The Intelligence of the transportation systems is an emerging area in the domain of technological research. There exists an imminent demand for reduced congestion in cities, improved traffic management, effectual predictive capacity of the vehicles to accurately categorize driver responses while making decisions for the witnessed driving styles, reduced risk of collision and better traffic reasoning have urged the development of Intelligent Transportation Systems (ITSs). This paper focuses on another significant aspect of ITSs called the ability of lane change prediction of the vehicular traffic flow. The effectiveness of the four machine learning (ML) models, namely, Support Vector Machine (SVM), Naive Bayes (NB), Random Forest (RF) and Decision Tree (DT) is evaluated for measuring the predictive capacity to detect change in lane using the high-fidelity data on vehicular traffic flow collected by the US Federal Highway Administration (FHWA) for the Peachstreet, Atlanta, GA. The different scenarios involve the measurement of accuracy and performance comparison to choose the optimal and accurate model for the task of lane change prediction. The results reveal that SVM outperforms the other three Machine Learning (ML) models in regard to the task of precise and accurate prediction of lane change of the vehicle.
The Facets of Machine Learning in Lane Change Prediction of Vehicular Traffic Flow
Algorithms for Intelligent Systems
Proceedings of International Conference on Intelligent Cyber-Physical Systems ; Chapter : 27 ; 353-365
2022-01-24
13 pages
Article/Chapter (Book)
Electronic Resource
English