Modern Trends in Non-Destructive Soil Compaction Assessment with a Novel Radio Wave–based Experimental Framework

Assessment of the geotechnical and pavement structures are to be done as regards to the long-term performance of the soil compaction. The conventional methods of destruction are traditional like core cutters test and sand cones test which are time lasting, and point based. This essay talks about the design and the laboratory experiments of a new radio wave-based non-destructive testing (NDT) system in conjunction with machine learning techniques in predicting soil bulk density. Experiments on the laboratory conditions were conducted based on varying degrees of compaction and moisture content on prepared soil samples. Sensors parameters like Received Signal Strength Indicator (RSSI) and Signalto Noise Ratio (SNR) results were compared to laboratory bulk density data. The model of multiple regression and machine learning were implemented and compared in the measures of R² , MAE and RMSE. Random Forest model gave R² = 0.9046, RMSE = 0.0404 g/cc and MAE = 0.0308 g/cc which indicates that there is huge ability to predict and even generalize. The proposed system will enable real-time, cost-efficient, and non-destructive system of the soil compaction and enable intelligent quality control of the construction process.