Towards Traffic Congestion-Free through Intelligent Traffic Control System

One of the most serious problems that take lots of attention during last decade is the massive increasing in number of vehicles, which leads to dramatically traffic congestion. Such problem in cities puts huge stress on various systems for decision making and infrastructure planning of many metropolitan areas. Therefore, traffic control becomes the main concern for different systems to overcome this gradual challenge as associated phenomena of the continuous growth of population and vehicles over inadequate road infrastructures. The conventional systems rely on using a static timer for traffic signal, but this solution has multiple limitations, as traffic signal changes its lights even there are no vehicles on the road, which causes fake jamming and waste time and resources. Thus, recent literatures show massive improvements of reducing traffic overcrowding through the usage of Internet of Things (IoT) technology to enable adaptive traffic signal timers. IoT facilitates dealing with the collected data for processing and analysis through integration with Cloud by the attached sensors and cameras on each signal node. In this paper, we assume each traffic signal alongside the road has multiple IoT sensors and a camera to detect the flow rate and type of different vehicles and send to Cloud as a powerful platform for perform accurate detection and information extraction through image processing approach such as YOLO algorithm. The results showed reduction on waiting time by 60% when manipulate traffic signal timer to accurately adapt signal with the real-time traffic flow alongside the roads. Such improvement relies on efficient usage of Q-learning approach to enhance the quality of decision making.