Third Generation Simulation Data (TGSIM): A Closer Look at The Impacts of Automated Driving Systems on Human Behavior

Talebpour, Alireza; Mahmassani, Hani S.; Hamdar, Samer H · 2024 · ROSA P / United States. Department of Transportation. Intelligent Transportation Systems Joint Program Office

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

This study addresses the critical gap in understanding human-automated vehicle interactions within mixed traffic environments. As the market penetration of Advanced Driver Assistance Systems (ADAS) and Automated Driving Systems (ADS) increases, existing traffic flow models and safety analyses lack sufficient empirical data to characterize how human drivers behave around automated vehicles. Previous datasets were either limited in scope, focused solely on automated vehicle behavior, or collected in closed-course environments. To remedy this, the Third Generation Simulation (TGSIM) project aimed to collect accurate trajectory datasets capable of characterizing human-automated vehicle interactions across diverse scenarios, including highway and city environments, and varying levels of automation (SAE Levels 1–3). The researchers employed three distinct videography methodologies to capture data in Chicago, IL, and Washington, D.C. Fixed-location aerial videography used a stationary helicopter hovering at high altitudes to capture wide roadway segments on I–90/I–94. Moving aerial videography involved a helicopter following strings of automated vehicles at lower altitudes on I–90/I–94 and I–294 to capture longer roadway segments. Infrastructure-based videography utilized synchronized, overlapping cameras mounted on overpasses and buildings to create comprehensive fields of view on I–395 and the George Washington University campus. The data processing framework involved preprocessing, deep-learning-based object detection and tracking, image stabilization, trajectory construction, and rigorous data cleaning and validation to ensure accuracy. A sample analysis of the collected data, specifically focusing on I-294, demonstrated the utility of the dataset for modeling automated vehicle driving behavior. The study calibrated car-following models, including the Intelligent Driver Model and a Constant Time Headway model, using both two-loop iterative and one-loop joint estimation methods. The results indicated that human drivers exhibit distinct behaviors in the presence of ADS-equipped vehicles, such as altered car-following patterns and headway choices. These behavioral changes can significantly impact traffic flow dynamics, suggesting that current traffic flow models require improvement to accommodate the presence of automated vehicles. The significance of this work lies in providing a robust methodology for extracting accurate vehicle trajectory data from real-world settings, which is essential for improving microscopic simulation tools and safety analysis methods. The report offers detailed lessons learned regarding data collection setup, instrumentation, and experimental design, providing guidance for future research. By capturing interactions across various traffic flow regimes and automation levels, the TGSIM dataset enables researchers to investigate the impacts of ADAS on human behavior, string stability, and traffic flow dynamics. This contributes to a deeper understanding of mixed traffic environments, supporting the development of novel congestion management strategies and more accurate traffic planning models as automated vehicle adoption grows.

Key finding

Human drivers exhibit different car-following behaviors in the presence of automated driving systems, necessitating improvements in traffic flow models and safety analysis methods.

Methodology

naturalistic

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• situational awareness
• traffic density
• driverless ads

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Methodological Resource: tool software, dataset resource
• Theoretical Contribution: computational model