NGnet-Based Topology Optimization of a 1-Bit Reconfigurable Intelligent Surface (RIS) Unit Cell

Abstract

This paper presents a topology optimization approach for the design of a 1-bit reconfigurable intelligent surface (RIS) unit cell based on a Normalized Gaussian Network (NGnet) combined with Particle Swarm Optimization (PSO). The unit cell operates in the 2.4 GHz Wi-Fi band and employs a single PIN diode to realize two reflection states. The design objective is to achieve a phase difference close to ${180}^{{\circ}}$ between the two states while maintaining high reflection magnitude. A symmetric configuration is imposed to place the switching element at the center of the unit cell, reducing the number of design variables. The optimized design exhibits stable phase behavior and consistent reflection performance over the operating frequency band, demonstrating the effectiveness of the proposed approach for RIS unit cell design.