Chaos Robustness and Computation Complexity of Piecewise Linear and Smooth Chaotic Chua's System

Abstract

Chaotic systems are often considered to be a basis for various cryptographic methods due to their properties, which correspond to cryptographic properties like confusion, diffusion and algorithm (attack) complexity. In these kinds of applications, chaos robustness is desired. It can be defined by the absence of periodic windows and coexisting attractors in some neighborhoods of the parameter space. On the other hand, when used as a basis for neuromorphic modeling, chaos robustness is to be avoided, and the edge-of-chaos regime is needed. This paper analyses the robustness and edge-of-chaos for Chua’s systems, comprising either a piecewise linear or a smooth function nonlinearity, using a novel figure of merit based on correlation coefficient and Lyapunov exponent. Calculation complexity, which is important when a chaotic system is implemented, is evaluated for double and decimal data types, where needed calculation time varies by a factor of about 1500, depending on the nonlinearity function and the data type. On the other hand, different data types result in different number precision, which has some practical advantages and drawbacks presented in the paper.