Smart traffic light control system based on traffic density and emergency vehicle detection

Abstract

Transportation is one of the main aspects of a country’s economy. Most economic sectors are laid upon detrimental results due to an unorganized transportation network. This is a crucial issue faced by developing countries. There is no doubt that highways should be built in order to maximize the throughput of the transportation network; nevertheless, expansion of existing roads is also not applicable in countries like Sri Lanka due to its ceasing land area with increasing population. Thus it is essential to switch to a more efficient, technologically advanced approach to solve this issue. In addition to the typical congestion scenarios, the prevailing pandemic situation has realized the importance of prioritizing ambulances when it is caught amidst a traffic jam. Pedestrians are another vital part of the road network. Effective and safe pedestrian crossing will ensure the reduction of road accidents while improving the existing heavy traffic. A smart traffic monitoring system integrated to control the traffic signals is the ideal solution in this context. This paper proposes a smart adaptive traffic monitoring and control system to detect vehicles and pedestrians and prioritize emergency vehicles. A new Convolutional Neural Network is trained with YOLOV3 architecture to achieve 91.3% detection precision.