AI-Enhanced 2-D-Material Terahertz Sensor for 6G IoT-Based Smart Agricultural Monitoring

Abstract

This article presents the design and characterization of a planar terahertz (THz) gas sensor based on three cascaded spiral resonators functionalized with advanced 2-D materials, namely graphene, black phosphorus (BPs), and vanadium dioxide (VO2). The use of these materials significantly enhances the sensor sensitivity and enables tunable multigas detection capabilities. The electromagnetic response of the proposed structure in the presence of target gas analytes is accurately modeled using the wave concept iterative process (WCIP), with simulation results showing excellent agreement with theoretical analysis. To further improve the sensor performance, a deep neural network (DNN) model is developed to predict optimal sensor configurations by learning the influence of key material parameters associated with graphene, BP, and VO2. The proposed DNN demonstrates high predictive accuracy, achieving an R2 score of 0.995, a mean absolute error (MAE) of 0.085, and a mean squared error (mse) of 0.010 for ammonia (NH3) using an 80% training dataset. The sensor exhibits high sensitivities of 13.87, 14.05, 14.21, 14.12, and 14 THz/RIU for methane (CH4), nitrous oxide (N2O), NH3, carbon dioxide (CO2), and hydrogen sulfide (H2S), respectively, while operating at room temperature with rapid response characteristics. Furthermore, a conceptual integration of the proposed THz sensor within a 6G-enabled internet of things (IoT) smart agriculture framework is discussed, highlighting its potential for real-time environmental monitoring and precision farming applications. These results demonstrate the strong potential of combining advanced 2-D materials, accurate electromagnetic modeling, and AI-driven optimization for the development of next-generation THz gas sensing platforms.