Closing the gaps: Complexity and uncertainty in the safety assurance and regulation of automated driving

Simon Burton; John Alexander McDermid

Closing the gaps: Complexity and uncertainty in the safety assurance and regulation of automated driving

Research output: Book/Report › Other report

Abstract

The increasing level of automation within an open context and use of artificial intelligence in cognitive cyber-physical Systems (CPS) is leading to emergent complexity and subsequently to uncertainties within the system assurance process. For example, in automated driving this is particularly true for the class of risks associated with the safety of the intended functionality (SOTIF) as described by the standard ISO 21448. In this report, we provide a definition of how complexity and uncertainty impacts the safety assurance of cognitive CPS. Based on this structured understanding of the problem, we propose an approach to managing the safety and regulating the deployment and operation of such systems in order to maintain an acceptable level of residual risk despite of, and with the intent of reducing, residual uncertainties. The approach includes criteria to guide decisions regarding the deployment and continuous assurance of the systems. The model used to structure these proposals includes a causal analysis of the factors impacting the complexity and resulting uncertainty (and, by extension, risk) that span the three layers of technical & human factors, management & operations and governance and regulation. These principles are generally applicable to a broad class of cognitive cyber-physical systems. However, in this report we focus on their application to automated driving systems.

Original language	English
Number of pages	24
Publication status	Published - 11 Jan 2023

Access to Document

https://www.york.ac.uk/media/assuring-autonomy/documents/Regulating_Safety_Complex_Systems.pdfLicence: Unspecified

Cite this

@book{97c87e591fdb4e13b1e7462070913afb,

title = "Closing the gaps: Complexity and uncertainty in the safety assurance and regulation of automated driving",

abstract = "The increasing level of automation within an open context and use of artificial intelligence in cognitive cyber-physical Systems (CPS) is leading to emergent complexity and subsequently to uncertainties within the system assurance process. For example, in automated driving this is particularly true for the class of risks associated with the safety of the intended functionality (SOTIF) as described by the standard ISO 21448. In this report, we provide a definition of how complexity and uncertainty impacts the safety assurance of cognitive CPS. Based on this structured understanding of the problem, we propose an approach to managing the safety and regulating the deployment and operation of such systems in order to maintain an acceptable level of residual risk despite of, and with the intent of reducing, residual uncertainties. The approach includes criteria to guide decisions regarding the deployment and continuous assurance of the systems. The model used to structure these proposals includes a causal analysis of the factors impacting the complexity and resulting uncertainty (and, by extension, risk) that span the three layers of technical & human factors, management & operations and governance and regulation. These principles are generally applicable to a broad class of cognitive cyber-physical systems. However, in this report we focus on their application to automated driving systems.",

author = "Simon Burton and McDermid, {John Alexander}",

year = "2023",

month = jan,

day = "11",

language = "English",

}

TY - BOOK

T1 - Closing the gaps: Complexity and uncertainty in the safety assurance and regulation of automated driving

AU - Burton, Simon

AU - McDermid, John Alexander

PY - 2023/1/11

Y1 - 2023/1/11

N2 - The increasing level of automation within an open context and use of artificial intelligence in cognitive cyber-physical Systems (CPS) is leading to emergent complexity and subsequently to uncertainties within the system assurance process. For example, in automated driving this is particularly true for the class of risks associated with the safety of the intended functionality (SOTIF) as described by the standard ISO 21448. In this report, we provide a definition of how complexity and uncertainty impacts the safety assurance of cognitive CPS. Based on this structured understanding of the problem, we propose an approach to managing the safety and regulating the deployment and operation of such systems in order to maintain an acceptable level of residual risk despite of, and with the intent of reducing, residual uncertainties. The approach includes criteria to guide decisions regarding the deployment and continuous assurance of the systems. The model used to structure these proposals includes a causal analysis of the factors impacting the complexity and resulting uncertainty (and, by extension, risk) that span the three layers of technical & human factors, management & operations and governance and regulation. These principles are generally applicable to a broad class of cognitive cyber-physical systems. However, in this report we focus on their application to automated driving systems.

AB - The increasing level of automation within an open context and use of artificial intelligence in cognitive cyber-physical Systems (CPS) is leading to emergent complexity and subsequently to uncertainties within the system assurance process. For example, in automated driving this is particularly true for the class of risks associated with the safety of the intended functionality (SOTIF) as described by the standard ISO 21448. In this report, we provide a definition of how complexity and uncertainty impacts the safety assurance of cognitive CPS. Based on this structured understanding of the problem, we propose an approach to managing the safety and regulating the deployment and operation of such systems in order to maintain an acceptable level of residual risk despite of, and with the intent of reducing, residual uncertainties. The approach includes criteria to guide decisions regarding the deployment and continuous assurance of the systems. The model used to structure these proposals includes a causal analysis of the factors impacting the complexity and resulting uncertainty (and, by extension, risk) that span the three layers of technical & human factors, management & operations and governance and regulation. These principles are generally applicable to a broad class of cognitive cyber-physical systems. However, in this report we focus on their application to automated driving systems.

M3 - Other report

BT - Closing the gaps: Complexity and uncertainty in the safety assurance and regulation of automated driving

ER -