Artificial Spacetimes that Control Robots.

From waveguides to cloaks, physicists possess incredible tools to manipulate the path of light. Here, we show these same concepts can be directly repurposed to control and understand robots. A formal connection between the two arises whenever a robot controls its speed using position alone. Like a photon, the robot’s future position must live on a local ‘light-cone’, effectively defining an ‘artificial spacetime’. This surprising yet formal link allows us to port techniques from general relativity, differential geometry, and optics to robotics, unlocking new ways for robots to rally, navigate, patrol, disperse, turn, or avoid obstacles like they were cloaked. Further, these schemes work in both simulation and the lab, which we demonstrate with experiments on silicon-based microrobots. More broadly, we argue the geometric foundations of modern physics offer powerful new ideas for robotics that are both mathematically grounded and computationally lightweight.