Proteinoid-polyaniline neuromorphic composites for audio recognition

Abstract

We present an innovative neuromorphic system using a proteinoid-polyaniline (PANI) composite for recognition of audio inputs of the English alphabet. Neuromorphic devices, which draw inspiration from the neural networks of the brain, have emerged as very promising potential solutions for efficient signal processing. The proteinoid-PANI composite was synthesized through a template-free method, resulting in a unique nanostructure consisting of both nanorods and nanospheres. Principal component analysis (PCA), spectrogram analysis, and temporal spiking response analysis were among the signal processing methods used to examine the composite's audio response to English alphabet stimuli. The system showed a moderate positive correlation between input and output signals, unique time-frequency response patterns, and convoluted spiking behaviour. In addition, the output amplitude showed less variation compared to the input, while maintaining the same temporal characteristics.
Microscopic analysis provided detailed information about the morphology of the composite. The nanorods displayed an optimal aspect ratio and had diameters of around 
100~nm, while the nanospheres varied in size, ranging from 200 to 500~nm in diameter. The nanostructure, morphological characteristics, and signal processing properties of the proteinoid-PANI composite demonstrate its potential for advanced applications in neuromorphic computing and signal processing, particularly in speech recognition and human-machine interaction.