on mathematically sophisticated control schemes specific to an environment, but these are often brittle, e.g., wheels in sand. Here we propose a new approach to creating robots that can move with life-like capabilities in complex cluttered environments, of interest to DoD. The approach is inspired by concepts in fundamental physics, namely gauge theory applied to locomotion by particle physicists in the 1980s (an approach recently validated by my group and collaborators), and our recent discoveries of mechanical ÒdiffractionÓ and geodesic dynamics in robot systems. Therefore, viewing robots as active and adapting ÒparticlesÓ through the lens of ÒModern PhysicsÓ -- quantum mechanics including statistical and hard/soft condensed matter physics and general relativity -- will facilitate discovery of novel emergent mechanical phenomena. We take this leap because we have discovered that robots (active & adaptive particles) can, should and do possess: 1) Wave/particle duality in that their persistent cyclic localized dynamics consists of waves traveling through the spatially finite ÒparticleÓ of the robotsÕ bodies; 2) Emergent properties corresponding to those in strongly-interacting systems arising from broken symmetries at bifurcations and phase transitions 3) Strong relativity-style coupling to the geometry of their landscape such that the robots cannot and should not be thought of as distinct therefrom. We posit that detailed Òphysics-styleÓ experimental and theoretical study of such interactions can form a key component of a (likely probabilistic) terradynamic locomotion framework, of interest across the DoD from search and rescue to reconnaissance and exploration. Such a framework would enable individual robots to blend into their locomotor environment and move with agility and grace in the most complex terrain. Such discoveries will inspire new simple control schemes via offloading complex control to such interactions, in effect better coupling the robot to the environment.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 04, 2021
- Source ID
- W911NF2110033
Entities
People
- Daniel Goldman
Organizations
- Army Contracting Command
- Georgia Tech Research Corporation
- United States Army