Physical embodiment enables information processing beyond explicit flow sensing in active matter.

Area of Science:

• Active matter physics
• Bio-inspired computation
• Robotics

Background:

• Microorganisms use sensory systems to react to their environment.
• Replicating sensing and behavioral adaptation in synthetic matter is challenging.
• Current synthetic active matter lacks sophisticated environmental response.

Purpose of the Study:

• To investigate if synthetic active particles can adapt to unobserved environmental changes.
• To explore the role of physical embodiment in information processing for active matter.
• To demonstrate implicit sensing through physical dynamics.

Main Methods:

• Utilized self-thermophoretic particles.
• Employed reinforcement learning for particle control.
• Trained particles to navigate unobserved hydrodynamic perturbations.

Main Results:

• Particles successfully adapted navigation strategies to counteract hidden flow fields.
• Demonstrated that physical embodiment alone enables adaptation without explicit sensing.
• Showcased embodied dynamics as an implicit sensing mechanism.

Conclusions:

• Physical embodiment serves as a computational resource for active matter.
• This approach has implications for autonomous microrobotic systems.
• Establishes a new paradigm for bio-inspired computation and synthetic intelligence.