Credible Safety Validation Toolchain for Automated Driving Functions
摘要
The development, validation and homologation of automated driving systems (ADS) is still a major challenge for vehicle manufacturers and suppliers. It requires massive testing in different test environments to proof that the system or vehicle is safe enough in all defined situations and the residual risk is acceptable. ADS Regulations, such as UNE 157 ALKS or the EU type approval regulation for ADS 2022/1426, specify three main pillars: This paper introduces a comprehensive safety validation tool chain for ADAS and ADS which has been developed and released by AVL. The toolchain aligns with relevant standards, like SOTIF (ISO 21448), and extensively implements the requirements of the three pillars. The toolchain enhances and automates the development and validation process of these systems and vehicles through several key components ensuring both efficiency and safety in the creation of cutting-edge automotive technologies: (a) scenario database creation and scenario management, (b) systematic requirement and ODD (operational design domain) based test planning, (c) sophisticated test plan optimization and execution in virtual or real-world test environments and (d) integrated assessment of safety and residual risk.