Towards an ontological framework for environmental survey hazard analysis of autonomous systems

Caleb-Solly, Praminda; Harper, Christopher

Abstract

This paper presents current progress in the development of Environmental Survey Hazard Analysis (ESHA), a method of preliminary hazard identification aimed at autonomous system application problems. In addition to performing their design mission, autonomous systems must be capable of reliable and predictable behaviour in their environments, particularly when facing potential hazards that are not explicitly included in their design specifications (’non-mission’ tasks).
ESHA differs from conventional hazard identification methods in that its scope explicitly covers the identification of non-mission interactions between a system and its environment and any associated hazards. Although of general use as a safety analysis technique, ESHA has been designed primarily to support a ”so far as is reasonably practicable” (SFAIRP) style of safety argument. However, early versions of the method were based on informal models, and therefore provided only weak support. This paper reviews the development of a formal ontological framework for ESHA, intended to provide much stronger basis for arguing the completeness and consistency of analyses.

• Presentation Conference Type: Conference Paper (published)
• Conference Name: Artificial Intelligence Safety 2021 (SafeAI 2021)
• Start Date: Feb 8, 2021
• End Date: Feb 8, 2021
• Acceptance Date: Dec 5, 2020
• Online Publication Date: Feb 8, 2021
• Publication Date: Feb 8, 2021
• Deposit Date: Mar 29, 2021
• Publicly Available Date: Mar 31, 2021
• Series ISSN: 1613-0073
• Keywords: safety, hazard analysis
• Public URL: https://uwe-repository.worktribe.com/output/7240303
• Publisher URL: http://ceur-ws.org/Vol-2808/Paper_41.pdf