Reliability of Radar Inspection for Detection of Pavement Damage

Abstract

ABSTRACT Ground Penetrating Radar based analysis of road pavement is actually performed by many agencies involved in road management. The radar image interpretation is the most difficult step for the researcher to overcome. Software assisted processes have usually such a low reliability, that a road pavement engineer advisor is needed. The main aim of this paper is to assess the reliability of an optimal signal processing algorithm for pavement inspection. The algorithm is based on signal computing time delays. A threshold analysis of the error is carried out to check any damage and, subsequently, if damage is predicted, a second threshold analysis, subsequently points out its expected nature. This paper examines the experimental survey that has been performed both in laboratory and on road. The results show a good performance of the numerical algorithm and electronic equipment assessing the reliability of the procedure. It is also implemented the Neyman-Pearson radar test. The Receiver's Operating Characteristic is computed in different numerical configurations. The proposed system demonstrates a significant performance as the probability of false alarm is always lower than 20 per cent. In this sense the methodological approach can be reasonably considered to be reliable. In conclusion the experimental outcomes show that after calibration the proposed procedure makes it possible to diagnose many damages of flexible pavement. Moreover the class of damage respect to the three main classes, as defined in the paper, can be identified reliably.