A smart algorithm for traffic lights intersections control in developing countries

Traffic jam is a problem that directly affects the quality of life of the population in large cities. This problem exacerbates at road intersections, where obsolete traffic control systems based on a static set of rules remain in use. We propose an algorithm that improves vehicular flow control at traffic-light intersections by optimizing a dynamic allocation of times. We train our own YOLO detector using a set of images captured from traffic cameras installed at a cross-road. Based on the number of vehicles detected in each intersection road, one set of rules was created and used by a fuzzy control. Since, at the local level, there are few traffic cameras installed on intersections. We build a simulated environment both to train our detector system and verify the efficiency of our algorithm.