ODDmeTOD – Safe Automated Driving Through Comparable ODD and TOD

About the Project

Automated vehicles are developed to operate under certain defined conditions. These conditions are described through the Automated vehicle's Operational Design Domain (ODD). To assess whether the vehicle can be used safely, the ODD needs to be compared with the actual traffic, infrastructure, environmental, and regulatory conditions that apply where the vehicle is to be deployed. These conditions are described as the Target Operational Domain (TOD).

Today, standards exist for structuring a vehicle's ODD, but corresponding established methods are lacking for defining TOD and systematically comparing the two domains. The assessment therefore risks becoming manual, time-consuming, and difficult to review.

Purpose and Objectives

The project's objective is to develop a PoC (Proof of Concept) of a methodological framework that translates societal and operational requirements into machine-readable TOD requirements. These shall be compared with the vehicle's ODD and form the basis for a traceable and verifiable chain of evidence for future approval.

The project focuses particularly on:

• developing a general method for defining a machine-readable TOD
• creating a harmonised and digital description of the vehicle's ODD
• developing algorithms for automatic comparison between ODD and TOD
• identifying gaps that may affect safety, operation, or regulatory compliance
• deriving representative and safety-critical scenarios
• defining minimum requirements for testing the selected Dynamic Driving Task (DDT)
• demonstrating the method through simulation and targeted physical testing
• producing material that can be used in standardisation and future approval processes.

Background and Needs

An automated vehicle's technical ability to drive in a certain type of environment does not automatically mean that it fulfils all requirements applicable to operate in that area. Road design, traffic regulations, weather conditions, traffic flows, other road users, and societal requirements may create conditions that lie outside the vehicle's ODD.

Existing standards, such as ISO 34503 and ASAM OpenODD, provide support for describing and structuring a vehicle's Operational Design Domain (ODD), including machine-readable representations of the conditions under which the system is intended to operate. However, there is no established method for deriving a Target Operational Domain (TOD) from the requirements of society, road infrastructure, and the operational environment. Consequently, there are no generally accepted methods for systematically comparing a TOD with a vehicle's ODD and identifying gaps, and also deriving and verifying that test scenarios provide sufficient coverage.This complicates safety assessment, regulatory review, and the scalable deployment of automated driving systems.

ODDmeTOD addresses this challenge by creating a transparent and repeatable chain from societal requirements and environmental data to domain comparison, scenario-based testing, and verifiable safety evidence

Demonstration in a Relevant Operational Environment

The project uses an industrial use case to develop and demonstrate the methods. A representative TOD is used to show how real road, traffic, environmental, and regulatory conditions can be described digitally and compared with an automated vehicle's ODD.

At the same time, the method shall be general, transferable, and independent of a specific geographical location. The results shall therefore be usable in more operational environments and for other automated transport systems.

The industrial use case is based on a logistics area in Torsvik, Jönköping, where the project can work with a coherent area rather than limiting the analysis to a single route.

Collaboration

The project brings together expertise in automated transport, safety assessment, standardisation, artificial intelligence, simulation, scenario generation, and physical testing.

The project partners are:

•  AstaZero, part of RISE
• Einride
• Applied Intuition
• University of Gothenburg
• Mid Sweden University

AstaZero coordinates the project. Einride contributes the industrial use case and the vehicle's operational requirements. Applied Intuition contributes platforms and expertise in simulation and digital validation. The University of Gothenburg and Mid Sweden University contribute research in areas including AI, scenario generation, software quality, and data-driven safety analysis.

ASAM and the Swedish Transport Agency participate as reference partners. Their perspectives strengthen the connection to international standardisation and future regulatory needs, respectively.

Expected Results

ODDmeTOD is expected to result in:

• a method for defining machine-readable TODs
• a digital and machine-readable ODD
• an open source algorithm for comparison between ODD and TOD
• methods for identifying and visualising relevant gaps
• a structure for a scenario database
• methods for scenario derivation and risk-based prioritisation
• minimum requirements for testing automated driving functions
• a visualisation of test coverage for independent review
• evidence for reliability in simulation-based verification
• recommendations and contributions to international standardisation

Expected Impacts

The project shall contribute to:

• safer and more transparent introduction of automated vehicles
• earlier identification of limitations and safety-related gaps
• more efficient development and verification processes
• better traceability between requirements, scenarios, tests, and results
• increased use of simulation in combination with targeted physical testing
• better foundations for independent safety assessment
• a foundation for larger future European research and innovation projects
   

Societal Benefit

ODDmeTOD contributes to traffic-safe automation by making it clearer under which conditions an automated vehicle can be used safely and when additional development, limitations, or testing are required.

The project supports:

• safer introduction of automated transport
• more consistent and reviewable safety assessments
• more efficient use of testing and development resources
• better collaboration between industry, research, test organisations, and authorities
• strengthened Swedish competitiveness in automated transport, verification, and standardisation

In the longer term, the methods may contribute to automated systems being scaled to new operational environments without loss of safety, traceability, or quality.