Efficient Integration of cross platform functions onto service-oriented architectures

TL;DR

This paper proposes a platform-agnostic application development and integration approach for automotive software, enabling efficient deployment across heterogeneous architectures like AUTOSAR and ROS 2.

Contribution

It introduces a standardized, hardware- and middleware-agnostic application design with semi-automated tools for cross-platform integration in automotive systems.

Findings

Successful integration of an example application onto AUTOSAR Adaptive and ROS 2

Demonstrated efficiency improvements through standardized interfaces and semi-automated tools

Validated the approach with relevant metrics showing enhanced development and integration efficiency

Abstract

The automotive industry is currently undergoing a major transformation with respect to the Electric/Electronic (E/E) and software architecture, driven by a significant increase in the complexity of the technological stack within a vehicle. This complexity acts as a driving force for Software-Defined Vehicles (SDVs) leading to the evolution of the automotive E/E architectures from decentralized configuration comprising multiple Electronic Control Units (ECUs) towards a more integrated configuration comprising a smaller number of ECUs, domain controllers, gateways, and High-Performance Computers (HPCs) [2]. This transition along with several other reasons have resulted in heterogeneous software platforms such as AUTOSAR Classic, AUTOSAR Adaptive, and prototypical frameworks like ROS 2. It is therefore essential to develop applications that are both hardware- and platform/middleware-agnostic to attain development and integration efficiency. This work presents an application development and integration concept to facilitate developing applications as Software as a Product (SaaP), while simultaneously ensuring efficient integration onto multiple software architecture platforms. The concept involves designing applications in a hardware- and software platform-agnostic manner and standardizing application interfaces [6]. It also includes describing the relevant aspects of the application and corresponding middleware in a machine-readable format to aid the integration of developed applications. Additionally, tools are developed to facilitate semi-automation of the development and integration processes. An example application has been developed and integrated onto AUTOSAR Adaptive and ROS 2, demonstrating the applicability of the approach. Finally, metrics are presented to show the efficiency of the overall concept.