The complexity of the operating environment and required technologies for highly automated driving is unprecedented. A different type of threat to safe operation besides the fault-error-failure model by Laprie et al. arises in the form of performance limitations. We propose a system theoretic approach to handle these and derive a taxonomy based on uncertainty, i.e. lack of knowledge, as a root cause. Uncertainty is a threat to the dependability of a system, as it limits our ability to assess its dependability properties. We distinguish uncertainties by aleatory (inherent to probabilistic models), epistemic (lack of model parameter knowledge) and ontological (incompleteness of models) in order to determine strategies and methods to cope with them. Analogous to the taxonomy of Laprie et al. we cluster methods into uncertainty prevention (use of elements with well-known behavior, avoiding architectures prone to emergent behavior, restriction of operational design domain, etc.), uncertainty removal (during design time by design of experiment, etc. and after release by field observation, continuous updates, etc.), uncertainty tolerance (use of redundant architectures with diverse uncertainties, uncertainty aware deep learning, etc.) and uncertainty forecasting (estimation of residual uncertainty, etc.).

Highly automated driving (HAD) vehicles are complex systems operating in an open context. Complexity of these systems as well as limitations and insufficiencies in sensing and understanding the open context may result in unsafe and uncertain behavior. The safety critical nature of the HAD vehicles demands to model limitations, insufficiencies and triggering conditions to argue safe behavior.Standardization activities such as ISO/PAS 21448 provide guidelines on the safety of the intended functionality (SOTIF) and focus on the performance limitations and triggering conditions. Although, SOTIF provides a non-exhaustive list of scenario factors that may serve as a starting point to identify and analyze performance limitations and triggering conditions, yet no concrete methodology is provided to model these factors.We propose a novel methodology to model triggering conditions and performance limitations in a scene to assess SOTIF. We utilize Bayesian network (BN) in this regard. The experts provide the BN structure and conditional belief tables are learned using the maximum likelihood estimator. We provide performance limitation maps (PLMs) and conditional performance limitation maps (CPLMs), given a scene. As a case study, we provide PLMs and CPLMs of LIDAR in a defined scene using real world data.

No abstract

Highly automated driving (HAD) vehicles are complex systems operating in an open context. Performance limitations originating from sensing and understanding the open context under triggering conditions may result in unsafe behavior, thus, need to be identified and modeled. This aspect of safety is also discussed in standardization activities such as ISO 21448, safety of the intended functionality (SOTIF). Although SOTIF provides a non-exhaustive list of scenario factors to identify and analyze performance limitations under triggering conditions, no concrete methodology is yet provided to identify novel triggering conditions.We propose a methodology to identify and model novel triggering conditions in a scene in order to assess SOTIF using Bayesian network (BN) and p-value hypothesis testing. The experts provide the initial BN structure while the conditional belief tables (CBTs) are learned using dataset. P-value hypothesis testing is used to identify the relevant subset of scenes. These scenes are then analyzed by experts who provide potential triggering conditions present in the scenes. The novel triggering conditions are modeled in the BN and retested. As a case study, we provide p-value hypothesis testing of BN of LIDAR using real world data.

No abstract