Towards building a safety case for Marine Unmanned Surface Vehicles: a Bayesian perspective

Marine Unmanned Surface Vehicles (MUSVs) are essential platforms for persistent and adaptable ocean monitoring and sampling. In order to operate these platforms in coastal areas or near oil and gas waters the MUSVs must meet statutorily and industry safety requirements. Given the novelty of these platforms, there is lack of evidence to support the claim that a given safety target can be met without any additional protection. Therefore, for safety critical operations, MUSVs require the implementation of a safety function. The development of a safety function must comply with IEC61508 safety standard, which requires a quantification of the safety integrity level. Compliance to IEC61508 is subject to subjective uncertainty. The nature of the technology in terms of mode of operation and the environment in which operates exacerbates this uncertainty. This paper presents a Bayesian belief network for formalizing the safety arguments underpinning MUSV compliance to IEC 615078 safety standard