This project developed a method using active microwave sensing technique to measure moisture, snow and ice accumulation on rural highways (Figure 1). The system uses a low-power microwave transmitter and incorporates neural network and pattern recognition techniques for assessing road surface conditions. The basic system was designed, built and, after laboratory testing, installed at an outdoor location to collect data. Pattern recognition techniques were applied to the data to identify road conditions based on microwave signatures, and yielded 80-90% accuracy in detecting ice, snow, wet and dry road conditions. The classifier’s accuracy was improved to over 95% by using a neural network technique. Several configuration modifications were made to the system to improve its performance. Field test of the system were conducted in cooperation with the Wyoming DOT during the 1997-98 winter season. Several companies have expressed interest in collaborating in commercializing the technology. However, additional design optimization and field tests are need to implement this technology. The project has received media attention through regional newspaper articles, TV and radio segments, and also has been described in journal articles including the October 1997 issue of Popular Science. The final report is available from the National Technical Information Service (NTIS # PB98-141187).

The final report for this IDEA project can be found at:
https://onlinepubs.trb.org/onlinepubs/archive/studies/idea/finalreports/highway/NCHRP031_Final_Report.pdf