Enhanced Implementation of Intelligent Transportation Systems (ITS) based on Machine Learning Approaches

Intelligent Transportation Systems (ITS) present a range of challenges that need to be addressed to ensure successful deployment, seamless integration, robust interconnection, and robust security. These challenges arise from the complex nature of ITS, the multitude of stakeholders involved, and the need to balance technological advancements with privacy and security concerns. Vehicle trajectory prediction is a critical task for improving traffic management and safety. It can be used to develop advanced traffic control systems, enhance autonomous vehicle navigation, and support real-world applications in intelligent transportation systems. This paper presents a detailed comparison and study on predicting vehicle trajectories using the NGSIM dataset, a publicly available dataset of high-quality vehicle trajectory data. A practical data set had been used to extract the main features of ITS, then we implemented two machine learning (ML) models for vehicle trajectory prediction by applying classification and regression tasks of both machine learning models, Random Forest (RF) and Support Vector Machine (SVM). We evaluated the models using classification accuracy, precision, recall, F1-score, mean squared error (MSE), and Rsquared. The results show that both models can predict vehicle velocity and type with reasonable accuracy. However, the random forest outperformed the SVM on the regression task, while the SVM outperformed the random forest on the classification task. We also studied key factors that influence the performance of vehicle trajectory prediction models and proposed suggestions for improving their performance.