Related Experiment Video
Updated: Aug 7, 2026
Evaluation of an Exclusive Spur Dike U-Turn Design with Radar-Collected Data and Simulation
Published on: February 1, 2020
Validation and analysis of real-world traffic accident black spots in a driving simulator environment
Kevser Ünsalan1, Ömer Faruk Cansız1, Mehmet Fatih Can2
1Department of Civil Engineering, Iskenderun Technical University, Hatay, Türkiye.
Objectives:
This study presents a comparative evaluation of conventional traffic accident black spot analyses based on historical crash records and driving simulator-based black spot analyses within a proactive road safety framework. The primary objective is to investigate whether driving simulator data can identify hazardous road segments beyond those detected through conventional crash-based approaches.
Methods:
In the first stage, traffic accident black spots were identified using Kernel Density Estimation (KDE) based on official police-reported fatal and injury crash records from the 2017-2020 period; property-damage-only crashes were excluded from the analysis. In the second stage, the examined road segment was replicated in a high-fidelity driving simulator environment, and controlled experimental driving sessions were conducted with licensed drivers. Driver behavior and vehicle control performance were assessed through critical risk events observed during the simulations, including run-off-road events, collisions with roadside safety elements, severe loss-of-control situations, and crash events preventing route completion. The spatial locations of these simulator-derived events were recorded and analyzed using KDE to identify clustered risk sections along the road.
Results:
The results demonstrated strong spatial consistency between several simulator-based risk locations and conventionally identified accident black spots. In addition, the simulator analyses revealed several hazardous road segments that were not identified through official crash records. These locations were associated with vehicle instability, control difficulties, and increased run-off-road tendencies, indicating the presence of latent safety risks before their manifestation in conventional accident statistics.
Conclusions:
The findings indicate that driving simulator-based black spot analyses complement and extend traditional crash-based approaches by enabling the early detection of potentially hazardous road segments. The proposed methodology offers a driver-oriented, proactive, and preventive framework that can support road safety planning, risk assessment, and infrastructure improvement studies.
