Accuracy of ground-penetrating radar for estimating rigid and flexible pavement layer thicknesses

In this paper, the accuracy of ground-penetrating radar (GPR) for estimating pavement layer thicknesses is studied on the basis of the investigation of 17 pavement sites in Virginia. The considered sites have different types of pavement systems (flexible, continuously reinforced, jointed concretes, and composite) and different ages (0 to 5 years; 10 to 15 years; older than 20 years with a surface less than 10 years; and older than 20 years with a surface older than 10 years). Because of the diversity of the test sections considered, the accuracy of the GPR thicknesses was studied for pavement age for the same type of pavement and against pavement type for sites of the same age category. For flexible pavements, the GPR thickness error was found to increase as the pavement's age increased (4.4% error for pavements 0 to 5 years old versus 5.8% error for pavements older than 20 years with surfaces older than 10 years). Moreover, for the same age category, flexible pavements were found to have a relatively high thickness error (4.4%) compared with the continuously reinforced concrete pavements (3.0%) and with the jointed plain concrete pavements (2.3%) because of the relative homogeneity of concrete for electromagnetic wavelength when compared with the different dielectric profile of aged hot-mix asphalt layer.