Design, development, and field evaluation of a dual-mode robotic-human system for in situ soil compaction assessment in agricultural fields

• Developed a high-throughput robotic penetrometer for reliable in-field compaction mapping. • Enables precision management, including variable-rate tillage depth to protect plant growth. • Benchmarked against the FieldScout SC900, showing significant depth-mean agreement (R 2 ≈0.97; RMSE≈103 kPa (15 PSI)). • Fills a key measurement gap by quantifying compaction in highly tilled, heterogeneous fields. • End-to-end pipeline enables site-specific remediation i.e. Controlled traffic farming lanes and subsoiling. Soil compaction is a major constraint in mechanized agriculture. Its accurate measurement is essential for understanding machinery impacts, optimizing crop growth, and guiding soil management practices. However, its strong spatial and depth-wise variability makes it difficult to characterize with traditional hand-held penetrometers alone. This study aimed to develop and evaluate a robotic penetrometer system for field-scale compaction mapping in sandy-loam potato–forage rotations on Prince Edward Island (PEI), Canada, over two consecutive growing seasons (2024 and 2025) at six depths. In 2024, we collected co-located penetration-resistance profiles with a FieldScout SC900 and the robotic penetrometer in electrical-conductivity (EC) zones derived from SWAT Box maps in two potato fields. In 2025, after calibration, the robot was deployed on a regular grid in two alfalfa fields that had grown potatoes in the previous year. In total, depth-wise mean profiles from the two instruments showed excellent agreement (R 2 ≈0.97), while point-wise paired measurements exhibited moderate but positive correlation (R 2 ≈0.60), reflecting micro-scale heterogeneity and differences in insertion mechanics. Spatial results revealed low to moderate compaction with localized hotspots in the 2024 potato fields, but extensive, vertically coherent high-resistance zones in the 2025 alfalfa fields, indicating persistent subsoil compaction following intensive potato traffic and limited subsequent tillage. The study demonstrates that a calibrated robotic penetrometer can reproduce conventional measurements while providing higher spatial resolution, deeper and more consistent penetration, and seamless integration with geospatial analysis. These capabilities support controlled-traffic farming, targeted subsoiling, and data-driven decision support for managing soil structure in traffic-sensitive cropping systems.