AI for safer and more sustainable traffic
Published on 05/07/2022 in News
Proximus subsidiary BeMobile worked with 5 project partners on an AI & IoT use case to improve traffic safety in adverse weather conditions, reduce congestion, and save on fuel.
Celtic-Next Secure and Accurate Road Weather Services (SARWS) is the name of the European research initiative in which BeMobile participates, together with bpost , KMI (Belgian meteorological unit) , the innovation specialists Verhaert Masters in Innovation and Inuits and IMEC , the Leuven research center for nanoelectronics and digital technology. VLAIO , the Flemish Agency for Innovation and Entrepreneurship, provided financial support.
Sensors for temperature, humidity and location
For the project, a fleet of bpost vans was equipped by Verhaert with the developed sensors and on-board AI. The idea was to measure the temperature, humidity, and location, classify how hard it rains and snows, and the slipperiness of the road. Prediction data can be forwarded to the back-end (cloud). An infrared camera was placed in the nose of the vehicle to record the temperature of the road surface.
At the same time, a dashcam continuously records images of the driving conditions. The entire system is also connected to the vehicle itself via its internal Controller Area Network (CAN bus). For example, you can check whether the windscreen wipers are working, or whether the windscreen demister and fog lights are switched on.
Data processing in the car
The data that is captured is immediately (partially) processed in the actual car by an AI system. This uses, for example, the images from the dashcam to identify three weather types: rain, fog or snow. At the same time, it also corrects the effects of the changing speed of the vehicle on the measured results. There is a simple reason why the data undergoes an initial and instant analysis on the way: this guarantees privacy and also limits the data volume that is sent to Inuit's management platform.
Collaboration with KMI and bpost drivers
This platform gives the various partners access to all sensor data. For example, Be-Mobile processes it in its traffic information system and releases algorithms that can determine whether it is raining and where. From there, the data goes on to the KMI, which they use to improve the accuracy of their weather model of our road network and thus better predict the state of the road. Be-Mobile also sends all data back to the bpost drivers. In order to validate the findings and further improve the predictions, they can provide manual feedback via a mobile app.
Finer-Detailed Road Weather Maps
By using AI and algorithms to process the sensor data, all partners made an interesting scientific contribution. These data will also be used for future projects and will, for example, provide much more accurate and finely detailed road maps.
Be-Mobile leads the Mobilidata project for safer traffic in Flanders.
Smart stock management, asset tracking and mobile warehouse staff. These are just a few of the solutions that Proximus has to offer in the Transport and Logistics sector. Want to know more?