Counteracting dynamic degradation with an N-dimensional time-varying delay coupled chaotic map

Abstract To address the performance degradation of chaotic systems under finite-precision constraints—a critical challenge in the design of secure and efficient pseudo-random number generators (PRNGs)—this paper proposes a novel delay-coupled chaotic framework that enhances dynamical complexity while preserving computational efficiency. Unlike many existing methods that suffer from short pseudo-periods and cyclic behavior due to quantization, our approach introduces time-varying delays and nonlinear perturbations into one-dimensional or high-dimensional chaotic maps. This strategy effectively suppresses dynamical degradation and extends the usable chaotic horizon. Theoretical analysis and extensive experiments demonstrate that the proposed model achieves longer periods, richer randomness, and superior robustness under low-precision conditions. We also introduce a new performance metric (‘Period/Time’) to evaluate computational efficiency, showing our method outperforms benchmark chaotic PRNGs in both speed and entropy. A scalable high-dimensional implementation based on a 3D-LST system confirms the method’s generality and practicality. This study contributes a structurally simple yet powerful solution to the design of quantization-resilient chaotic systems for secure communications and embedded cryptography.