Controlling Risk for Highly Automated Transportation Systems Operating in Complex Open Environments

CONTROL builds trust in autonomous systems for road and rail.

Ensuring Quality, Understanding Mobility Systems. Discover how the research project CONTROL makes uncertainties manageable and accelerates cross-domain, forward-looking innovations in the mobility sector.

Autonomous systems in transportation must be able to respond safely to a wide range of possible scenarios in open, dynamic environments, including rare and unpredictable events. Existing methods fail to perform reliably in complex and less predictable operational areas.

The CONTROL research project addresses this challenge across the automotive, commercial vehicle, and rail sectors.

On the road, situations such as suddenly appearing pedestrians in areas with limited visibility, unexpected construction zones, faulty or contradictory traffic signals, extreme weather conditions affecting sensors or technical malfunctions in the vehicle, are key concerns. These scenarios are currently not systematically safeguarded. In the rail sector, additional challenges arise, such as people or obstacles on the tracks.

Autonomous systems must reliably detect and assess such safety-critical situations. CONTROL is developing a cross-industry, systemic approach to this. The project focuses on autonomous vehicles including cars, trucks, buses, and trains as an integrated system for evaluating and managing uncertainties.

Research approach

Detecting, assessing, and managing uncertainties. CONTROL lays the foundation for safe autonomous driving in open-world environments.

The diversity and dynamics of open environments exceed the limits of traditional safety assurance methods. That’s why CONTROL takes a new approach: the system continuously evaluates how safely it can operate in a given situation and responds to uncertainty with cautious actions, such as adjusting speed or executing controlled evasive maneuvers.

CONTROL builds on the foundations laid by the PEGASUS and VVM projects, which systematically derived test cases and secured complex operational domains. As part of the VDA Leitinitiative for autonomous and connected driving, CONTROL extends this path by adding the capability to actively assess and manage uncertainties.

Instead of analyzing individual components in isolation, the autonomous vehicle is understood as a holistic system, including adaptive methods for evaluating uncertainties during both development and runtime. A central element is the creation of a cross-industry safety argumentation framework that systematically captures, assesses, and integrates uncertainties into the architecture of autonomous systems.

CONTROL will become a key building block for the software-defined vehicle (SDV), laying the groundwork for faster development, updates, and accelerated innovation. The metrics, models, and tools developed are scalable, cross-domain, and industrially applicable, with direct relevance for road, rail, and beyond.

Two domains – one objective: Unified metrics and testing frameworks

Impulse

From Research to Practice. CONTROL deploys the developed tools into action.

The approaches developed in CONTROL are being applied in real-world road and rail environments. The entire perception chain, from sensors and data processing to evaluation, is treated as a cohesive system. Conditions such as weather, surroundings, or technical faults are modeled and addressed using adaptive methods.

The metrics, models, and tools emerging from the project, including digital twins and uncertainty models, support the development, evaluation, and validation of autonomous systems. They form the foundation for a cross-industry safety argumentation and are tested in various environments: in computer simulations, lab setups such as test settings, and in demonstrators and test vehicles that model real-world scenarios.

Robust systems, accelerated innovation, and practical safety ensure direct benefits for industry, research, and society.