IoT solutions for Sustainable Cities: An Online Adaptation for the Driver Intent Inference Algorithm

In this paper we introduce an adaptation to the "driver intent inference algorithm for urban intersections". This algorithm has been proven to detect potential right turns of vehicles by estimating the probability of a driver to turn right, and we propose to use it for reducing the number of cycling deaths at an intersection. We extend this algorithm following the IoT design principles and thus, with this approach, cyclists' safety no longer depends only on actions taken inside the vehicles, but also can use additional safety solutions based on standards and available information shared about the vehicles and drivers in vehicular networks. Our approach proposes to process the inference algorithm outside the vehicle, considering cloud and edge computing. We use predicting models for identifying driver's intention of turning right at intersections and the use of edge connected devices running our algorithm for alerting cyclists of possible collisions, thus preventing as many collisions as possible in intersections.