Accelerated Modifications using Physics-substantiated Safety Arguments (AMPS)

A new approach is proposed for Accelerated Modifications using Physics-substantiated Safety Arguments (AMPS). We leverage evidence-based certification methodology and Safety Arguments (SAs) to map modifications to system architecture and behaviors. We integrate formal methods to build SAs that trace the impact of modifications through complex system interactions. AMPS connects SAs to critical design cases and system physics. The links with design cases maps SAs to engineering models and implements model-based recertification. The link with physics ensures SA sufficiency and overcomes confirmation/cognitive biases. AMPS includes tools to analyze SAs, identify gaps and plan recertification. AMPS integrates into the DoD planning and acquisition cycle.We will leverage our experience with SA modeling user interfaces and data structures to design a cross-enterprise certification portal. Stakeholders will use the portal to generate SAs, map them to certification requirements and safety evidence. We will implement a deep learning text analytics subsystem to automatically identify certification requirements from thousands of pages of regulations. When information required to builds SAs is embedded in design documents, the deep learning subsystem will automatically identify pertinent information. AMPS will automatically pull data from engineering tools such as DOORS to support SAs. AMPS will drive SA reuse and Digital Twin.