Amr Abdelraouf

This study proposes a three-stage framework for real-time crash likelihood and severity prediction. Firstly, a real-time crash likelihood prediction model was developed. Secondly, a real-time crash severity clustering model was proposed to cluster the crashes into different severity levels. Thirdly, a severity clustering validation model was developed to assess the performance of the proposed severity clustering model. Extensive data processing techniques were employed to collect real-time features from State Road 408 in Orlando, Florida, and a total of 6,750,072 events (625 crash events and 6,749,447 non-crash events) along with 24 real-time features were used. To develop the crash likelihood prediction model, nine machine-learning techniques were attempted, and the convolutional neural network model was found to provide the best result with respect to the sensitivity (0.916), false alarm rate (0.111), and area under the receiver operating characteristic curve (0.967). Davies–Bouldin index criteria were used to find the detector location that generated the most accurate traffic information to cluster the crashes into severity levels, and based on this traffic information, k-means clustering was applied to develop the severity clustering model. Finally, a severity clustering validation model was developed after investigating nine machine-learning techniques to validate the developed severity clustering model, and the decision tree model provided the best results based on three levels of sensitivity and specificity values. The developed framework has the potential to help traffic management centers to warn road users or develop transportation systems management and operations strategies in real time to avoid crashes or minimize the severity and, thus, can significantly contribute to improving road safety.

This research presents a robust approach for automated safety diagnosis using sensor fusion techniques. This work fuses the outputs of a roadside low-resolution camera and a solid-state LiDAR. For vehicle classification and detection in videos, the YOLO-v5 object detection model was utilized. The raw 3D point clouds generated by the LiDAR are processed by two manual steps—ground plane transformation and background segmentation, and two real-time steps—foreground clustering and bounding box fitting. Taking the generated 2D bounding boxes of both camera and LiDAR, we associate the common bounding box pairs by thresholding on the Euclidean distance threshold of 6 ft between the centroid pairs. We perform weighted measurement updates based on the root mean square error of each of the sensor’s detection compared with manually labeled ground truths. The fused measurements are tracked by using linear constant velocity Kalman filter. With the generated trajectories, we compute post encroachment time at pixel-level conflicts based on the generated vehicle trajectories. We have proposed a complete bipartite graph-matching strategy of vehicle parts along with the conflict angle to obtain conflict types—rear-end, sideswipe, head-on, and angle conflict. A case study on a signalized intersection is presented. The output of the proposed framework performs with 97.384% precision and 95.316% recall. It is better than both single-sensor-based systems in relation to detection count and localization. It is expected that the proposed method can be employed to diagnose road safety problems and inform the required countermeasures.