The ESPLANADE project targeted the complex question of showing that an automated road vehicle is safe. This problem is significantly different from safety argumentation for manually driven vehicles. Since the automated driving system (ADS) has complete control of the vehicle when activated, part of its function must be to drive safely. There are several methodological problems that need to be mastered in order to find out how to perform safety argumentation for an ADS. The scope of this project was to provide methods to help solve these problems. The following topics related to safety assurance of an ADS were investigated: How to do safety analysis for Human-ADS interaction? How to perform risk assessment and define safety goals (top-level safety requirements)? How to determine operational capability and distribute decision in the ADS architecture? How to handle incomplete redundancy for sensor systems in the safety argumentation? How to ensure completeness and consistency in requirements refinement? The results include several novel methods as well as new application areas for existing methods. The ESPLANADE project ran from January 2017 to March 2020 with the partners Aptiv, Comentor, KTH, Qamcom, RISE, Semcon, Systemite, Veoneer, Volvo Cars, Volvo Technology, and Zenuity. 18 scientific papers were produced, of which 16 are at the time of writing published in academic peer-reviewed conferences or journals. Additionally, 13 deliverables in the form of project reports were written. This final report is a summary of the project results and contains excerpts from the deliverables.
@misc{EsplanadePublicReport_esplanade,
title = {ESPLANADE Project - Public Report},
author = {Warg, Fredrik},
year = {2020},
month = {06},
abstract = {The ESPLANADE project targeted the complex question of showing that an automated road vehicle is safe. This problem is significantly different from safety argumentation for manually driven vehicles. Since the automated driving system (ADS) has complete control of the vehicle when activated, part of its function must be to drive safely. There are several methodological problems that need to be mastered in order to find out how to perform safety argumentation for an ADS. The scope of this project was to provide methods to help solve these problems.
The following topics related to safety assurance of an ADS were investigated:
How to do safety analysis for Human-ADS interaction?
How to perform risk assessment and define safety goals (top-level safety requirements)?
How to determine operational capability and distribute decision in the ADS architecture?
How to handle incomplete redundancy for sensor systems in the safety argumentation?
How to ensure completeness and consistency in requirements refinement?
The results include several novel methods as well as new application areas for existing methods.
The ESPLANADE project ran from January 2017 to March 2020 with the partners Aptiv, Comentor, KTH, Qamcom, RISE, Semcon, Systemite, Veoneer, Volvo Cars, Volvo Technology, and Zenuity. 18 scientific papers were produced, of which 16 are at the time of writing published in academic peer-reviewed conferences or journals. Additionally, 13 deliverables in the form of project reports were written. This final report is a summary of the project results and contains excerpts from the deliverables.},
keywords = {automated driving systems, safety, risk assessment, operational design domain, human interaction safety analysis, requirements refinement},
howpublished = {Project partners/Vinnova},
url = {https://www.vinnova.se/globalassets/mikrosajter/ffi/dokument/slutrapporter-ffi/trafiksakerhet-och-automatiserade-fordon-rapporter/2016-04268engelska.pdf?cb=20201007132619},
note = {Publication data: https://warg.org/fredrik/publ/}
}