Abstract/Summary

Active matter systems have long been explored for computational potential, utilising their unique behaviours for esoteric applications. Electroactive Polymer (EAP) hydrogels are an active matter material currently explored as soft robotic actuators, however not yet explored for their potential computational potential as a form of active matter. Interesting alternate computing mediums exist but studies have yet to present significant framework for materials integrating computation with physical action. This study exploits EAP responses, finding that EAP behaviour can be utilised for automaton and reservoir computing frameworks. Through reproduction of previous EAP works memory-like behaviour was found within the EAPs, confirming their potential as computational resources, given appropriate framework. Under sequential electrical stimulation, the EAP mechanical responses were represented in a probabilistic Moore automaton framework and expanded through shaping the reservoir’s energy landscape. The EAP automaton reservoir’s computational ability was compared with digital computation to assess EAPs as computational resources, showing that the EAP’s computation via reaction to stimuli can be presented through automaton structures, revealing a potential bridge between and controller in EAP’s. This computation was then further expanded through additional inputs. Through this exploration similarities in behaviour between EAP hydrogels and biological neurons were discovered, among these were evidence of the Free Energy Principle (FEP), raising the question; by applying theories of Biological Neural Network (BNN) learning, such as FEP active inference, to learning within a different medium whose behaviour is also governed by FEP, can emergent learning be achieved with alternative mediums? EAP hydrogels were embedded in the simulated game-world of Pong via custom Multi-Electrode Arrays and feedback between motor commands and stimulation. Through performance analysis within the game environment emergent learning was observed, driven by ionic behaviours of the hydrogels. These observations enforced the theories of FEP within learning and its ability to allow learning in other mediums.