Towards an Operational Design Domain Framework for Autonomous Navigation in Oil & Gas Facilities

Automated and autonomous robots present significant opportunities for operations in Oil & Gas facilities by reducing human exposure to dull, dirty and dangerous tasks in hazardous environments. So far, automated quadrupeds and crawlers have been widely trialed, but due to the complexity of the operation domain and high risks involved, scalable deployment has been limited. One of the main reasons for this is the lack of agreed tests and metrics in academia and industry for operational feasibility and safety of automated/autonomous navigation in Oil & Gas facilities. To tackle this, we believe that the industry can leverage the operational design domain (ODD) research conducted for autonomous road vehicles, which led to a significant number of standards and metrics to ensure that automated/autonomous decision-making is safe and reliable. Using these standards, many simulation and benchmark datasets have been developed to quantify the safety of autonomous decision-making in various conditions with dynamic agents. In this paper, a similar approach is taken to provide an initial taxonomy for operational design domain frameworks for Oil & Gas facilities. To achieve this, we adapted the ODD framework developed by NHTSA (National Highway Traffic Safety Association). It is believed that this can pave the way for building benchmarking tools and quantifiable metrics to accelerate safe deployment of fully autonomous ground-based robotic operations in Oil & Gas facilities.