Table 1.

Description of reservoir computing components with application within the EAP gel network

Component	Description	Application to Gel
Excitation layer	Input to the reservoir, converting from defined input symbols to stimulations in the reservoir network.	Inputs are defined as −1,1 and are applied via electrical stimulation across the width of the gel. This is the same system as the excitation layer used in the Moore Machine framework.
Reservoir	Fixed, non-linear system that is used to map inputs to higher dimensional space.	EAP hydrogel suspended in the ionic solution via surface electrodes, stored as a virtual database through the PMA.
Readout layer	Output from the reservoir. Converts from the reservoir network to defined output symbols. This layer is tuned to train the reservoir computing network.	Bending angle of the gel as viewed from the side perpendicular to the electric field, binarized (thresholded) into the defined output symbols of −1,0,1. Tuning is performed through the optimization of thresholds.