Fraunhofer Institute for Transportation and Infrastructure Systems IVI
A collaborative situation report sees more than sensors alone
How can vehicle sensors and infrastructure sensors work together to create one reliable situation report—and make automated driving safer through it? This is the question that the VALISENS project partners spent three years researching in detail by investigating how the environment of automated vehicles can be detected so that all road users are kept safe. During the project's closing event at Fraunhofer IVI in Dresden, the consortium presented their results in front of an expert audience and showcased in a demonstration on the institute's test track how research can be transferred directly to practical application.
Preventing accidents and strengthening trust
At the core the VALISENS project was the achievement of comprehensive, robust and networked environment perception via sensors by merging fields of view, thus ensuring the safety of automation functions at levels 3 to 5 even in complex environments. The aim: preventing accidents involving vulnerable road users (VRUs) and strengthening trust in autonomous systems.
Together with their application partners LiangDao, NXP Semiconductors, XITASO, and smartmicro, Fraunhofer IVI researchers developed multiple individual system components that demonstrate their strengths in interaction. Three test vehicles, one mobile rover, infrastructure sensors contribute their respective environment information. This data is fused in a decentralized way in an Edge Cloud and subsequently exchanged between the systems via standardized V2X communications. All networked units, whether vehcile or infrastructure, are able to seamlessly integrate the information received and improve their individual environment model.
The combination of these diverse data sources creates a comprehensive environment perception system that is capable of producing a precise situation report: Each individual system "sees" what the other systems detect. The different sensor types, such as LiDAR, radar, RGB and infrared cameras, each have their own strengths which combine to balance out the weaknesses of individual technologies. In this way, fusion helps to improve the overall quality of perception.
In order to guarantee the sensors' reliability and detect sensory degradation at an early stage, digital self-testing and methods of remote inspection and predictive maintenance are applied.
Validation in digital test beds
Within the scope of multiple measuring campaigns, the project partners carried out practical test of the implementations in the digital test beds of Dresden and Ingolstadt (»Erste Meile«). In Dresden, the so-called smart intersection, which is fully equipped with sensors, was available for this purpose. Here, it was possible to study the interaction of the sensors combined with those of the institute's test vehicles in regular road traffic. An excursion to the smart intersection was also part of the live tour undertaken during the closing event.
It is the joint aim of the institute and their application partners to continue investigating the potential of sensor fusion within further projects, thus promoting the mission of achieving accident-free mobility in the future.
VALISENS received funding from the Federal Ministry for Economic Affairs and Energy within the scope of the KoPa4Mobility program.
