Chaotic Oscillator Prediction Based on Artificial Neural Network and its Realization on FPGA

Abstract

Chaotic oscillators are widely used in applications due to their unpredictability. These applications include cryptography, true random number generators, and secure communication. On the other hand, artificial Neural Networks (ANNs) are known to be used in modeling and predicting the output of chaotic oscillators. This paper proposes a single switch Jerk chaotic oscillator prediction based on ANN and implemented on FPGA. One-step ahead prediction is implemented on FPGA using a feedforward neural net-work(FFNN). The ANNs performance is measured using Mean Square Error (MSE) and Root Mean Square Error (RMSE) and reaching 5.50E-6 and 2.3452E-3 for MSE and RMSE respectively in one-step prediction. The FFNN is implemented on FPGA based on a 32-bit fixed-point format and using an approximation hyperbolic tangent function by Piece-wise-Linear (PWL) achieving 88.11 MHz.