Bird’s eye view trajectory reconstruction of naturalistic crashes and near-crashes in the SHRP2 NDS

Project Description This database contains reconstructed bird's eye view trajectories from SHRP2 NDS, including 10,919 safe baseline trips and 8,111 trips involving safety-critical events (crashes and near-crashes). Of the total trips, 3,893 safe baseline and 6,664 safety-critical trips have the trajectories of both subject vehicles and surrounding objects reconstructed. The data include kinematic time series detailing vehicle dynamics, structured tabular features describing event conditions and incident characteristics, along with metadata descriptions. All the data are stored in open file formats of HDF5 (.h5), CSV (.csv), and PDF (.pdf). Data Specification General Structure The dataset is organized into two primary archives: ReconstructedTrajectories includes reconstructed trajectories (Ego_birdseye.h5, Surrounding_birdseye.h5), event metadata (metadata_birdseye.csv), extended Kalman Filter parameters (ekf_parameters.csv), event categorization (event_counts.csv), and plots for trajectory visualizations. SafetyCriticalTestSet contains event-specific metadata (event_meta.csv), environmental data (environment.csv), and detailed trajectory data (event_data.h5). Variables The datasets provide comprehensive detail on events, vehicle states, and environmental contexts, including: Ego Vehicle Data (Ego_birdseye.h5): raw and reconstructed data such as timestamps, speed, yaw rate, acceleration, braking, steering, turn signals, and filtered positional coordinates. Surrounding Vehicle Data (Surrounding_birdseye.h5): radar-detected and reconstructed positions, velocities, and headings of nearby vehicles. Event Metadata (event_meta.csv): vehicle dimensions, event timings (start, impact, reaction times), severity, conflict type, and narratives. Environmental Conditions (environment.csv): lighting, weather, surface condition, and traffic density. A pdf file for a detailed data dictionary as attached.