Scenario Description Language for Automated Driving Systems: A Two Level Abstraction Approach

Abstract

Please refer to published version for the most recent bibliographic citation information. If a published version is known of, the repository item page linked to above, will contain details on accessing it.

Analyzing Real-world Accidents for Test Scenario Generation for Automated Vehicles

“…In this paper, we discuss the test scenario generation for ADSs using accident analysis of the UK's accident database (STATS19), which helped identify parameters and context contributing to high severity (including fatal) accidents. Subsequently, the test scenarios were formatted in a scenario definition language [15], and store them in the UK's National Scenario Database. This paper is organized in six sections.…”

Test Framework for Automatic Test Case Generation and Execution Aimed at Developing Trustworthy AVs from Both Verifiability and Certifiability Aspects

“…After scenario generation, the scenario content needs to be represented in a human and machine-readable format. A two-abstraction approach for scenario description [6] was used in the OmniCAV project which meet both regulatory and development needs of ADS testing. The structured scenarios will then be stored in a scenario database for storage, sharing, analysis and query purposes.…”

“…ISO22737 has been developed for low-speed automated driving systems (LSAD) and the EuroNCAP provides a set of testing scenarios for the safety assurance of vehicles. An example of EuroNCAP scenario converted into logical scenario was previously illustrated in this paper section IV-B [6]. Option 8 includes the scenarios that occur during real world trials and deployments.…”

“…After generating the scenario content, an adequate scenario description language (SDL) is used to represent the content and enable its sharing and execution. A two-level abstraction approach of SDL, as depicted in Figure 5, has been previously developed and published as part of the OmniCAV project [6]. It was developed after analyzing the inputs received from various stakeholders such as AV developers, test engineers, regulators.…”

“…This led to the shift from distance-based approach to a scenario-based approach for ADS safety assurance. Various studies have been published on different aspects of the scenario-based testing workflow such as scenario generation [3][4] [5], scenario description format [6][7] [8], and scenario analysis [9][10] [11]. The auto-generated test cases from ADS scenarios can also be replicated in other domains where specific definition languages and ranged values are used. )…”