查看文章

Transportation agencies use a variety of methods to evaluate the condition of pavements and the characteristics of their use. Methods range from in-pavement sensors to ride quality characterizations that require specially instrumented probe vehicles. However, the extensive labor, specialized training, and high cost associated with the large variety of existing approaches limit their ability to scale in frequency and coverage. Infrequent monitoring diminishes the accuracy of assessing infrastructure needs and of evaluating maintenance effectiveness. This research developed a sensing approach that extends the capability of in-pavement sensors beyond their ability to measure just loading and condition parameters. Specifically, the approach links the output of durable in-pavement strain sensors to common roughness indices. However, to maintain their accuracy throughout the life cycle of the pavement, models that use their output must be calibrated periodically. Therefore, this research also developed a localized roughness measurement method based on connected vehicle sensing to calibrate the models. Field experiments validated that the relative roughness indices of the two methods agreed within 3.3%. This result demonstrates that it is possible to build smart roads with embedded strain sensors that can also report roughness levels continuously, without requiring special probe vehicles.