Performance Evaluation of Traffic Speed Deflectometer Based on Virtual Standard Test Road

It is challenging to find actual standard test roads (ASTRs) evaluate the measuring performance of Traffic Speed Deflectometer (TSD), leading misjudgment its evaluation results. To address this issue, a method construct virtual (VSTRs) using in-service proposed. Firstly, principle TSD based on Laser-Doppler effect measure pavement deflection introduced, and shortcomings traditional methods are reviewed summarized. Secondly, an adaptive construction for VSTRs proposed K-means clustering algorithm. layers Falling Weight (FWD) measurements reconstructs them as with stable deflections but different ranges. Thirdly, two new indexes, coefficient variation $CV_{k}$ influence speed notation="LaTeX">$ICS_{k}$ , replace such repeatability correlation coefficient. Finally, feasibility evaluating indexes verified by experiments. Results show that newly constructed solve problem ASTRs obtain, reducing experimental requirements TSD, achieve efficient scientific TSD.