With the upcoming introduction of vehicle and infrastructure connectivity, a new paradigm shift will take place in the field of integral safety systems. After the mitigation of crash injuries in passive safety and the avoidance respectively mitigation of collisions in active safety, now cooperative maneuver planning for the avoidance of collisions and dangers by sharing information via appropriate connectivity services will be the next step towards Vision Zero. In general, cooperation requires awareness of the common situation as well as awareness of the operational design domains, intentions, and capabilities of the others. The criticality of a traffic situation with respect to collision probability and severity is a parameter, which is relevant for both, the situation awareness and the expression the ODD, intentions, and capabilities. Hence there should be a common understanding and accepted definition of collision criticality when it comes to cooperative collision avoidance. The given paper will propose such a common definition based on Bayesian approaches, which is general and compatible for any kind of traffic participants (cars, truck, (powered) two-wheelers, pedestrians, etc.) and which can be used for distributed, interactive algorithms in collaborative collision avoidance. In addition, accompanying development and validation procedures are described, allowing a common assessment and judgment for the effectiveness rating of the information exchange across different types of traffic participants. That way we gain an universal approach, allowing to build cooperative ADAS and AD systems and functions with significantly improved performance in collision avoidance, incorporating information, which cannot only be acquired by the vehicles onboard sensors. Instead, also information like the observations of other traffic participants and the infrastructure, the other’s intentions and capabilities can be considered in cooperative maneuver planning for collision and danger avoidance. The common framework also allows development and validation across different brands, vehicles, and road user types. A general limitation of connected services for safety critical applications is the trustworthiness of the information coming from different other sources. Hence the level of trust and the standards of information preparation must be incorporated into the message content as well. To build up trust, all contributors including infrastructure roadside units must comply to according safety standards like SOTIF and similar. The proper design and modeling of the exchanged information and messages content is a crucial part of such. In fact, this has been considered here too. Anyhow, information design must gain broader acceptance in the community first before entering standards and norms. The given papers strive to contribute to such a discussion for establishment and understanding of the aspired information exchange. The presented approaches are partially result of the project COPE – Collective Perception (see www.project-cope.eu), which already have been presented in principle. The given paper puts the focus on the details of the criticality rating and the underlying mathematical principles including the way these can be applied in practical use. Some formal criteria and standards for building up trustworthiness in the case the information preparation and exchange provided in form of a digital twin of the infrastructure, have been treated within the project DIGEST. Summing up, vehicle and infrastructure connectivity promises to deliver some significant contribution on the way to Vision Zero. Nevertheless, this requires new approaches in collaborative and cooperative algorithms development and application. The given paper aims to deliver some contribution to begin setting some standards in the generation of supportive information exchange for active collision and danger avoidance with accordingly connected ADAS and AD systems and functions.
Dr. Andreas Kuhn, CEO/CTO, ANDATA