Harnessing swarm behaviours of active particles to build intelligent microrobots

Swarm intelligence governs the coordinated behaviour of interacting entities. In this work, we aim to employ the collective intelligence of confined microparticles to build highly deformable microrobots. Confinement- and interaction-driven dynamics are crucial for our purpose, as the microparticles are encapsulated within a soft membrane. To explore mechanical sensing, we performed proof-of-concept experiments by confining Janus microparticles inside a rigid, elliptical polydimethylsiloxane (PDMS) well. This controlled setup allowed us to study particle-boundary interactions. We observed a tendency for particles to accumulate in the high-curvature region, indicating their ability to sense and respond to curvature. These particles prefer to stay and move along the boundaries, thus giving rise to particle number fluctuations at high-curvature zones.

Furthermore, we comprehended these findings with numerical simulations (in Julia) of active Brownian particles in 2D elliptical confinements with steric interactions among the particles. Moreover, incorporating various interaction potentials and no confinement, we observed phenomena such as cluster formation and flocking. These dynamics depend on particle velocity, size, and packing fraction. We anticipate that our results may be applicable to the design of particle-based microrobots, where environment-induced membrane deformations could guide the movement of an encapsulated active particle swarm.