Novel Pattern Detection Algorithm for Monitoring Phase Change of Moisture on Concrete Pavement Using Surface Temperature Data

Abstract

The presence of snow and ice on the pavement are major winter hazards for roadway traffic safety. Pavement temperature sensors are usually employed as a standard road weather element in many road weather information systems. A novel pattern detection algorithm has been developed to monitor the phase change of moisture on the pavement surface by analyzing a single surface temperature reading. The results of both indoor weather chamber tests and field tests under realistic highway in-service traffic conditions demonstrate that the event of moisture phase changes (e.g., dry, wet, snowy, and icy states) can be detected by converting a complicated raw sensor time history signal into a simpler binary “spike train” time history signal. These spikes can be tuned by the user to detect only the events of interest that are usually entangled with various known and unknown ambient trends in field sensor data. Since the procedures of the developed data processing algorithm are data driven and not limited to specific sensor types or physical problems, the methodology used in this study potentially can be applied widely in many field-monitoring applications.