Abstract
Precision agriculture demands continuous and accurate monitoring of soilmoisture (M) and key macronutrients, including nitrogen (N), phosphorus (P),and potassium (K), to optimize yields and conserve resources. Wireless soilsensing has been explored to measure these four components; however, currentsolutions require recalibration (i.e., retraining the data processing model) tohandle variations in soil texture, characterized by aluminosilicates (Al) andorganic carbon (C), limiting their practicality. To address this, we introduceSoilX, a calibration-free soil sensing system that jointly measures six keycomponents: {M, N, P, K, C, Al}. By explicitly modeling C and Al, SoilXeliminates texture- and carbon-dependent recalibration. SoilX incorporatesContrastive Cross-Component Learning (3CL), with two customized terms: theOrthogonality Regularizer and the Separation Loss, to effectively disentanglecross-component interference. Additionally, we design a novel tetrahedralantenna array with an antenna-switching mechanism, which can robustly measuresoil dielectric permittivity independent of device placement. Extensiveexperiments demonstrate that SoilX reduces estimation errors by 23.8% to 31.5%over baselines and generalizes well to unseen fields.