Chaotic Oscillator for a Chaotic Oscillatory Neural Network Hardware Implementation

Abstract

Neural networks have opened up a wide range of new abilities. However, increasing complexity narrows the range of application due to calculation time and power consumption growth. That is the reason why the world famous companies are starting to develop a hardware implementation for different types of neural networks. This article presents the results of creating models for hardware implementation of the chaotic oscillatory neural network. The hardware implementation needs small, fast, energy efficient, and tunable circuit for one-dimensional chaotic oscillator. This article proposes one-dimensional chaotic oscillator circuit with improved characteristics that is completely suitable for usage in chaotic oscillatory neural network hardware implementation. Using the proposed oscillator, the model of chaotic oscillatory neural network has been developed and tested.