Long-Horizon Traffic Forecasting via Incident-Aware Conformal Spatio-Temporal Transformers
2026-03-17 • Machine Learning
Machine Learning
AI summaryⓘ
The authors worked on predicting traffic conditions for multiple hours ahead, which is hard because traffic changes randomly and incidents happen unexpectedly. They used special models called Spatio-Temporal Transformers combined with a technique to estimate how uncertain their predictions are. By creating a dynamic map of how traffic spots connect each hour and including information about accidents and road conditions, their method captures traffic changes better than older approaches. They tested their model with simulations of trips in Columbus, Ohio, and found it gave more accurate and reliable forecasts than other methods.
Multi-horizon forecastingSpatio-Temporal TransformerAdaptive Conformal PredictionCoefficient of VariationDynamic adjacency matrixTraffic incident severitySUMO simulationMonte Carlo simulationTravel-time distributionVehicle Under Test (VUT)
Authors
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.