A Novel Machine Learning Based Traffic Congestion Recognition System

One of the most pressing problems in transportation nowadays is road congestion. Congestion and inattention due to constant haste are one of the main sources of road accidents. Rapidly evolving electronics have made it possible to equip our vehicles with wide variety of small, reliable and accurate sensors. The occurrence of V2X enabled vehicles and infrastructural elements to share their data thus a complex image can be constructed about the state of the traffic. Our paper presents a novel traffic congestion system that compares both supervised and unsupervised models. The presented system detects congested road sections and forecasts congestions based on previous sections. The system reaches up to 96% accuracy.