| Abstract |
Among many methods presented for traffic congestion detection and control, the traffic light control system is effective, however, produces a slew of issues, including excessive delays and high energy consumption. To increase the efficiency, the traffic light length must be adapted dynamically corresponding to real-time traffic. Computer vision techniques have been commonly used for traffic detection. In this article, to regulate the signal for traffic, a Two-Dimensional Convolutional Neural Network (2D-CNN) is constructed for traffic detection and the TensorFlow is used to implement the 2D-CNN. The live traffic data is then used to regulate traffic lights. To validate the proposed method, it is compared with Vehicular Adhoc Network (VANET) adaptive traffic light control, using Simulation of Urban Mobility (SUMO) platform. The simulation's outcome shows that the proposed method is more effective at controlling traffic signals and reduces traffic congestion. The results showed 96% accuracy on the testing dataset. |
