Create
The Create block entails acquiring the necessary data and knowledge to create scenarios. This creation of scenarios takes place in external SCDBs, and is not covered in the scope of the SUNRISE project. Nevertheless, the Create block forms an integral part of the SUNRISE SAF.
Scenarios can be created using 2 different approaches: data-based and knowledge-based. The Data-Based approach extracts scenarios from existing sources (such as databases), whereas the Knowledge-Based approach generates scenarios from more abstract sources (such as expert knowledge). The Knowledge-Based approach does not require actual data collection, and is therefore often used to complement data-based approaches. It is considered especially valuable for identifying edge cases that may not appear in a Data-Based approach.
Examples of data-based scenario sources are:
- Accident Databases – These databases contain historical accident data that can be analyzed to extract concrete accident scenarios. They provide real-world evidence of safety-critical situations that have actually occurred on roads.
- Insurance Claim Records – This is considered an optimal source for collision and near-miss accident trends. Insurance data captures both actual accidents and near-miss events, providing valuable insights into safety-critical scenarios that may not be captured in official accident reports.
- Naturalistic Driving Data – Real-world driving data collected from vehicles equipped with sensors. This includes data from normal driving conditions collected over extensive distances covering highways, rural and urban roads. It represents the majority of driving conditions that a CCAM system would encounter.
- Drone Database – Provides naturalistic road user trajectory data, particularly useful for intersection scenarios. Drones offer a unique perspective for capturing traffic interactions and behaviors from above.
- Synthetic Simulation Data – While not strictly “real-world,” synthetic data from simulations can supplement real data by generating scenarios that may be rare or dangerous to collect in reality.
Examples of knowledge-based scenario sources are:
- Systems Theoretical Process Analysis (STPA) – STPA analyzes the characteristics of an ADS architecture to identify potential system hazards and failures. This systematic approach examines how system components interact and where failures might occur, creating scenarios based on theoretical system behavior rather than observed data.
- Formal Verification Methods – Highway code rules and traffic regulations can be used to frame each highway code rule as a hypothetical driving scenario with corresponding behavior and ODD (Operational Design Domain) elements. It focuses on maneuver parameters that are near the boundary of rule violations, producing scenarios that represent these potential violations.
- ODD-Based Generation – It combines required elements of the Operational Design Domain (ODD), and limits possible scenario combinations based on ODD restrictions, ensuring scenarios remain within the system’s intended operational boundaries.
- Standards and Regulatory Documents – Official standards and regulatory frameworks can be a relevant scenario source. This manual process contributes well-defined scenarios, based on established safety standards and regulatory requirements that systems must meet.
- Expert Knowledge – Domain experts and safety specialists can provide the basis for creation of scenarios based on their professional experience, safety expertise, and understanding of potential hazardous situations.
SAF Application Guidelines for the Create block are combined with those of the Format and Store blocks, and can be found on the Scenario page.