New integrated autonomous solution for underground monitoring and mapping

Soil contamination monitoring and precision nutrient assessment require measurement systems capable of operating directly in subsurface environments while ensuring reliable spatial referencing of collected data. This paper presents a novel integrated autonomous solution for underground monitoring and mapping, combining in-line chemical sensing with magnetic-field-based localization in a unified architecture designed for deployment on drilling or burrowing robotic platforms. The proposed system integrates: (i) a near-infrared (NIR) spectroscopy module coupled with a convolutional neural network (CNN) for the estimation of urea–nitrogen content; (ii) a solid-state ion-selective electrochemical sensing unit for selective potassium detection in soil-derived aqueous extracts; and (iii) an AC magnetic localization subsystem based on surface transmitting anchors and a triaxial underground receiver. The architecture is designed to ensure spatio-temporal integration between measurement and positioning, enabling spatially consistent subsurface chemical mapping in GNSS-denied environments. The NIR–CNN approach demonstrated reliable urea-N estimation (R2=0.79, RMSE = 3.57%), while the electrochemical subsystem provided selective and repeatable potassium measurements with a more repeatable potentiometric response in the additively manufactured sensors. The magnetic localization algorithm achieved sub-decimeter accuracy in simulation (RMSE ≈8.8 cm when the anchors’ position is known with negligible accuracy) and sub-meter performance (RMSE ≈40 cm, in presence of anchors’ positioning errors) in outdoor experimental validation, confirming the feasibility of real-time underground positioning within operational constraints of drilling platforms. The results demonstrate the effectiveness of the proposed integrated sensing and localization framework for subsurface environmental monitoring, supporting future applications in precision agriculture, soil contamination assessment, and autonomous underground exploration.