Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Long-Horizon Traffic Forecasting via Incident-Aware Conformal Spatio-Temporal Transformers

Mayur Patil, Qadeer Ahmed, Shawn Midlam-Mohler, Stephanie Marik, Allen Sheldon, Rajeev Chhajer, Nithin Santhanam

Abstract

Reliable multi-horizon traffic forecasting is challenging because network conditions are stochastic, incident disruptions are intermittent, and effective spatial dependencies vary across time-of-day patterns. This study is conducted on the Ohio Department of Transportation (ODOT) traffic count data and corresponding ODOT crash records. This work utilizes a Spatio-Temporal Transformer (STT) model with Adaptive Conformal Prediction (ACP) to produce multi-horizon forecasts with calibrated uncertainty. We propose a piecewise Coefficient of Variation (CV) strategy that models hour-to-hour traveltime variability using a log-normal distribution, enabling the construction of a per-hour dynamic adjacency matrix. We further perturb edge weights using incident-related severity signals derived from the ODOT crash dataset that comprises incident clearance time, weather conditions, speed violations, work zones, and roadway functional class, to capture localized disruptions and peak/off-peak transitions. This dynamic graph construction replaces a fixed-CV assumption and better represents changing traffic conditions within the forecast window. For validation, we generate extended trips via multi-hour loop runs on the Columbus, Ohio, network in SUMO simulations and apply a Monte Carlo simulation to obtain travel-time distributions for a Vehicle Under Test (VUT). Experiments demonstrate improved long-horizon accuracy and well-calibrated prediction intervals compared to other baseline methods.

Long-Horizon Traffic Forecasting via Incident-Aware Conformal Spatio-Temporal Transformers

Abstract

Reliable multi-horizon traffic forecasting is challenging because network conditions are stochastic, incident disruptions are intermittent, and effective spatial dependencies vary across time-of-day patterns. This study is conducted on the Ohio Department of Transportation (ODOT) traffic count data and corresponding ODOT crash records. This work utilizes a Spatio-Temporal Transformer (STT) model with Adaptive Conformal Prediction (ACP) to produce multi-horizon forecasts with calibrated uncertainty. We propose a piecewise Coefficient of Variation (CV) strategy that models hour-to-hour traveltime variability using a log-normal distribution, enabling the construction of a per-hour dynamic adjacency matrix. We further perturb edge weights using incident-related severity signals derived from the ODOT crash dataset that comprises incident clearance time, weather conditions, speed violations, work zones, and roadway functional class, to capture localized disruptions and peak/off-peak transitions. This dynamic graph construction replaces a fixed-CV assumption and better represents changing traffic conditions within the forecast window. For validation, we generate extended trips via multi-hour loop runs on the Columbus, Ohio, network in SUMO simulations and apply a Monte Carlo simulation to obtain travel-time distributions for a Vehicle Under Test (VUT). Experiments demonstrate improved long-horizon accuracy and well-calibrated prediction intervals compared to other baseline methods.
Paper Structure (22 sections, 48 equations, 8 figures, 6 tables)

This paper contains 22 sections, 48 equations, 8 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Traffic Forecasting Models and Graph Learning
  4. Long-horizon forecasting
  5. Transformer Architectures for Traffic Forecasting
  6. Incident context features
  7. Uncertainty Quantification
  8. Proposed Methodology
  9. Graph Formulation and Data Availability
  10. Hourly Adaptive Adjacency Matrix Formulation
  11. Crash Attributes Integration in Adjacency Matrices
  12. Hourly Crash Risk Signals
  13. Formulating Incident-Aware Adjacency Matrix
  14. Model Architecture and Adaptive Conformal Prediction (ACP)
  15. Results and Discussion
  16. ...and 7 more sections

Figures (8)

  • Figure 1: Static vs. piece-wise graph priors: 1a) one fixed-CV $\hat{c}$ drives a log-normal travel-time prior; the resulting adjacency $A_{ij}$ is reused across hours, 1b) an hourly profile $CV(h)$ yields hour-conditioned priors and distinct adjacencies $A_{ij}(h)$. Edge colors/thicknesses are schematic.
  • Figure 2: Spatio-Temporal Transformer (STT-ED) architecture with hour-conditioned adaptive adjacency and Adaptive Conformal Prediction (ACP) for multi-horizon traffic forecasting
  • Figure 3: Sampled Traffic Route
  • Figure 4: Crash-induced change in adjacency weights across selected hours
  • Figure 5: Traffic Prediction with Uncertainty Bounds (Node 1)
  • ...and 3 more figures