Criticality Metrics for Automated Driving: A Review and Suitability Analysis of the State of the Art

Abstract

The large-scale deployment of automated vehicles on public roads has the potential to vastly change the transportation modalities of today's society. Although this pursuit has been initiated decades ago, there still exist open challenges in reliably ensuring that such vehicles operate safely in open contexts. While functional safety is a wellestablished concept, the question of measuring the behavioral safety of a vehicle remains subject to research. One way to both objectively and computationally analyze traffic conflicts is the development and utilization of so-called criticality metrics. Contemporary approaches have leveraged the potential of criticality metrics in various applications related to automated driving, e.g. for computationally assessing the dynamic risk or filtering large data sets to build scenario catalogs. As a prerequisite to systematically chooseThe research leading to these results is funded by the German Federal Ministry for Economic Affairs and Energy within the projects 'VVM -Verification & Validation Methods for Automated Vehicles Level 4 and 5' and 'SET Level -Simulation-based Development and Testing of Automated Driving', based on a decision by the Parliament of the Federal Republic of Germany.

Towards Safe and Sustainable Autonomous Vehicles Using Environmentally-Friendly Criticality Metrics

“…Regarding this, to assess how critical a traffic situation is, the existent research works found in the literature focus on the use of so-called criticality metrics for automated driving [7,8]. However, because AVs are operating in a complex traffic environment where a high number of actors are present, such as AVs, non-AVs, and pedestrians, to name only a few, it is imperative to not only identify the suitable criticality metrics that can mitigate dangerous situations as it is currently done in the literature [7,8] but also to implement and evaluate them efficiently regarding their environmental impact as well. This is of high importance, especially when the transportation sector is known to be a key contributor to climate change, accounting for more than 35% of carbon dioxide emissions in the United States alone [9].…”

“…Therefore, in this paper, we present a mathematical analysis of 43 criticality metrics [7,8] to determine if they are well-defined as well as if they are working as intended within their scope to easily facilitate their selection for criticality assessment in the context of AV safety evaluation. Furthermore, due to recent emergent paradigms, such as Green AI [13], which encourage researchers to move towards more sustainable methods that are environmentally friendly and inclusive, we also propose several green metrics that are used to create a novel green-based criticality metric, which is suitable for evaluating a critical scenario not only regarding safety but also regarding the environmental impact in a car-following scenario.…”

“…Furthermore, from the perspective of AI training, we assess whether existing criticality metrics are suitable to serve as a component of the objective function, (which was already mentioned in [7] (Section 3.1.1)), for example, of the reward function in RL. To this end, the used criticality metrics must work as intended so that actions and, therefore, policies themselves that represent the desired behavior are flagged as optimal.…”

Using Ontologies for the Formalization and Recognition of Criticality for Automated Driving

“…Aspects of criticality are measured using criticality metrics: Definition III.2 (Criticality Metric). [4] A criticality metric is a function κ : S × R + → O that measures for a given traffic scene S ∈ S at a time t ∈ R + aspects of criticality on a predetermined scale of measurement O ⊆ R ∪ −∞, +∞. Scenario level criticality metrics extend this definition from scenes to scenarios.…”

“…For example, such reasoning allows for an automated assessment of the situational risk. In the area of automated driving, risk evaluations are often performed using criticality metrics [4], e.g. to guide decisions of the driving automation towards states of minimal risk or to derive relevant test cases within a safety case.…”