KisSME

To develop new vehicles with a 4/5 automation level, achieving a selective acquisition of relevant and critical driving situations, significant environmental data as well as raw data of the vehicle sensor system while driving are indispensable. This data is needed to validate, improve and reproduce the decisions made by artificial intelligence (AI), with the aim of thus achieving the necessary test coverage for future functionalities.

Within the KIsSME project, AI-based algorithms will be applied to enable on-board systems to recognize relevant and critical scenarios in real time and to selectively acquire raw data and scenario descriptions for this purpose. The AI-based algorithms will enable an inherent learning capability that continuously improves the recognition of critical situations and the associated relevant data to increase the information density of the data used for testing when developing level 4/5 automated systems, while simultaneously significantly reducing the associated data volume required as well as the effort needed to ensure data protection.