Performance Assurance
The SUNRISE Safety Assurance Framework (SAF) is a harmonized structure of processes and procedures that allows stakeholders to determine whether a CCAM system meets a set level of safety, for introduction and during deployment.
The SAF is based on the multi-pillar approach from UNECE’s New Assessment/Test Method for Automated Driving (NATM). The SAF includes an audit of the safety assessment process, the safety management system as followed by the CCAM manufacturer, and the used test methods and test tools. The framework adopts a scenario-based approach, where tests (virtual testing, track testing and real-world testing) are based on scenarios taking into account the CCAM system’s Operational Design Domain (ODD) for assessing the safety of the system. Additionally, it provides procedures to monitor that the CCAM’s safety is maintained during the system’s lifecycle. The scenarios that feed into the SAF are knowledge based and/or data driven.
The SAF includes, but is not limited to, processes to:
- enable user input on system ODD & behaviour, further external requirements, and test objectives
- generate the relevant scenarios,
- allocate test scenarios to the different test methods,
- execute the test scenarios,
- assess and analyse the test results to come to a statement about safety assurance, and
- monitor the CCAM system and provide information of the system to the stakeholders to ensure safety during deployment.
In the SUNRISE project, the focus is on Performance Assurance using a scenario-based approach and incorporating virtual environments. In addition, it features a federated layer which allows users to obtain scenarios from multiple qualified scenario databases. Now, the main building blocks of the Performance Assurance workflow will be examined. At a high level, it comprises the Scenario, Execute, and Analyse stages. Essentially, this means that to conduct the safety assurance process, a set of scenarios is needed, an environment for executing these scenarios, and a process to analyse the execution results to determine whether the system is safe.