Large Range Soil Moisture Sensing for Inhomogeneous Environments using Magnetic Induction Networks

Water resource has become one of the most precious resources in recent decades. Agriculture accounts for about 80% of the total water usage in US. There is a demanding need for efficient irrigation and water management systems built for sustainable water utilization in smart agriculture. Real time insitu soil moisture sensing is a vital part for smart agriculture. Traditional electromagnetic (EM) based soil moisture sensing relies on EM based wireless sensor or ground penetrating radar (GPR) system. Based on the receiving signal strength and delay, tomographic techniques are used to derive the dielectric parameters of the soil, which are then into soil moisture distribution using empirical model. However, the EM signal attenuate sharply during underground propagation because of high operating frequency and lossy medium. In order to counter the disadvantage for underground sensing, we propose a Magnetic Induction (MI) based large range soil moisture sensing scheme in inhomogeneous environments. Here, we present the topology of the sensing system and analyze the channel model. The sensing process is based on transformed model, the conductivity and permittivity distribution are derived using SIRT algorithm. Through COMSOL simulation and analytical results, our proposed soil moisture sensing method achieves a root mean square error (RMSE) of 0.06 m(3)/m(3) in 40 m 2D scale inhomogeneous environment range.