Criticality Metrics for Automated Driving: A Review and Suitability Analysis of the State of the Art
DOI: 10.1007/s11831-022-09788-7
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This paper addresses the challenge of selecting appropriate criticality metrics for the verification, validation, and implementation of automated driving systems (AVs). While functional safety is well-established, measuring the behavioral safety of AVs in open, unstructured environments remains an open research problem. Criticality metrics serve as computational tools to quantify traffic conflicts and dynamic risk, enabling objective safety assessments. However, the diversity of existing metrics and their varying properties makes it difficult for practitioners to identify which metrics are suitable for specific applications. The authors aim to provide a systematic method for selecting metrics that align with application requirements, thereby ensuring efficient, valid, and reliable safety measurements. The study employs a review and suitability analysis methodology. First, the authors categorize common applications of criticality metrics along the automotive V-model, dividing them into implementation tasks (e.g., objective functions for planning, run-time monitoring, and minimal risk maneuvers) and verification/validation tasks (e.g., requirement elicitation, scenario classification, data-driven scenario elicitation, testing, and safety argumentation). Based on these applications, the authors derive specific properties and requirements for criticality metrics, such as sensitivity, validity, and computational efficiency. They then conduct an extensive review of the state of the art, unifying the description of various metrics and evaluating their capabilities against the derived properties. Finally, the authors propose a suitability analysis framework, an expert-based process that maps application requirements to metric capabilities, and demonstrate its utility through an exemplary execution. The findings highlight that no single metric is universally suitable for all automated driving contexts. For instance, the Time To Collision (TTC) metric is effective for car-following scenarios but lacks validity in intersection scenarios due to its reliance on single-point kinematics. The review reveals that metrics vary significantly in their ability to detect criticality across different scenario classes, necessitating the use of multiple metrics in combination to achieve adequate coverage. The proposed suitability analysis provides a structured blueprint for practitioners to identify potent sets of metrics for specific applications, addressing the gap between metric capabilities and application needs. The results of the review and the suitability analysis are made available in an open repository to facilitate community contributions and continuous refinement. The significance of this work lies in its contribution to the trustworthiness and safe deployment of automated vehicles. By providing a unified review and a methodical tool for metric selection, the paper enables more rigorous safety argumentations and efficient testing processes. It bridges the gap between theoretical metric properties and practical application requirements, offering a scalable approach for the industry to validate AV safety. This framework supports the development of robust safety mechanisms by ensuring that the computational tools used to assess criticality are well-suited to the specific risks and scenarios encountered by automated vehicles.
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | OpenAlex-citations | — | — | 1 | 2026-06-18 |
| archive | success | unpaywall | — | — | 2 | 2026-06-25 |
| extract | success | cached | — | — | 2 | 2026-06-26 |
| clean | success | clean | — | — | 1 | 2026-06-18 |
| chunk | success | chunk | — | — | 1 | 2026-06-18 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-18 |
| promote | success | — | — | — | 1 | 2026-06-18 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-26 |
| tag | success | vector_similarity | — | — | 6 | 2026-06-18 |
| verify | success | — | — | — | 1 | 2026-06-26 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Empirical Findings: crash risk outcomes
- Methodological Resource: validation psychometrics, metric or index