Advanced Markov Modeling and Simulation for Safety Analysis of Autonomous Driving Functions up to SAE 5 for Development, Approval and Main Inspection

Häring, Ivo; Satsrisakul, Yupak; Finger, Jörg; Vogelbacher, Georg; Köpke, Corinna; Höflinger, Fabian; Gelhausen, Patrick

doi:10.3850/978-981-18-5183-4_r03-02-012-cd

Cited by 3 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another example is the failure probabil-ities provided for the lidar and the camera, for which the source publication of the data states that their values are arbitrary and might not represent the real world. An estimation of failure rates computed from those probabilities was performed by Häring et al 6 . However, since their study was focused on methods for safety assessment of AVs through Markov modeling, they did not need precise and/or accurate values and assumed a time frame of 10 years to convert the probabilities to rates, which may not reflect the experimental conditions of the source publications.…”

Section: Related Workmentioning

confidence: 99%

Components and Their Failure Rates in Autonomous Driving

Richter,

Walz,

Dhanani

et al. 2023

Proceeding of the 33rd European Safety and Reliability Conference

View full text Add to dashboard Cite

Autonomous driving has been among the most actively researched topics over the past decades. Today, automotive vehicles are already equipped with driving assistance systems with partial autonomous driving capabilities. Thus, the need for quantitative and qualitative assessment of automated driving functions becomes increasingly vital. The used hardware and software must undergo vigorous safety assessments with regard to reliability and safety. This must be done under careful consideration of driving scenarios and environmental conditions. The safety of the intended functionality (SOTIF) standard, which is developed under the corresponding ISO 21448 standard for road vehicles, lies at the center of these considerations. SOTIF deals with the question of how a target function needs to be specified, developed, verified, and validated so that it can be considered sufficiently safe. As a good starting point, we suggest regarding the individual failure probabilities for each of the components comprising the autonomous driving system. Based on the failure probabilities of each component, it is possible to make assumptions about the failure probability of the system as a whole and even identify possible deficiencies. In this contribution, we aim to identify the typical components needed for an autonomous vehicle (AV) and further provide a comprehensive overview of failure probabilities for said components. Certainly, it would go beyond the scope of this work to create a statistically firm data basis by individually testing all components until failure, especially when taking into consideration that the failure probabilities of each component vary over time and with environmental conditions. Instead, the relevant factors with regard to the typical failure modes are identified and relevant data is accumulated from publications that reflect the current state of the art.

show abstract

Section: Related Workmentioning

confidence: 99%

Components and Their Failure Rates in Autonomous Driving

Richter,

Walz,

Dhanani

et al. 2023

Proceeding of the 33rd European Safety and Reliability Conference

View full text Add to dashboard Cite

show abstract

“…Markov models are used in a wide range of domains including medicine success prediction (Sonnenberg and Beck 1993), predictive reliability assessment (Yosri et al 2021), and autonomous driving (AD) safety assessment (Nyberg 2018) (Bonderson 2018) (Heinrich, Plinke, and Hausschild 2017) (Häring et al 2022).…”

Section: Introductionmentioning

confidence: 99%

Dynamically Resolving and Abstracting Markov Models for System Resilience Analysis

Häring,

Sandela,

Walz

et al. 2023

Proceeding of the 33rd European Safety and Reliability Conference

View full text Add to dashboard Cite

Regarding the modeling of quasi-static systems with minor failures for failure prediction and maintenance, Markov models have shown to be very successful. Finite discrete state models can be considered as best practice in this domain, often even assumed to be homogeneous. The question arises if Markov models are also capable to model resilience of systems including major disruptions, where great fractions of the system and its functionality fail. To this end, analytical propositions are made that define model extensions. An initial scalable system is defined, including expected refinements and abstractions. In further phases, major disruptions occur. The disruptions can cause branching points opening routes to model extensions or abstractions. Also independent of disruptions, new states and transitions are introduced or merged for model granularity adoption. Overall system behavior can be interpreted in terms of system improvement with or without new system states or functionalities and corresponding transitions, reaching the ex-ante system state as before the disruption, reaching a deteriorated system state, or finally various degraded and failed overall system states. Definitions such as states, absorbing states and critical transitions are reinterpreted or extended to allow for dynamically resolving or abstracting the Markov model. The main results are extended definitions and derivations when compared to traditional Markov models. Based on the analytical expressions, an example is provided where the formalism could be applied with advantage for autonomous driving safety assessment by considering increasing or decreasing levels of resolution of subsystems or subfunctions.

show abstract