Intelligent automation of road pavement compaction processes using satellite navigation and sensor technologies

Abstract

Density requirements for structural layers of pavement and subgrade are formulated in Russian regulatory documents as a continuous characteristic of the whole layer volume, whereas verification relies on spot sampling covering fractions of a percent of the working section area. The mismatch between the object of regulation and the object of measurement is removed by continuous recording of the compacted medium response through onboard roller sensors, with each reading referenced in space by satellite navigation means. The study quantifies which parameters of the navigation and measurement chain govern the reliability of a continuous compaction map and to what extent the transition to such a map changes the decision to stop compaction. Calculations were performed on an author's model of a working section of 3,680 square metres rasterised with a 0,5 metre cell, reproducing drum lane patterns, lane guidance error and measurement channel noise. The calibration relationship between the compaction coefficient and the stiffness index was derived from seven control points of a test strip and describes the sample with a determination coefficient of 0.9751 at a residual standard deviation of 0.00372. The stiffness saturation criterion with a 5% increment threshold is triggered at the eighth pass, when the required coefficient of 0.98 has been reached over 87,17% of the area against a reference value of 90%; the acceptance threshold of 70% of the target index corresponds to a compaction coefficient of 0.9628 and should be raised to 82,5% for Russian regulation. The shift from autonomous satellite positioning to real-time kinematic mode reduces the share of area receiving fewer passes than designed from 40,49% to zero, and the under-compacted area from 813,7 to 430,3 square metres. Variance decomposition attributes 87,63% of the compaction coefficient spread to soil moisture heterogeneity and 10,63% to the edge zone. The reliability of a continuous compaction map is governed by machine lane guidance accuracy and navigation solution rate, while the resolution of the stiffness sensor remains redundant; the benefit of continuous control lies in eliminating the probabilistic nature of acceptance, since the reduction in roller work is estimated at about four percent.