A Deep Learning Approach to Dynamic Traffic Flow Prediction and Management in VANETs

The evolution of the Internet of Things (IoT) facilitates the emergence of the Internet of Vehicles (IoV) and Intelligent Transportation Systems (ITS). A crucial component of ITS is the vehicular ad hoc network (VANET) featuring smart vehicles (SV). This study introduces a dynamic traffic regulation method within VANET utilizing Deep Neural Networks (DNN) and Bat Algorithms (BA). The DNN is employed to direct vehicles through highly congested routes, thereby minimizing average delays and enhancing efficiency. To assess traffic congestion among network nodes, the BA is integrated with the IoT and implemented over VANETs. Experiments were carried out to evaluate the proposed method's effectiveness based on various parameters, including packet delivery ratio, average latency, and throughput, with results compared against several machine learning (ML) algorithms. The simulation outcomes indicate that the proposed approach effectively manages real-time traffic conditions while consuming less energy and incurring lower delays compared to existing techniques.