Density Based Real-time Smart Traffic Management System along with Emergency Vehicle Detection for Smart Cities

Traffic congestion is one of the major modern-day crisis in the world. There are many reasons behind this problem, among which the common reasons are poor traffic management, cars changing lanes, unplanned stoppage, dysfunctional traffic lights, drivers not following rules, emergency vehicle priorities not met etc. To overcome such situations traffic police is placed and the traffic congestion is handled by them manually. But in congested cities, it is very tough to handle huge traffic by a traffic police manually. As more and more vehicles are being commissioned in an already congested traffic system, there is an urgent need for a whole new traffic control system using advanced technologies to utilize the already existent infrastructures to its fullest extent. In this work, we create a fully automated system for traffic control based on traffic density with the help of a machine learning algorithm. We used foreground background subtraction to identify the vehicles in each lane. Using K-nearest neighbour algorithm we computed the density of each lane. Using KNN algorithm we found the accuracy as 99.04% and recall as 73.18%. We then create a database with the density values of each lane using phpmyadmin. The density values are fetched by NodeMCU from the cloud and traffic signals are activated based on the largest density in a round robin fashion. We further improvise the system for prioritizing emergency vehicles in the congestion. We use the Yolo object detection algorithm to detect emergency vehicles like ambulances so that traffic can be cleared up for them.