HEADSTART D4.2 Evaluation results of application and demonstration

This deliverable presents the work done in HEADSTART work package 4: Task 4.2 “Assessment of test results”. This report describes both, the evaluation of the results of the test executed within the link projects (chapter 2) and the HEADSTART methodology (chapter 3) as a whole. The outcome of HEADSTART D4.1 [1] together with the evaluation of the test results from the linked projects show that the methodology created by the HEADSTART project is suited to perform the safety assessment of a Connected and Automated Vehicle (CAV). The HEADSTART methodology wasn’t always implemented to its full extend in the linked projects. This was largely related to the fact that the linked projects are not part of the HEADSTART project itself and operate on their own. Together, however, the different linked projects make it possible to evaluate the different aspects of the HEADSTART methodology, including HEADSTART use cases, HEADSTART Key Enabling Technologies (KETs) and HEADSTART key user groups. The different lessons learned when applying the HEADSTART methodology for the assessment of different CAVs in the linked projects are documented in this and other HEADSTART deliverables, for example HEADSTART D4.1 [1].
In the second part of this report (Chapter 3) the HEADSTART methodology, that has been created over the course of the HEADSTART project, is evaluated. The evaluation showed that the building blocks of the HEADSTART methodology is generic and can be widely used. Dedicated settings or parameters for scenarios and testing might be required, however. The various variables depend on the specific function or vehicle under test (CAV). Depending on the capabilities, different Key Enabling Technologies (KETs) are relevant and need to be take into account in the assessment. Besides, also the use case and to which key user group it applies plays a role in selection of scenarios, tests and evaluation criteria. The universal approach of the HEADSTART methodology makes it possible to apply it to more complex CAVs and environments, where more scenarios are relevant, by scaling the number of scenarios.