A combination of DE and SVM with feature selection for road icing forecast
CAR'10 Proceedings of the 2nd international Asia conference on Informatics in control, automation and robotics - Volume 2
| Hi-index | 0.00 |
During the period from late autumn to early spring, vast areas in North America, Western Europe, and many other countries experience frequent snow, sleet, ice, and frost. Such adverse weather conditions lead to dangerous driving conditions with consequential effects on road transportation in these areas. A numerical forecasting system is developed for automatic prediction of slippery road conditions at road station sites in northern Europe and North America. The system is based on a road conditions model forced by input from an operational atmospheric limited area model. Synoptic information on cloud cover and observations of temperature, humidity, water and ice on the road from the road station sites are taken into account in a sophisticated initialization procedure involving flux corrections for individual locations. The system is run initially at the Rhode Island University with promising results. Currently, new forecasts 3 h ahead are produced every 20 minutes for 14 road station sites.