Intelligent Detecting of Emergency Vehicles on the Road Networks: Available Datasets Assessment

Intelligent transport systems have received high attention from researchers as part of smart cities recently. Several advanced technologies have been used to enhance the smart behavior of vehicles over road networks. Machine learning and Artificial intelligence mechanisms have widely been used in this field of research. In order to allow the traveling vehicles to make the best decisions based on the real-time traffic distribution during their trips. Analyzing the real-time traffic characteristics and context over the road network has received extensive attention. In this work, we mainly aim to assist and investigate the available datasets used to analyze the context of traffic in terms of detecting emergency vehicles. The source, characteristics, and limitations of these datasets are investigated. Moreover, the effects of these datasets on the performance evaluation of the developed detection models in this field are discussed. From this study, we can mainly infer that the available datasets suffer several drawbacks and limitations that severely affect the accuracy and performance of the trained models in this field of research. Thus, we recommend preparing a more realistic and representative dataset.