Last week I visited the Auto[nom]Mobil conference in Hanau. Besides the great organization of the event by carhs training gmbh, I enjoyed receiving insights in the development of safe automated driving vehicles in the various countries. Here are my key take-aways from the conference:
- Dr. Johannes Springer from Deutsche Telekom AG pointed out that when defining the operational design domain of an AD vehicle, one should always consider the connectivity of the vehicle, besides road types & conditions
- Dr. Yong Gessner from WeRide presented the readiness of L4 driverless technology in China as well as their first application in Europe, an autonomous electric shuffle at the French Open this year in Paris.
- Dushyant Wadivkar from TorcRobotics presented a robust autonomous trucking validation strategy, in which requirement-driven, design-driven, scenario-driven and data-driven approaches are combined
- Dr. Andreas Zeller from Bosch XC revealed that in order to verify road safety for AD vehicle, the developed SafeLatLon metric comprises conflict avoidance metrics, conflict metrics and collision avoidance metrics and collision metrics.
During my talk, I presented an approach to create a Safety Assurance Case for self-driving vehicles that employ Machine Learning algorithms, aiming to compile a robust safety argument. This approach goes beyond the standard safety requirements outlined in ISO 26262, ensuring that the intelligent learning systems are both reliable and consistent. The Safety Assurance Case method is designed to demonstrate that AI-driven features have been thoroughly investigated, thus fostering trust and responsibility as we advance towards the future of automated driving.
“Obrigado” as they say in Lisbon, Peter Hafmar! Proud for our collaboration to be profiled on both the Corporate Innovation and the AI stages at Web Summit!