InTrans / Aug 30, 2021
Aurora Program research projects consider optimal RWIS density and location
A series of research projects to determine the optimal quantity and placement of road weather information systems (RWISs) led by Tae J. Kwon, an assistant professor at the University of Alberta, has recently concluded.
The Aurora Program-sponsored Phase III project developed a systematic, yet transferrable, method for estimating key road surface condition variables between RWIS stations using large-scale data and advanced modeling techniques such as Geostatistics for spatial inference and mapping and Deep Learning for image recognition.
The projects culminated in techniques that transportation agencies can use to expand their road surface condition (RSC) spatial coverage substantially and enhance their ability to perform winter road maintenance activities, ultimately providing the general public with a greater level of service in terms of safety and mobility.
This is critical as RWIS technologies are relatively expensive to maintain and operate and are therefore only installed at a limited number of locations to provide real-time and near-future surface condition information to make timely maintenance related decisions.
The limited number of RWIS stations along with the need to monitor spatially large road networks with vastly varied conditions necessitate a strategic and scientific approach to the continuous and accurate monitoring of RSCs during inclement weather events.
The project also sought to automate the process of image recognition to fill in the spatial gap of unmonitored areas using RWIS and other mobile sensing technologies (e.g., maintenance fleet dash cameras and automatic vehicle locations). In particular, because most RWIS stations have cameras that provide a direct view of surface conditions but require users to manually view the images, the ability to automate the process would allow agencies to use the data more effectively and improve the level of service they provide.
Future phases of the project are planned to further validate the methodology and improve image recognition by covering a larger range of road, weather, environmental conditions, as well as testing alternative deep learning models for real-time implementation. These efforts are expected to provide winter maintenance personnel with newfound knowledge and analytical tools required to make better use of available resources, resulting in better maintenance and improvements to their highway infrastructure thereby promoting improved winter mobility and safety.
To learn more about the projects, go here.