Glatzki, Felix Fabian (2023)
Methodology for Specifying and Testing Traffic Rule Compliance for Automated Driving.
Technische Universität Darmstadt
doi: 10.26083/tuprints-00023569
Dissertation, Erstveröffentlichung, Verlagsversion
Kurzbeschreibung (Abstract)
The introduction of highly-automated driving functions promises to increase safety and comfort, but the safety validation remains an unsolved challenge. Here, the requirement is that the introduction does not reduce safety on public roads. This dissertation addresses one major aspect of road safety: traffic rule compliance. Even an automated vehicle must comply with existing traffic rules. The developed method enables automated testing of traffic rule compliance of automated driving functions.
In the first part of the thesis, the state of the art for describing and formalizing behavioral rules is analyzed. A special challenge is posed by the different traffic rules depending on the traffic region. With existing approaches, a separate description and formalization of the behavior rules is necessary for each traffic region or even for individual traffic areas. This shows the necessity to develop new approaches for the abstraction and transferability of the behavioral rules in order to reduce the effort of testing and ensuring traffic rule compliance. The rule compliance criteria are to be integrated into the behavior specification within the functional specification. The objective of this thesis is to develop a method to formalize the limits of traffic rule compliance, based on which fail criteria for system testing are defined and applied.
For this purpose, existing traffic rules are analyzed as a basis to identify which behavior constraints are imposed by the static traffic environment. Based on this, a semantic description that is transferable between traffic domains and that links the boundaries of traffic rule compliance to the static traffic environment is developed. The method involves deriving behavioral attributes from which the semantic behavior description is constructed. These behavioral attributes construct the behavior space that describes the boundaries of legally allowed behavior. Furthermore, methods for automated derivation of behavioral attributes from high definition maps are developed, thus extracting the behavioral requirement from an operational design domain. It is investigated which functionalities an automated vehicle has to provide to comply with the behavioral attributes. The attributes are then formalized to obtain quantifiable failure criteria of traffic rule compliance that can be used in automated testing. Finally, building on the state of the art, a test strategy for validating traffic rule conformance is presented. The explicit availability of the behavioral limits results in an advantage in the influence analysis of possible parameters for these tests.
Finally, the developed method is applied to existing map material and to test drives with an automated vehicle prototype in order to investigate the practical applicability of the approach as well as the resulting gain in knowledge about traffic rule compliance testing. The developed approach allows to derive the behavioral specification with respect to traffic rule conformance as an essential part of the functional specification independent of the application domain. It is proven that the approach is able to test the traffic rule conformance of an automated vehicle in different test scenarios within an application domain. By applying the developed methodology, it was possible to identify defects in the investigated test vehicle with respect to rule understanding and compliance.
Typ des Eintrags: | Dissertation | ||||
---|---|---|---|---|---|
Erschienen: | 2023 | ||||
Autor(en): | Glatzki, Felix Fabian | ||||
Art des Eintrags: | Erstveröffentlichung | ||||
Titel: | Methodology for Specifying and Testing Traffic Rule Compliance for Automated Driving | ||||
Sprache: | Englisch | ||||
Referenten: | Winner, Prof. Dr. Hermann ; Czarnecki, Prof. Dr. Krzysztof | ||||
Publikationsjahr: | 2023 | ||||
Ort: | Darmstadt | ||||
Kollation: | XIX, 147 Seiten | ||||
Datum der mündlichen Prüfung: | 17 Januar 2023 | ||||
DOI: | 10.26083/tuprints-00023569 | ||||
URL / URN: | https://tuprints.ulb.tu-darmstadt.de/23569 | ||||
Kurzbeschreibung (Abstract): | The introduction of highly-automated driving functions promises to increase safety and comfort, but the safety validation remains an unsolved challenge. Here, the requirement is that the introduction does not reduce safety on public roads. This dissertation addresses one major aspect of road safety: traffic rule compliance. Even an automated vehicle must comply with existing traffic rules. The developed method enables automated testing of traffic rule compliance of automated driving functions. In the first part of the thesis, the state of the art for describing and formalizing behavioral rules is analyzed. A special challenge is posed by the different traffic rules depending on the traffic region. With existing approaches, a separate description and formalization of the behavior rules is necessary for each traffic region or even for individual traffic areas. This shows the necessity to develop new approaches for the abstraction and transferability of the behavioral rules in order to reduce the effort of testing and ensuring traffic rule compliance. The rule compliance criteria are to be integrated into the behavior specification within the functional specification. The objective of this thesis is to develop a method to formalize the limits of traffic rule compliance, based on which fail criteria for system testing are defined and applied. For this purpose, existing traffic rules are analyzed as a basis to identify which behavior constraints are imposed by the static traffic environment. Based on this, a semantic description that is transferable between traffic domains and that links the boundaries of traffic rule compliance to the static traffic environment is developed. The method involves deriving behavioral attributes from which the semantic behavior description is constructed. These behavioral attributes construct the behavior space that describes the boundaries of legally allowed behavior. Furthermore, methods for automated derivation of behavioral attributes from high definition maps are developed, thus extracting the behavioral requirement from an operational design domain. It is investigated which functionalities an automated vehicle has to provide to comply with the behavioral attributes. The attributes are then formalized to obtain quantifiable failure criteria of traffic rule compliance that can be used in automated testing. Finally, building on the state of the art, a test strategy for validating traffic rule conformance is presented. The explicit availability of the behavioral limits results in an advantage in the influence analysis of possible parameters for these tests. Finally, the developed method is applied to existing map material and to test drives with an automated vehicle prototype in order to investigate the practical applicability of the approach as well as the resulting gain in knowledge about traffic rule compliance testing. The developed approach allows to derive the behavioral specification with respect to traffic rule conformance as an essential part of the functional specification independent of the application domain. It is proven that the approach is able to test the traffic rule conformance of an automated vehicle in different test scenarios within an application domain. By applying the developed methodology, it was possible to identify defects in the investigated test vehicle with respect to rule understanding and compliance. |
||||
Alternatives oder übersetztes Abstract: |
|
||||
Status: | Verlagsversion | ||||
URN: | urn:nbn:de:tuda-tuprints-235692 | ||||
Sachgruppe der Dewey Dezimalklassifikatin (DDC): | 600 Technik, Medizin, angewandte Wissenschaften > 620 Ingenieurwissenschaften und Maschinenbau | ||||
Fachbereich(e)/-gebiet(e): | 16 Fachbereich Maschinenbau 16 Fachbereich Maschinenbau > Fachgebiet Fahrzeugtechnik (FZD) 16 Fachbereich Maschinenbau > Fachgebiet Fahrzeugtechnik (FZD) > Fahrerassistenzssysteme 16 Fachbereich Maschinenbau > Fachgebiet Fahrzeugtechnik (FZD) > Sicherheit 16 Fachbereich Maschinenbau > Fachgebiet Fahrzeugtechnik (FZD) > Testverfahren |
||||
Hinterlegungsdatum: | 21 Apr 2023 12:09 | ||||
Letzte Änderung: | 24 Apr 2023 05:14 | ||||
PPN: | |||||
Referenten: | Winner, Prof. Dr. Hermann ; Czarnecki, Prof. Dr. Krzysztof | ||||
Datum der mündlichen Prüfung / Verteidigung / mdl. Prüfung: | 17 Januar 2023 | ||||
Export: | |||||
Suche nach Titel in: | TUfind oder in Google |
Frage zum Eintrag |
Optionen (nur für Redakteure)
Redaktionelle Details anzeigen |