A magnetic induction network for high-resolution, real-time soil moisture monitoring in complex subsurface environments

• A novel MI network enables high-resolution, real-time soil moisture sensing. • Multi-frequency phase unwrapping enhances accuracy and reduces phase ambiguities. • COMSOL simulations show a 25–50 % improvement in sensing accuracy over 40 m. • Indoor tests reveal a 38 % improvement in accuracy compared to traditional methods. • MI network offers scalable, real-time precision agriculture solutions. Water resource management in smart agriculture demands accurate, real-time soil moisture monitoring across large and heterogeneous fields. However, existing sensing technologies, such as time-domain reflectometry (TDR), capacitance probes, ground-penetrating radar (GPR), and single-link wireless underground systems, suffer from limited spatial coverage, reduced performance in conductive soils, and an inability to resolve heterogeneous moisture patterns in real time. This study aims to develop and validate a scalable magnetic induction (MI) network–based sensing system that enables real-time, high-resolution volumetric soil moisture mapping across large and heterogeneous agricultural fields, thereby overcoming the limited coverage and instability of conventional underground sensing technologies. In particular, the system combines three main elements: (i) a distributed magnetic induction (MI) networking architecture designed for reliable underground communication; (ii) a channel frequency response (CFR)-based protocol that supports real-time multi-link operations; and (iii) a multi-frequency phase unwrapping algorithm that ensures precise phase reconstruction across diverse soil conditions. Comparative evaluation through COMSOL Multiphysics full-wave simulations shows that the system reduces average sensing errors by 25–50 % over a 40 m range compared with conventional wireless underground sensing. Laboratory testbed experiments confirm a 38 % accuracy improvement over state-of-the-art approaches, achieving a spatial RMSE of ∼0.02 m 3 /m 3 . This work is the first experimental demonstration of using MI networking for 3D real-time soil moisture imaging, offering a robust and scalable solution for precision agriculture. By enabling distributed, non-invasive subsurface monitoring, the system has the potential to improve irrigation scheduling and enhance water-use efficiency by an estimated 15–25 % in large-scale farming operations.