Scholarlink

Validation and approval of automated driving systems require the application of extensive virtual testing methods. Within the International Alliance for Mobility Testing and Standardization (IAMTS), a four-step reference process including methods for the correlation of the virtual and real world was proposed. This process needs to be executed based on the exact needs of the coverage of the virtual world and the corresponding formalized requirements. The process is a bottom-up approach starting with a subsystem (brake system, a radar model, or artifacts) as standalone towards the integrated system including all sub-components. To save time and costs, it is important to only model the relevant systems and the relevant behavior for the dedicated operational design domain (ODD), use cases and scenarios. A classification of these models according to their individual features supports the selection significantly. While fidelity classifications for the vehicle dynamics are standardized within the ISO/TC 22/SC 33, a classification scheme of the perception system is still missing. To address this gap, a proposal feature-based classification scheme for automotive perception sensors together with a classification scheme of environment simulation features was provided. The sensor model classification presented is comparable to existing classifications, such as general model fidelities for mechanical systems or to the sensor model classification as presented by the SetLevel project (Förderkennzeichen BMWK 19A19004S). The proposed approach is by no means exhaustive or complete, but it is easily extendable to further features of sensor model. Nevertheless, this classification method provides a simple way to check which features are needed in the simulation based on the ODD, the driving function requirements and the corresponding sensors. Even taking into account ISO21448 and ISO26262, this classification helps to determine which features need to be validated particularly intensively in order to be able to use the simulation in compliance with EU regulation 2022/1426.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Novel Classification Scheme for Perception Sensors Incorporated into Simulation Validation Process for AD

  • 章节

  • 发表日期：

    2026年

查看全文
  • Christopher Wiegand,
  • Jakob Reckenzaun,
  • Tobias Düser,
  • Jonas Freyer

摘要

Validation and approval of automated driving systems require the application of extensive virtual testing methods. Within the International Alliance for Mobility Testing and Standardization (IAMTS), a four-step reference process including methods for the correlation of the virtual and real world was proposed. This process needs to be executed based on the exact needs of the coverage of the virtual world and the corresponding formalized requirements. The process is a bottom-up approach starting with a subsystem (brake system, a radar model, or artifacts) as standalone towards the integrated system including all sub-components. To save time and costs, it is important to only model the relevant systems and the relevant behavior for the dedicated operational design domain (ODD), use cases and scenarios. A classification of these models according to their individual features supports the selection significantly. While fidelity classifications for the vehicle dynamics are standardized within the ISO/TC 22/SC 33, a classification scheme of the perception system is still missing. To address this gap, a proposal feature-based classification scheme for automotive perception sensors together with a classification scheme of environment simulation features was provided. The sensor model classification presented is comparable to existing classifications, such as general model fidelities for mechanical systems or to the sensor model classification as presented by the SetLevel project (Förderkennzeichen BMWK 19A19004S). The proposed approach is by no means exhaustive or complete, but it is easily extendable to further features of sensor model. Nevertheless, this classification method provides a simple way to check which features are needed in the simulation based on the ODD, the driving function requirements and the corresponding sensors. Even taking into account ISO21448 and ISO26262, this classification helps to determine which features need to be validated particularly intensively in order to be able to use the simulation in compliance with EU regulation 2022/1426.