Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Learning from geometry-aware near misses to real-time COR: A corridor-wide grouped random parameters GEV framework

Mohammad Anis, Yang Zhou, Dominique Lord

TL;DR

A geometry-aware two-dimensional time-to-collision near-miss extraction framework that integrates it with a hierarchical Bayesian grouped random parameter unified generalized extreme value model (HBSGRP-UGEV) to estimate short-term COR in urban corridors and indicates that the HBSGRP-UGEV framework outperforms the fixed-parameter HBSFP-UGEV model.

Abstract

Real-time corridor-wide crash-occurrence risk (COR) prediction is challenging because existing near-miss extreme value theory (EVT) models often oversimplify collision geometry, neglect vehicle-infrastructure (V-I) interactions, and inadequately account for spatial heterogeneity in traffic and roadway conditions. This study develops a geometry-aware two-dimensional time-to-collision (2D-TTC) near-miss extraction framework and integrates it with a hierarchical Bayesian grouped random parameter unified generalized extreme value model (HBSGRP-UGEV) to estimate short-term COR in urban corridors. The proposed framework builds on prior grouped EVT formulations while explicitly accommodating both vehicle-vehicle (V-V) and vehicle-infrastructure (V-I) near-miss processes within a unified corridor-wide modeling structure. High-resolution trajectories from the Argoverse-2 dataset were analyzed across 28 sites along Miami's Biscayne Boulevard to extract extreme near-miss events. The model incorporates vehicle dynamics and roadway features as covariates, with partial pooling across segments and intersections to capture corridor-wide heterogeneity. Results indicate that the HBSGRP-UGEV framework outperforms the fixed-parameter HBSFP-UGEV model, reducing the deviance information criterion (DIC) by up to 7.5 percent for V-V interactions and 3.1 percent for V-I interactions. Predictive validation using receiver operating characteristic area under the curve (ROC-AUC) demonstrates strong classification performance, with values of 0.89 for V-V segments, 0.82 for V-V intersections, 0.79 for V-I segments, and 0.75 for V-I intersections.

Learning from geometry-aware near misses to real-time COR: A corridor-wide grouped random parameters GEV framework

TL;DR

A geometry-aware two-dimensional time-to-collision near-miss extraction framework that integrates it with a hierarchical Bayesian grouped random parameter unified generalized extreme value model (HBSGRP-UGEV) to estimate short-term COR in urban corridors and indicates that the HBSGRP-UGEV framework outperforms the fixed-parameter HBSFP-UGEV model.

Abstract

Real-time corridor-wide crash-occurrence risk (COR) prediction is challenging because existing near-miss extreme value theory (EVT) models often oversimplify collision geometry, neglect vehicle-infrastructure (V-I) interactions, and inadequately account for spatial heterogeneity in traffic and roadway conditions. This study develops a geometry-aware two-dimensional time-to-collision (2D-TTC) near-miss extraction framework and integrates it with a hierarchical Bayesian grouped random parameter unified generalized extreme value model (HBSGRP-UGEV) to estimate short-term COR in urban corridors. The proposed framework builds on prior grouped EVT formulations while explicitly accommodating both vehicle-vehicle (V-V) and vehicle-infrastructure (V-I) near-miss processes within a unified corridor-wide modeling structure. High-resolution trajectories from the Argoverse-2 dataset were analyzed across 28 sites along Miami's Biscayne Boulevard to extract extreme near-miss events. The model incorporates vehicle dynamics and roadway features as covariates, with partial pooling across segments and intersections to capture corridor-wide heterogeneity. Results indicate that the HBSGRP-UGEV framework outperforms the fixed-parameter HBSFP-UGEV model, reducing the deviance information criterion (DIC) by up to 7.5 percent for V-V interactions and 3.1 percent for V-I interactions. Predictive validation using receiver operating characteristic area under the curve (ROC-AUC) demonstrates strong classification performance, with values of 0.89 for V-V segments, 0.82 for V-V intersections, 0.79 for V-I segments, and 0.75 for V-I intersections.

Paper Structure

This paper contains 25 sections, 29 equations, 12 figures, 8 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Near-miss detection framework
  3. Vehicle dynamics model
  4. Vehicle geometry representation
  5. V-V near-miss detection
  6. V-I near-miss detection
  7. COR estimation
  8. Extreme value theory (EVT)
  9. Univariate generalized extreme value (UGEV)
  10. Hierarchical Bayesian Structure Grouped Random Parameters (HBSGRP) model
  11. COR estimation
  12. Data description
  13. Motion forecasting dataset
  14. HD map
  15. Trajectory preprocessing and derivation of vehicle dynamics
  16. ...and 10 more sections

Figures (12)

  • Figure 1: Computation of 2D-TTC for near-miss scenarios
  • Figure 2: COR time series generation
  • Figure 3: AV data transformed to real-world coordinates and aligned: (a) trajectories with lane boundaries and crosswalks, (b) validation with satellite imagery, and (c–f) directional traffic patterns highlighting interaction hotspots
  • Figure 4: Effect of B-spline smoothing on a sample trajectory, showing removal of noise while preserving geometric fidelity
  • Figure 5: Segmented study corridor along Biscayne Boulevard, Miami, with lane boundaries and crosswalks. Insets show direction-specific segments (S1–S9) and intersections (I1–I10)
  • ...and 7 more figures