Deep Neural Networks (DNN) will emerge as a cornerstone in automotive software engineering. However, developing systems with DNNs introduces novel challenges for safety assessments. This paper reviews the state-of-the-art in verification and validation of safety-critical systems that rely on machine learning. Furthermore, we report from a workshop series on DNNs for perception with automotive experts in Sweden, confirming that ISO 26262 largely contravenes the nature of DNNs. We recommend aerospace-to-automotive knowledge transfer and systems-based safety approaches, e.g., safety cage architectures and simulated system test cases.

No abstract

No abstract

No abstract

EAST-ADL is an Architecture Description Language (ADL) initially defined in several European-funded research projects and aligned with AUTOSAR and ISO26262. It provides a comprehensive approach for defining automotive electronic systems through an information model that captures engineering information in a standardized form. Aspects covered include vehicle features, requirements, analysis functions, software and hardware components and communication. The representation of the system's implementation is not defined in EAST-ADL itself but by AUTOSAR. However, traceability is supported from EAST-ADL's lower abstraction levels to the implementation level elements in AUTOSAR. In this article we describe EAST-ADL in detail, show how it relates to AUTOSAR as well as other significant automotive standards and present recent research work on using and advancing EAST-ADL, the functional safety standard ISO 26262, heterogeneous multi / many core architectures, security and for multi-objective optimization.