A5080479612- Traffic and Road Safety
- Traffic control and management
- Traffic Prediction and Management Techniques
- Autonomous Vehicle Technology and Safety
- Transportation Planning and Optimization
- Human-Automation Interaction and Safety
- Vehicle emissions and performance
- Infrastructure Maintenance and Monitoring
- Urban Transport and Accessibility
- Face and Expression Recognition
- Automotive and Human Injury Biomechanics
- Tensor decomposition and applications
- Advanced Data Compression Techniques
- Vehicular Ad Hoc Networks (VANETs)
- Conducting polymers and applications
- Imbalanced Data Classification Techniques
- Advanced Clustering Algorithms Research
- Remote Sensing and LiDAR Applications
- IoT and GPS-based Vehicle Safety Systems
- Transportation Safety and Impact Analysis
- Artificial Intelligence in Healthcare
- Human Mobility and Location-Based Analysis
- Automated Road and Building Extraction
- Injury Epidemiology and Prevention
- Bayesian Methods and Mixture Models
Oak Ridge National Laboratory
2021-2024
National Transportation Research Center
2021-2023
University of Central Florida
2018-2022
With the help of traffic detectors widely deployed along arterial roads and intersections, real-time data are collected updated in a very short time period, which makes it possible to conduct analysis at signalized intersections. Among them, crash risk prediction is one most promising challenging research topics. This study attempts predict by considering series dependency with employment long short-term memory recurrent neural network (LSTM-RNN) algorithm. Also, synthetic minority...
Traffic violations of pedestrians at intersections are major causes road crashes involving pedestrians, especially red-light crossing behaviors. To predict the pedestrians’ intentions, video data from real traffic scenes collected. Using detection and tracking techniques in computer vision, some characteristics, including location information, generated. A long short-term memory neural network is established trained to intentions. The experimental results show that model has an accuracy rate...
Speed prediction is a crucial yet complicated task for intelligent transportation systems. The challenge derives from the complex spatiotemporal dependencies of traffic parameters. In past few years, deep neural networks have achieved best speed performance. However, most models depend on short-term input sequences to predict short/long-term (e.g., predicting next hour using data hour). These fail consider daily and weekly periodic behavior traffic. Another problem posed by lack...
Proactive traffic safety management systems can reduce crashes by identifying crash precursors, evaluating real-time risks, and implementing suitable interventions. The basic prerequisite for developing such a system is to propose reliable risk evaluation model that takes flow data as input. Previous studies have primarily focused on prediction using some statistical or machine-learning methods. However, further quantitative classification of risks been ignored. In this study, we conduct...
The advent of connected vehicle (CV) technology offers new possibilities for a revolution in future transportation systems. With the availability real-time traffic data from CVs, it is possible to more effectively optimize signals reduce congestion, increase fuel efficiency, and enhance road safety. success CV-based signal control depends on an accurate computationally efficient model that accounts stochastic nonlinear nature flow. Without necessity prior knowledge system’s architecture,...
In the context of pro-active traffic management, real-time crash risk evaluation is one most critical components. Signalized intersections are well-known high-risk locations because variety movements, modes, and their interactions. Unlike access-controlled freeways, flow at signalized presents cyclical characteristics, which temporally separated by signals. Therefore, data preparation for prediction should be based on signal cycle rather than a predefined fixed time interval (e.g., 5...
Min cut is an important graph partitioning method. However, current solutions to the min problem suffer from slow speeds, difficulty in solving, and often converge simple solutions. To address these issues, we relax into a dual-bounded constraint and, for first time, treat as nonlinear optimal transport problem. Additionally, develop method solving based on Frank-Wolfe (abbreviated DNF). Notably, DNF not only solves size constrained but also applicable all problems. We prove that convex...
With the challenges of increasing traffic congestion, concept managed lanes (MLs) has been gaining popularity recently as a means to effectively improve mobility. MLs are usually designed be left-lane concurrent with an at-grade access/exit. Such design forms weaving segments since it requires vehicles change multiple general purpose (GPLs) enter or exit ML. The could have negative impact on safety in GPLs. This study provides comprehensive investigation different lengths for each lane...