Urban Traffic Monitoring and Modeling System: An IoT Solution for Enhancing Road Safety

Understanding the potential of emerging digital technologies for improving road safety

“…The growing number of studies employing image processing [77,[79][80][81] and machine learning [76,[82][83][84][85][86][87] is not surprising due to the promising application of artificial intelligence techniques such as computer vision (using machine learning or deep learning algorithms) to automate the process of monitoring driver behaviour from images and videos. Machine learning was used to profile driver behaviour [79,83,84], predict driver manoeuvres at intersections [87], and evaluate the level of drowsiness [76].…”

“…Image processing was also employed to study and predict speeding [81] and outcome/severity of accidents [80]. It was also combined with IoT and AI by and Jabbar, Shinoy [82] to detect driver drowsiness with 82% accuracy. Smartphone-based applications were also utilised for analysing driver behaviour and to inform collision warning systems [88].…”

“…However, as a result of relatively low ratings of studies related to driver and pedestrian behaviour, limited conclusions can be drawn regarding the applicability of digital technologies for monitoring or predicting such factors. Except for one on driver [82] and pedestrian behaviour [91] respectively, none of the studies reported the accuracy of the proposed techniques. However, a prominent finding was that machine learning and image processing has the potential to be employed for this purpose, although further research is needed to determine the feasibility, desirability and effectiveness of such techniques, particularly from a data protection perspective.…”