A Study of Road Surface Ice Prediction Based on Highway Operations and Safety

To improve the high-risk situation of driving on icy road in cold weather, a more economy applicable measure for road ice forecasting is proposed compared with the Road Weather Information System widely used abroad. A temperature-controlled laboratory with a short section of road was used to simulate the scene driving on the highway. According to the changes of temperature and precipitation, the corresponding changes of temperature close to the surface, road surface temperature and road surface ice were recorded to show the rules of the road surface temperature changes and icing development. The main purpose was to build the road surface ice prediction model. After the network training of the experiment samples using a BP neural network, the road surface ice prediction model is put forward, especially for the ice prediction of some high-risk sections, including bridges, tunnels, shady sections etc. Besides, the method contributes to the development and application of low cost icy road prediction techniques.