Design and FPAA simulation of multi-scroll attractors in a memristor-based Hopfield neural network

This paper presents a memristor-based Hopfield neural network (MHNN) designed to generate multi-scroll chaotic attractors. By employing a novel flux-controlled memristor model, the system exhibits complex nonlinear dynamics, including bifurcations, variations in Lyapunov exponents, and multi-scroll chaotic behavior. To evaluate the randomness and cryptographic suitability of the generated sequences, statistical tests from the NIST SP 800‐22 suite are applied, confirming their high unpredictability. Additionally, the proposed MHNN system is implemented using both discrete analog electronic circuitry and a Field Programmable Analog Array (FPAA) platform, thereby validating the theoretical findings through hardware realization. This hardware facilitates the evaluation of system feasibility and reliability while enabling the rapid prototyping of analog and mixed signal-circuits. These results demonstrate the potential of the MHNN for secure communications, neuromorphic applications, and hardware-based chaos generation.