- AutorIn
- Clemens Markus Hruschka TU Bergakademie Freiberg#Volkswagen AG
- Titel
- Risk assessment for integral safety in operational motion planning of automated driving
- Zitierfähige Url:
- https://nbn-resolving.org/urn:nbn:de:bsz:105-qucosa2-767559
- Datum der Einreichung
- 09.04.2021
- Datum der Verteidigung
- 01.09.2021
- Abstract (EN)
- New automated vehicles have the chance of high improvements to road safety. Nevertheless, from today's perspective, accidents will always be a part of future mobility. Following the “Vision Zero”, this thesis proposes the quantification of the driving situation's criticality as the basis to intervene by newly integrated safety systems. In the example application of trajectory planning, a continuous, real-time, risk-based criticality measure is used to consider uncertainties by collision probabilities as well as technical accident severities. As result, a smooth transition between preventative driving, collision avoidance, and collision mitigation including impact point localization is enabled and shown in fleet data analyses, simulations, and real test drives. The feasibility in automated driving is shown with currently available test equipment on the testing ground. Systematic analyses show an improvement of 20-30 % technical accident severity with respect to the underlying scenarios. That means up to one-third less injury probability for the vehicle occupants. In conclusion, predicting the risk preventively has a high chance to increase the road safety and thus to take the “Vision Zero” one step further.
- Verweis
- Uncertainty-adaptive risk based motion planning in automated driving / Hruschka, C. M., Schmidt, M. and Zug, S.
In: IEEE International Conference on Vehicular Electronics and Safety (ICVES) (2019)
DOI: 10.1109/ICVES.2019.8906305 - Risk assessment for integral safety in automated driving“, In: IEEE International Conference on Intelligent Autonomous Systems (ICoIAS) / Hruschka, C. M., Töpfer, D. and Zug, S.
In: IEEE International Conference on Intelligent Autonomous Systems (ICoIAS) (2019), pp. 102–109
DOI: 10.1109/ICoIAS.2019.00025 - Verhaltensentscheidung auf der Grundlage einer kontinuierlichen Risikobewertung als ein Schlüssel zum sicheren automatisierten Fahren / Gerhardus, M., Hruschka, C. M., Schmidt, M. et al.
In: Assistenzsysteme, Automatisierungssysteme und eingebettete Systeme (AAET) (2018) - Verfahren und Vorrichtung zum Bestimmen einer Kollisionswahrscheinlichkeit eines Fahrzeugs mit einem Objekt / Hruschka, C. M., Schmidt, M., Töpfer, D., Volkswagen AG
Patentschrift DE102018211514B4 - Verfahren und Vorrichtung zum Bestimmen einer Kollisionswahrscheinlichkeit eines Fahrzeugs mit einem Objekt / Hruschka, C. M., Schmidt, M., Töpfer, D., Volkswagen AG
Patentschrift DE102018211513A1 - Control of an autonomous or semi-autonomous vehicle / Hruschka, C. M., Schmidt, M., Töpfer, D., Volkswagen AG
Patentschrift DE102019212666A1 - Freie Schlagwörter (DE)
- Automatisierte Fahren, Integrale Sicherheit, Kollisionsvermeidung durch Ausweichen und Bremsen, Kollisionsmitigation durch Lenken und Bremsen, Kritikalitätsmetrik, Risikobewertung, Kollisionswahrscheinlichkeit, Schwereprädiktion, Unfallschwere, Bewegungsplanung, Trajektorienplanung, Kollisionsvermeidung, Kollisionsmitigation, Trefferlagenoptimierung, Unsicherheitsbeachtung
- Freie Schlagwörter (EN)
- automated driving, integral safety, collision avoidance by braking and swerving, collision mitigation by steering and braking, criticality measure, risk assessment, collision probability, severity prediction, accident severity, motion planning, collision avoidance, collision mitigation, impact point localisation, uncertainty consideration
- Klassifikation (DDC)
- 004
- Normschlagwörter (GND)
- Autonomes Fahrzeug
- Bahnplanung
- Unfallrisiko
- Verkehrsunfall
- Unfallverhütung
- Schaden
- Verletzung
- Minimierung
- GutachterIn
- Prof. Dr. Sebastian Zug
- Prof. Dr. Marcus Baum
- BetreuerIn Hochschule / Universität
- Prof. Dr. Sebastian Zug
- Den akademischen Grad verleihende / prüfende Institution
- TU Bergakademie Freiberg, Freiberg
- Version / Begutachtungsstatus
- publizierte Version / Verlagsversion
- URN Qucosa
- urn:nbn:de:bsz:105-qucosa2-767559
- Veröffentlichungsdatum Qucosa
- 14.01.2022
- Dokumenttyp
- Dissertation
- Sprache des Dokumentes
- Englisch
- Lizenz / Rechtehinweis
- CC BY 4.0
- Inhaltsverzeichnis
Abstract Acknowledgements Contents Nomenclature 1.1 Background 1.2 Problem statement and research question 1.3 Contribution 2 Fundamentals and relatedWork 2.1 Integral safety 2.1.1 Integral applications 2.1.2 Accident Severity 2.1.2.1 Severity measures 2.1.2.2 Severity data bases 2.1.2.3 Severity estimation 2.1.3 Risk assessment in the driving process 2.1.3.1 Uncertainty consideration 2.1.3.2 Risk as a measure 2.1.3.3 Criticality measures in automated driving functions 2.2 Operational motion planning 2.2.1 Performance of a driving function 2.2.1.1 Terms related to scenarios 2.2.1.2 Evaluation and approval of an automated driving function 2.2.2 Driving function architecture 2.2.2.1 Architecture 2.2.2.2 Planner 2.2.2.3 Reference planner 2.2.3 Ethical issues 3 Risk assessment 3.1 Environment model 3.2 Risk as expected value 3.3 Collision probability and most probable collision configuration 4 Accident severity prediction 4.1 Mathematical preliminaries 4.1.1 Methodical approach 4.1.2 Output definition for pedestrian collisions 4.1.3 Output definition for vehicle collisions 4.2 Prediction models 4.2.1 Eccentric impact model 4.2.2 Centric impact model 4.2.3 Multi-body system 4.2.4 Feedforward neural network 4.2.5 Random forest regression 4.3 Parameterisation 4.3.1 Reference database 4.3.2 Training strategy 4.3.3 Model evaluation 5 Risk based motion planning 5.1 Ego vehicle dynamic 5.2 Reward function 5.3 Tuning of the driving function 5.3.1 Tuning strategy 5.3.2 Tuning scenarios 5.3.3 Tuning results 6 Evaluation of the risk based driving function 6.1 Evaluation strategy 6.2 Evaluation scenarios 6.3 Test setup and simulation environment 6.4 Subsequent risk assessment of fleet data 6.4.1 GIDAS accident database 6.4.2 Fleet data Hamburg 6.5 Uncertainty-adaptive driving 6.6 Mitigation application 6.6.1 Real test drives on proving ground 6.6.2 Driving performance in simulation 7 Conclusion and Prospects References List of Tables List of Figures A Extension to the tuning process