Discrete-time neural synchronization between an Arduino microcontroller and a Compact Development System using multiscroll chaotic signals

Abstract

In this paper, we present the synchronization of a chaotic system using a discrete-time recurrent high order neural network. This is done by using a Genesio & Tesi oscillator circuit in discrete-time embedded into an Arduino microcontroller that provides the state space variables. A discrete-time recurrent neural network is designed to synchronize the dynamics of the chaotic oscillator. This neural network is trained using a time-varying training algorithm where it is used the Extended Kalman Filter. Two state space variables are captured in real-time in ADC inputs of a compact development system, where these signals are synchronized by the recurrent high order neural network in discrete-time. The proposed work allows synchronization of interactions associated between the neural convergence and the chaotical plant state. The obtained real-time results, and the statistical analyses on the synchronization process validate the possible application in chaos-based communications systems.