Defining Operational Design Domain for Autonomous Systems: A Domain-Agnostic and Risk-Based Approach

The integration of Artificial Intelligence (AI) into industrial systems with high levels of automation has introduced significant uncertainty and complexity. In particular, work by the automotive industry, on autonomous vehicles has led to the emergence of the Operational Design Domain (ODD) concept, which delineates the expected operating domain of such vehicles, departing from conventional automotive Use Case-based approaches. However, this ODD's automotive-centric approach has hindered its broader application, lacking the comprehensive guidance on the system engineering methodologies required for its definition. This paper presents a domain-agnostic definition of ODD, grounded in established system frameworks and emphasizing a systemic risk-based engineering to make it applicable to multiple domains. A case study from the maritime domain illustrates the benefits and applicability of the proposed methodology. By providing a systematic framework, this research facilitates the adoption of ODD principles beyond the automotive sector, fostering the development of AI-based products and services across diverse industrial domains. The ODD represents a key aspect of systems engineering for autonomous systems, integrating considerations of technology, environment, regulation, and user expectations.