RAS safety framework

There exist several emerging approaches for systematic safety assurance of robotics and autonomous systems. These range from implementation of processes and safety cases, to low-level analysis of particular autonomous functions. Even with these advancements, there remains a paucity of guidance of how to identify, capture and communicate hazards that emerge as a result of autonomous system operation. This paper presents the RAS Safety Framework – an end-to-end assurance process that makes use of digital twins and constraint rules to represent safety claims and provide evidence for them. The RAS Safety Framework represents a fundamental shift in paradigm of assuring ‘behavioural hazards’. The approach is applied to a case study to demonstrate the utility and benefit of having a structured, traceable safety case and digital twin for complex autonomous systems.