ABSTRACT
An important aspect of a Smart City framework is its transportation system and the existence of an efficient traffic management system. Efficiency in a traffic management system is characterized—among others—by the network’s level of road safety. This paper studies a road safety oriented contribution to Smart Cities by quantifying driver’s risk perception in relation to vehicle to vehicle interaction. The objective of this study is to propose an automatic driving behavior monitoring mechanism for an urban environment, which identifies near-crash phenomena by capturing rear-end potentials at a microscopic level, while furthermore to induce driving behavioral aspects, valuable for understanding drivers’ perception on rear-end collision risk. The disaggregated data utilized in the study were obtained by inductive loop detectors in the urban network of Nicosia, Cyprus. The data gathered from the loop detectors was post-processed and a risk index based on rear-end potential was derived, which was used to classify drivers into four risk levels describing whether given their individual characteristics, drivers would engage in a potential rear-end collision. The proposed risk index results showed that 65% of the car-following events were considered as potentially unsafe. It was also shown that when engaged in car-following situations with Heavy Goods Vehicles-HGVs mean speeds of the following vehicles are lower. The proposed methodology enables the identification of potential near-crash events in an urban environment in real time. The information and knowledge collected by real-time data processing are key aspects of an efficient traffic management system and consequently a Smart City as a whole.
