Home / Current Issue / Paper 1705076
A Novel Framework for Predicting Driver Behavior at Unsignalized Intersections Using Machine Learning
Subject area: Science,Engineering and Technology · Area of research: Machine Learning
Abstract
Unsignalized intersections are critical points in road networks where the absence of traffic signals significantly increases the risk of accidents due to unpredictable driver behavior. Understanding and predicting driver decisions in such scenarios is crucial for improving traffic safety and designing intelligent transportation systems. This study presents a novel framework for predicting driver behavior at unsignalized intersections using machine learning techniques. The framework integrates multiple data sources, including vehicle dynamics, road geometry, and environmental factors, to create a comprehensive dataset for modeling driver behavior. Feature engineering techniques are employed to extract relevant features from the data, such as speed, acceleration, gap acceptance, and proximity to other vehicles. Several machine learning algorithms, including Random Forest, Gradient Boosting Machines, and Neural Networks, are evaluated for their predictive performance. The proposed framework incorporates advanced methods like hyperparameter tuning and cross-validation to optimize model accuracy and robustness. Results from simulations and real-world datasets demonstrate the framework's ability to achieve high predictive accuracy, outperforming existing methods. The study also highlights the interpretability of machine learning models, providing insights into key factors influencing driver decisions at unsignalized intersections. By identifying high-risk scenarios and enabling proactive interventions, this framework has significant implications for enhancing traffic management and safety systems. Furthermore, its scalability and adaptability make it suitable for integration into autonomous vehicle systems and smart city initiatives. Future research directions include the integration of real-time data from connected vehicles and IoT devices to improve prediction accuracy and the application of reinforcement learning to model more complex driver decision-making processes. This novel approach bridges the gap between machine learning applications and traffic behavior analysis, paving the way for safer and more efficient road networks.
Keywords
Driver Behavior Prediction, Unsignalized Intersections, Machine Learning, Traffic Safety, Feature Engineering, Intelligent Transportation Systems, Connected Vehicles, Autonomous Systems
How to cite this paper
@article{1705076,
author = {Sikirat Damilola Mustapha, Abidemi Adeleye Alabi},
title = {A Novel Framework for Predicting Driver Behavior at Unsignalized Intersections Using Machine Learning},
journal = {Iconic Research And Engineering Journals},
year = {2023},
volume = {7},
number = {3},
pages = {689-709},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1705076.pdf},
abstract = {Unsignalized intersections are critical points in road networks where the absence of traffic signals significantly increases the risk of accidents due to unpredictable driver behavior. Understanding and predicting driver decisions in such scenarios is crucial for improving traffic safety and designing intelligent transportation systems. This study presents a novel framework for predicting driver behavior at unsignalized intersections using machine learning techniques. The framework integrates multiple data sources, including vehicle dynamics, road geometry, and environmental factors, to create a comprehensive dataset for modeling driver behavior. Feature engineering techniques are employed to extract relevant features from the data, such as speed, acceleration, gap acceptance, and proximity to other vehicles. Several machine learning algorithms, including Random Forest, Gradient Boosting Machines, and Neural Networks, are evaluated for their predictive performance. The proposed framework incorporates advanced methods like hyperparameter tuning and cross-validation to optimize model accuracy and robustness. Results from simulations and real-world datasets demonstrate the framework's ability to achieve high predictive accuracy, outperforming existing methods. The study also highlights the interpretability of machine learning models, providing insights into key factors influencing driver decisions at unsignalized intersections. By identifying high-risk scenarios and enabling proactive interventions, this framework has significant implications for enhancing traffic management and safety systems. Furthermore, its scalability and adaptability make it suitable for integration into autonomous vehicle systems and smart city initiatives. Future research directions include the integration of real-time data from connected vehicles and IoT devices to improve prediction accuracy and the application of reinforcement learning to model more complex driver decision-making processes. This novel approach bridges the gap between machine learning applications and traffic behavior analysis, paving the way for safer and more efficient road networks.},
keywords = {Driver Behavior Prediction, Unsignalized Intersections, Machine Learning, Traffic Safety, Feature Engineering, Intelligent Transportation Systems, Connected Vehicles, Autonomous Systems},
month = {September},
}