From Isolated to Distributed: Specifying, Validating and Verifying Safety-Critical CPS

Cyber-Physical Systems (CPS) integrate computational and physical processes and interact with the real world in complex and highly dynamic ways. These systems face demanding challenges, particularly in safety-critical domains such as automotive, healthcare, and maritime applications. However, most are still implemented as standalone devices, which makes them difficult to upgrade, extend, and maintain. The transition from isolated to distributed systems is especially challenging for safety-critical CPS (sCPS), as it requires a high level of trustworthiness and reliability while also ensuring compliance with regulatory requirements. This paper addresses a key challenge for established industries: transforming isolated sCPS into distributed systems spanning the edge-cloud continuum. It presents a methodology that begins with a trusted, standalone device and distributes selected system functions across edge and cloud tiers, while preserving its trustworthiness and functionality. The methodology supports a structured, stepwise transformation process based on early evaluation and incremental concretisation of requirements. Validation and verification activities systematically address cross-cutting concerns such as safety, security, performance, privacy, and regulatory compliance. The approach was developed in the context of the EU TRANSACT project (https://transact-ecsel.eu) and builds on the TRANSACT architecture-a universal, distributed solution architecture for safety-critical CPS.