Problem Statement

The increasing automation of road traffic is fundamentally changing the interaction between human and automated traffic participants. This transition raises not only technical but also societal challenges: How does the migration from a human-centered to an automated traffic environment progress? What new interaction patterns and safety requirements arise? What happens when the automation level of a vehicle is misjudged? And how can the associated risks be proactively minimized?

These complex interaction patterns primarily occur at the tactical level of driving behavior— a level that current microscopic traffic flow simulations are not yet able to model adequately. To integrate such behavioral models, explicit modeling of the tactical layer is required. However, this is currently missing in existing modeling and software architectures.

Objective

The research project aims to comprehensively investigate and model the migration from a human-centered to an increasingly automated traffic environment. In individual subprojects, key aspects of the interaction between humans and automated vehicles at various levels of automation are analyzed. This includes internal processes such as perception and decision-making as well as external interactions like communication and cooperation. Based on these analyses, precise models are developed to describe and systematically categorize these interactions.

The developed submodels are integrated into a microscopic traffic flow simulation. This simulation enables the realistic representation of traffic flow and interactions between humans and vehicles within a road network. A key challenge is incorporating the complexity of the tactical level of driving behavior into the simulation while managing computational and data-intensive demands.

Finally, the simulation will be used to investigate various potential migration paths. The goal is to identify and mitigate critical scenarios at an early stage, ensuring the safe, efficient, and socially accepted introduction of automated vehicles.

Methods

In Subproject 6, the focus is on integrating behavioral and interaction models developed in other subprojects into a microscopic traffic flow simulation. The open-source tool SUMO serves as the simulation platform. The emphasis is on the tactical level of driving behavior, which governs planning and decision-making during driving.

A major challenge is that existing model and software architectures for microscopic simulations lack explicit modeling of the tactical level. Tactical decisions are often implemented as special cases at the operational level, complicating the integration of advanced behavioral models. To address this, a "Clean Architecture for Traffic Flow Simulation" is being developed, providing a clear separation between tactical and operational models.

This architecture enables realistic modeling and simulation of tactical driving maneuvers and cooperation between human and automated road users. These aspects are examined based on vehicle trajectories to gain a detailed understanding of the underlying interaction patterns. Scenarios such as overtaking, merging onto a highway, or pedestrian crossings in urban areas are analyzed through simulation. The aim of this simulation-based research is to develop a deeper understanding of potentially critical migration paths and lay the groundwork for their proactive design.