Towards a Software Component to Perform Situation-Aware Dynamic Risk Assessment for Autonomous Vehicles

Reich, Jan; Wellstein, Marc; Sorokos, Ioannis; Oboril, Fabian; Scholl, Kay-Ulrich

doi:10.1007/978-3-030-86507-8_1

Cited by 10 publications

(4 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the context of autonomous aviation systems, [10] proposes a method for quantifying system assurance using perception component uncertainty and dynamic BNs. For autonomous vehicles, [25] offers a framework for dynamic risk assessment, using BNs to predict the behavior intents of other traffic participants. Unlike these works, this paper considers uncertainty from Bayesian deep learning components beyond perception.…”

Section: Uncertainty-based Dependability Frameworkmentioning

confidence: 99%

Towards Dependable Autonomous Systems Based on Bayesian Deep Learning Components

Arnez,

Espinoza,

Radermacher

et al. 2023

Preprint

View full text Add to dashboard Cite

As autonomous systems increasingly rely on Deep Neural Networks (DNN) to implement the navigation pipeline functions, uncertainty estimation methods have become paramount for estimating confidence in DNN predictions. Bayesian Deep Learning (BDL) offers a principled approach to model uncertainties in DNNs. However, in DNN-based systems, not all the components use uncertainty estimation methods and typically ignore the uncertainty propagation between them. This paper provides a method that considers the uncertainty and the interaction between BDL components to capture the overall system uncertainty. We study the effect of uncertainty propagation in a BDL-based system for autonomous aerial navigation. Experiments show that our approach allows us to capture useful uncertainty estimates while slightly improving the system's performance in its final task. In addition, we discuss the benefits, challenges, and implications of adopting BDL to build dependable autonomous systems.

show abstract

Section: Uncertainty-based Dependability Frameworkmentioning

confidence: 99%

Towards Dependable Autonomous Systems Based on Bayesian Deep Learning Components

Arnez,

Espinoza,

Radermacher

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Dynamic Risk Assessment (DRA) relies on situation awareness to support run-time decision-making of an ADS, e.g. [31,124,127,128]. In addition to situation awareness, DRA capabilities need to connect a given situation to the safety requirements of the system, or at least to some kind of (quantitative) risk measurement.…”

Section: Dynamic Risk Assessmentmentioning

confidence: 99%

“…The notion of precautionary safety could also be merged with a framework for DRA in order to achieve a dynamic adaptation of the policy, based on the current risk levels of the ADS, which could help to achieve an even more performant system. This is partly exemplified in [127], but with the main difference that the requirements on the ADS are not posted as quantitative elements. However, the risks are dynamically estimated in [127], including a probabilistic formulation of the uncertainties, which is the same as suggested in [12] and [13].…”

Section: Precautionary Safetymentioning

confidence: 99%

“…This is partly exemplified in [127], but with the main difference that the requirements on the ADS are not posted as quantitative elements. However, the risks are dynamically estimated in [127], including a probabilistic formulation of the uncertainties, which is the same as suggested in [12] and [13]. For example, the jay-walking avoidance use case analysed by Rodrigues de Campos et al presents a very crude way of DRA, where the considered two different road types are associated with different exposure levels.…”

Section: Precautionary Safetymentioning

confidence: 99%

See 1 more Smart Citation

Holistic Perspectives on Safety of Automated Driving Systems - Methods for Provision of Evidence

Gyllenhammar¹,

Campos²,

Törngren³

2022

Preprint

View full text Add to dashboard Cite

<p>In recent years, the enormous investments in Automated Driving Systems (ADSs) have distinctly advanced ADS technologies. Despite promises made by several high profile auto-makers, it has however become clear that the challenges involved for deploying ADS have been drastically underestimated. This paper focuses on the challenge of providing sufficient evidence to support the safety claims of ADSs. The provisioning of such evidence clearly relates both to technical maturity of ADS systems (including actual experiences from deploying such systems), and on the development of methodologies for reasoning about ADS safety claims. Contrary to previous generations of automotive systems, common design, development, verification and validation methods for safety critical systems do not suffice to cope with the increased complexity and operational uncertainties of an ADS. Therefore, the aim of this paper is to provide an understanding of existing methods focusing on the development of a safe ADS and, most importantly, identifying the associated challenges and gaps. </p> <p>We present eight challenges, collectively distinguishing ADSs from safety critical systems in general, and discuss the existing methods in the light of these eight challenges. Based on this discussion, a set of research gaps are identified.</p>

show abstract

Structuring Research Related to Dynamic Risk Management for Autonomous Systems

Adler,

Reich,

Hawkins

2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Towards a Software Component to Perform Situation-Aware Dynamic Risk Assessment for Autonomous Vehicles

Cited by 10 publications

References 5 publications

Towards Dependable Autonomous Systems Based on Bayesian Deep Learning Components

Towards Dependable Autonomous Systems Based on Bayesian Deep Learning Components

Holistic Perspectives on Safety of Automated Driving Systems - Methods for Provision of Evidence

Structuring Research Related to Dynamic Risk Management for Autonomous Systems

Contact Info

Product

Resources

About