Defining Operational Domain and Specifying Operational Design Domains: Current Practices, Standards, and a Systematic Approach

Kurzfassung

The operation of Automated Vehicle (AV) systems is limited by their sensors, actuators, and design. As a result, AVs cannot safely operate within a wide range of environmental, scenery, and traffic conditions. Therefore, a subset of the Operational Domain (OD) shall be specified, within which the AV is designed to operate safely. This subset of OD is called the Operational Design Domain (ODD), which is specified using an ODD specification. Also, understanding the OD and ODD is crucial for effective scenario generation in Virtual Test Drives for AVs. Creating an ODD specification for a specific AV system is difficult due to the vast range of OD attributes and the wide range of parameters affecting the AV's operation. It is beneficial to start with a preliminary ODD specification that is created methodically and systematically rather than creating one from scratch through trial and error. This approach saves time and effort in the early stages of development. Also, it streamlines the development process and enhances the safety and functionality of AV systems. Our presentation begins with an overview of the latest advancements in research and standards in this field, demonstrating how formal representations of these concepts can enhance the development process. By leveraging the current best practices, we aim to bridge the gap between theoretical frameworks and practical applications, ensuring a robust foundation for AV testing and validation. Furthermore, we introduce our systematic approach to specifying an ODD, designed to address the challenges posed by diverse sensor configurations, complex operational environments, and varied driving assistant functionalities with specific regulations. Our methodology focuses on identifying and excluding non-safe areas within the operational domain, thereby minimizing risks and simplifying the testing process. This targeted strategy narrows the OD to a manageable subspace, ultimately saving time and effort in AV development and testing, and paving the way for safer and more efficient deployment of automated driving technologies.