DEVELOPMENT OF A MOBILE COMPLEX FOR MAPPING SOIL COMPOSITION USING GPS AND ARTIFICIAL INTELLIGENCE

This article provides an extensive analytical overview of the scientific, methodological and technological foundations for developing an automated mobile system designed for high-precision soil composition mapping using GPS navigation and artificial intelligence. The relevance of this research is particularly significant for Kazakhstan, where soil conditions vary widely across regions – from chernozems to chestnut soils, saline soils and degraded landscapes. Such diversity requires advanced digital tools capable of performing rapid and accurate soil assessment.The literature review reveals the strong potential of VIS-NIR spectroscopy, mathematical modelling techniques, machine learning algorithms and geoinformation systems for improving soil diagnostics. Emphasis is placed on evaluating model performance indicators such as the coefficient of determination (R²) and root mean square error (RMSE), as well as the role of RTK-GPS in achieving centimeter-level geospatial accuracy. The integration of AI-enabled spectral analysis with GPS-based georeferencing ensures the generation of high-resolution soil maps (3-7 m), allowing precise identification of nitrogen, phosphorus, potassium, organic carbon and other vital soil attributes.Furthermore, automated mobile platforms significantly reduce time and labor costs, increase analytical efficiency and provide valuable decision-making tools for sustainable agriculture. They support targeted fertilizer application, soil fertility management, monitoring of erosion and salinization processes and assessment of agroecological risks. The findings confirm that implementing AI- and GPS-integrated mobile systems in Kazakhstan can substantially enhance soil monitoring infrastructure, optimize agricultural resource management and strengthen long-term environmental sustainability.