Leveraging an Optimized AI ‐Driven Model for Estimating IoV Traffic Flow

ABSTRACT

Traffic congestion in urban areas has become a major challenge due to rapid urbanization and the increasing number of vehicles, resulting in travel delays, environmental pollution, and reduced transportation efficiency. Existing traffic forecasting methods often struggle to capture complex temporal patterns and achieve consistent prediction performance in dynamic Internet of Vehicles (IoV) environments. This study proposes an Improved Dolphin Swarm‐optimized Dynamic Recurrent Neural Network (IDS‐DRNN), which integrates metaheuristic optimization with deep temporal learning for traffic flow prediction. The model is developed using a publicly available Kaggle Traffic Flow Monitoring Dataset comprising 4000 traffic records and 15 traffic‐related features, including vehicle density, speed, weather conditions, and network parameters. Min‐Max normalization is employed for data preprocessing, while principal component analysis (PCA) is used to reduce feature dimensionality and improve computational efficiency. The proposed model is implemented and evaluated using the Python environment. Experimental results indicate that IDS‐DRNN achieves an accuracy of 97.5%, precision of 96.5%, recall of 97.1%, RMSE of 1.01, and MAE of 0.48, demonstrating improved prediction performance compared with the baseline methods considered in this study. Although the Improved Dolphin Swarm optimization stage introduces additional computational complexity during model training, the trained model supports efficient inference for traffic flow prediction in IoV environments. Overall, the proposed IDS‐DRNN shows promising potential for supporting intelligent traffic management and reducing traffic congestion; however, further validation using larger and more diverse datasets is required to confirm its generalizability and reliability.