MagTile: A Modular Platform for Controlling Electromagnetically Actuated Agents for Ethorobotics Applications

Abstract

Ethorobotics, the study of animal–robot interactions, has advanced behavioral research by enabling controlled, repeatable experiments through the use of biomimetic robotic agents. Existing platforms, however, often lack flexibility, scalability, and computational tools for in-silico experimentation, limiting their easy deployment across diverse experimental configurations. Here, we introduce MagTile, a modular and scalable platform for controlling ensembles of electromagnetically actuated agents. Our design utilizes standard components and 3-D-printed parts, enabling multiagent actuation without the need for complex onboard electronics. We formulate a mathematical model of MagTile and analyze its convergence properties using Lyapunov stability theory. The model serves as the foundation for establishing a computational framework, facilitating simulation, testing, and optimization of experimental conditions prior to physical deployment, thereby improving efficiency and reducing trial-and-error experimentation. The effectiveness of our platform was validated through a series of illustrative experiments showcasing multiagent coordination, vision-based collision avoidance, and collective behavior using biomimetic fish.