Chaos synchronization using photonic reservoir computing based on a single delayed semiconductor laser

Chaos synchronization has attracted significant interest as a promising technique for secure communications. However, its practical implementation is fundamentally limited by the dependence on precise parameter matching between the transmitter and receiver. In this work, we propose an approach based on photonic reservoir computing (PRC) with a single nonlinear dynamical node to overcome this limitation by learning the complex nonlinear dynamics of the chaotic carrier directly from the transmitter. Once trained, the PRC enables accurate reconstruction of the chaotic carrier at the receiver side without parameter matching. The performance of chaos synchronization is systematically investigated over a wide parameter space, and the optimal operating point of the PRC is identified and analyzed from a physical perspective. Under optimal conditions, high-quality chaos synchronization is achieved, with a maximum correlation coefficient of 0.9969 between the reconstructed and original chaotic carriers. Based on this performance, secure transmission of the quadrature amplitude modulation-16 (16 QAM) message is successfully demonstrated, validating the proposed scheme as an effective and practical solution for secure communication.