Vienna, December 2019:
In the research project, a forecasting tool for real-time predictions of road surface temperature and hoarfrost integrating measurement data and sensors as well as weather forecasts for the optimization of winter service and traffic safety at ASFINAG is being developed.
Projec title
Development of a forecast tool for real-time predictions of road temperature and hoarfrost events by integrating measurement data / sensors and weather forecasts with thermal fingerprint & AI.
Customer
ASFINAG Service GmbH
Type
Research collaboration
Duration
May 2019 to October 2020
Partner/Role
HOFFMANN CONSULT with Alexander Haberl und Valentin Donev with project conception and scientific project management and implementation.
Project goals
The reliable prediction of road temperature and hoarfrost is essential for winter maintenance planning, preventive application with efficient deicing agent use resulting in high availability and safety on motorways and expressways in Austria. Implementation of a prediciton tool for real-time forecasts of road surface temperature and hoarfrost events by integrating measurement data / sensors, weather forecasts in an administrator interface as well as integration into the IT environment of ASFINAG and their weather portal Wetter 2.0.
Results
Model development, integration and implementation of high-precision forecast models with road temperature and hoarfrost prediction at 250 weather sections, downscaling of the forecasts to the entire network of 2,200 km and implementation of the Admin GUI for the functional control and optimization of the sensors in the existing network as well as the forecasts. Among others the following methods have been used:
Regression models, neural networks
Artificial intelligence and ensemble learning
Application design and development
Comprehensive statistical analysis
Sensor tests and result check
Function tests and forecast validation
The result is more efficient winter maintenance and increased traffic safety.