Informational active matter

Bryan VanSaders; Vincenzo Vitelli

doi:10.48550/arxiv.2302.07402

Many biomolecular systems can be viewed as ratchets that rectify environmental noise through measurements and information processing. As miniaturized robots cross the scale of unicellular organisms, on-board sensing and feedback open new possibilities for propulsion strategies that exploit fluctuations rather than fight them. Here, we study extended media in which many constituents display a feedback control loop between measurement of their microstates and the capability to bias their noise-induced transitions. We dub such many body systems informational active matter and show how information theoretic arguments and kinetic theory derivations yield their macroscopic properties starting from microscopic agent strategies. These include the ability to self-propel without applying work and to print patterns whose resolution improves as noise increases. We support our analytical results with extensive simulations of a fluid of `thinker' type particles that can selectively change their diameters to bias scattering transitions. This minimal model can be regarded as a non-equilibrium analogue of entropic elasticity that exemplifies the key property of this class of systems: self-propulsive forces grow ever stronger as environmental noise increases thanks to measurements and control actions undertaken by the microscopic constituents. We envision applications of our ideas ranging from noise induced patterning performed by collections of microrobots to reinforcement learning aided identification of migration strategies for collections of organisms that exploit turbulent flows or fluctuating chemotactic fields.

Informational active matter

Files and links (1)

Abstract

Metrics

Details

Informational active matter

Files and links (1)

Abstract

Metrics

Details

Drexel University Social media