High‐Capacity Corner State Encoding and Dual‐Plane Switching Encryption Based on Topological Photonic Crystals

This study introduces a tunable high‐capacity corner state encoding photonic crystal structure designed for Beyond‐5G (B5G) communication. By arranging photonic crystals with distinct topological properties into arrays, localized confinement of light at specific frequencies at the corners of the structure is achieved. The hybrid deep learning model integrating convolutional neural networks and long short‐term memory networks are employed to precisely predict photonic crystal array parameters and corner state encoding performance within the B5G band. A nested single‐frame photonic crystal array is constructed and manipulate the relative phases of four excitation sources to selectively strengthen or weaken localized states at each corner. Simulation and experiment results validate the effectiveness of the proposed high‐capacity corner state encoding method. Furthermore, a dual‐plane corner state encryption system significantly enhancing the efficiency is developed, security and practical applicability of B5G communication.