Search

Developing countries such as India face many problems when it comes to the management of existing traffic system and its inefficient planning and scheduling leads to traffic congestion, increased pollution level and transit delays. This requires a significant amount of attention to build a smart traffic management system. Taking this into consideration a framework is proposed to determine traffic density and use the corresponding data to further modulate the traffic signal proficiently. This will be carried out by applying two different methodologies namely Background Subtractor MOG and TensorFlow Object Detection API. A video from surveillance camera is taken as input which will be converted into frames using OpenCV and after processing the detected vehicles are counted, thus the density is obtained. This is followed by comparing the results obtained by the mentioned methodologies. A vital application of the determined traffic densities can be in controlling traffic signals in a smarter way, therefore leading in decreased traffic congestion and uniform flow of traffic. [ABSTRACT FROM AUTHOR]

The main instrument for urban traffic control is the traffic light settings, which is applied in this research in order to improve the traffic dynamics. The goal is to develop a model for urban traffic control with bigger outgoing traffic flows than ingoing ones in order to be decreased the vehicle queues at crossroads. As a difference from the usual practice of using only one control influence of the traffic light settings, here two control settings - traffic light cycles and green light durations are implemented as control influences of an arterial network of five crossroads. The implementation of this approach is on the basis of the bi-level optimization methodology. It integrates two goal functions (of both hierarchical levels), more variables and constraints in comparison with the one-level optimization. The optimization problems of the bi-level optimization have been defined. The bi-level solutions have been compared with the one-level optimization ones. The numerical simulations give preference to bi-level optimization. [ABSTRACT FROM AUTHOR]