Codes for our paper "HUTFormer: Hierarchical U-Net Transformer for Long-Term Traffic Forecasting". Please refer to READM.md and Replication.pdf file for more details.

The dataset focuses on vehicle conflicts on the motorway, it has both vehicle kinematics data between the ego-vehicle and neighbouring vehicles, as well as traffic-based variables. The traffic variables are collected from 500m road segments and are averaged over a 5 min period before the ego-vehicle passed at every time step. If several consecutive loop detectors were not functioning, the missing cells are stored as -1. The data uses Modified Time-To-Collision as the conflict indicator. It was reconstructed a posteriori by synchronising collected driving data with corresponding traffic data that would theoretically be available in real time. This approach ensures temporal and spatial consistency between datasets, enabling realistic simulation of real-time data collection scenarios.

Replication code for 'Domain-Enhanced Dual-Branch Model for Efficient and Interpretable Accident Anticipation'.

The uploaded files include:

Description file:

Description of the dataset and access to the replication codes.

Datasets

1) Hexagon level datasets used for the clustering analysis in the paper.

2) City-level data used for the main regression analysis.

Code

Code is accessible via `https://github.com/brookefzy/global-city-road-crash`

During the eye-tracking-based data collection —using a Tobii Pro eye camera— participants were presented with a total of nine image pairs depicting streetscapes in their "before" (current) and "after" (following the proliferation of AVs) states for a specific street segment. As a preliminary step, a 60-second calibration procedure was conducted, during which participants were instructed to visually track a moving dot on the screen. Following the briefing and calibration, the image pairs were displayed for data collection, with each pair shown on the screen for 12 seconds.

In terms of the data collected via eye cameras, four important measures can be defined:

• Total Fixation Duration (TFD): the cumulative duration of all fixations within the defined Area of Interest (AOI), reflecting the overall level of visual attention.
• Average Fixation Duration (AFD): the mean duration of individual fixations within an AOI.
• Fixation Count (FC): the total number of fixations recorded within the AOI.

• Time to First Fixation (TTFF): the time elapsed from the onset of the stimulus to the first fixation within the AOI.