Large-scale Soil Moisture Monitoring: A New Approach

Soil moisture is a critical factor for understanding the interactions and feedback between the atmosphere and Earth's surface, particularly through energy and water cycles. It also plays a key role in land climate and hydrological processes. Recent advancements in autonomous robotic applications for precision agriculture have introduced significant solutions, particularly in remote sensing. Currently, these platforms enable autonomous soil parameter measurement and on-site data collection, which is essential for resource optimization and data-driven agricultural decision-making. However, challenges persist, especially in real-time soil moisture monitoring—a key focus for improving irrigation efficiency, water use, and crop yields. Soil moisture measurement in-situ techniques include the accurate oven-drying method and soil moisture sensors, while satellite remote sensing uses optical, thermal, and microwave imaging to estimate surface soil moisture from a broader perspective. However, fully autonomous robotised sampling procedures for optimising the process, increasing repeatability and overall accuracy, as well as increasing the reachability of the sampling of remote areas, are still not utilized.

Soil moisture measurement in-situ techniques include the accurate oven-drying method and soil moisture sensors, while satellite remote sensing uses optical, thermal, and microwave imaging to estimate surface soil moisture from a broader perspective. However, fully autonomous robotised sampling procedures for optimising the process, increasing repeatability and overall accuracy, as well as increasing the reachability of the sampling of remote areas, are still not utilized.

Measuring soil moisture presents a significant challenge due to its reliance on human labour, which is required to cover extensive areas for sensor measurements manually. Additionally, soil moisture measurements at a specific point vary with time and environmental conditions, making these values unstable. While satellites offer a potential solution to some of these issues, their accuracy is affected by environmental factors such as cloud cover and dense vegetation, while they only describe the upper soil layer. Moreover, ground measurements of surface soil moisture are still necessary for calibrating and training the satellite systems. To address these challenges, we propose an adaptable in situ method for automating soil moisture measurements.

Our approach introduces AgriOne, an autonomous soil moisture measurement robot leveraging a surface-aware data collection framework to achieve precise and efficient soil moisture assessments, thereby minimizing reliance on permanent sensors and reducing associated costs and labour. The hardware of AgriOne consists of a UGV Husky A200 from Clearpath Robotics loaded with the soil moisture sensor TEROS12 from Meter Group. The sensor is mounted on a linear actuator probe, as shown in the figure.  

To evaluate the proposed approach, we conducted two field experiments in different locations under different weather and soil conditions. The experiments were successful in both cases, and we collected 70 and 66 measurements, respectively, of surface soil moisture. For the first experiment, we covered an area of 380m² in 57 minutes, and for the second experiment, we covered an area of 800m² in 2,5 hours. The results showed proof of concept because of the workability of the AgriOne robot and the reliability of the data collection framework.