In this paper, we study the effect of the numerical solution accuracy on the digital implementation of differential chaos generators. Four systems are built on a Xilinx Virtex 4 FPGA using Euler, mid-point, and Runge-Kutta fourth order techniques. The twelve implementations are compared based on the FPGA used area, maximum throughput, maximum Lyapunov exponent, and autocorrelation confidence region. Based on circuit performance and the chaotic response of the different implementations, it was found that less complicated numerical solution has better chaotic response and higher throughput. © 2011 IEEE.
Zidan M.A., Radwan A.G., Salama K.N.
Chaos generator; Chaotic generators; Chaotic response; Circuit performance; Confidence region; Digital implementation; Fourth order; Maximum Lyapunov exponent; Maximum through-put; Numerical solution; Numerical techniques; Runge-Kutta; Differential equations; Microelectronics; Runge Kutta methods; Lyapunov methods
Proceedings of the International Conference on Microelectronics, ICM, Art. No. 6177395, Doi: 10.1109/ICM.2011.6177395