Safe forward-looking restraint system deployment for automated driving

• Applicant: Professor Dr.-Ing. Thomas Brandmeier
• Subject Area: Traffic and Transport Systems, Intelligent and Automated Traffic
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 545554847

The increase in automation and electrification of future vehicles, along with new technologies such as artificial intelligence, present not only opportunities but also challenges for vehicle safety. New interior concepts with a greater degree of freedom for occupant posture and movements, as well as the absence of the engine compartment, are changing the current safety structure. To achieve the "Vision Zero", the European goal of zero traffic fatalities, a technological leap is therefore necessary. The rise in the usage of forward-looking sensors offers the possibility to detect critical situations before an accident occurs and to enable the early activation of irreversible restraint systems before the crash, thereby enhancing occupant protection. Since this is a critical safety measure where errors are not acceptable, the primary objective of this project is to ensure the reliability of a predictive system. To overcome current system limitations, a safeguarding function, which monitors the immediate proximity (ultra-close vehicle environment) of the vehicle before impact will be developed. This particularly considers aspects such as hardware and software redundancy and explainability. A parallel safeguarding path will not only secure measurements disrupted by factors like weather conditions, but also AI-based and difficult-to-test black-box systems, which are increasingly being used today. To establish the safeguarding path, an additional robust safing sensor is required, which provides valid information just before the collision. The goal is to obtain essential measurement points and extend continuous coverage of the vehicle environment into the ultra-close range. This requires a sensor with a reduced cycle time (<10 ms) for the 0-5 meter range, a centimeter-level distance resolution and minimal complexity. A key aspect of the project is the development of methods to validate the crash object and specify the crash constellation, in combination with environmental perception and prediction based on existing AD/ADAS sensors. Considering the varying measurement parameters, levels of detail, cycle times, and measurement distance ranges of the information, the challenge lies in intelligently combining different types of information within the safeguarding function. By the end of the project, the goal is to combine the long-standing development and field experience of the application partner, Continental, in the areas of restraint systems and sensors with the extensive research on safe perception and prediction conducted by THI. This collaboration aims to achieve the project objectives through continuous knowledge transfer and to showcase the developed functionalities in a demonstrator. The demonstrator, which will be validated on a test track, is expected to reach Technology Readiness Level 5.

• DFG Programme: Research Grants (Transfer Project)
• Application Partner: Continental Automotive Technologies GmbH