An OpenSource CI/CD Pipeline for Variant-Rich Software-Defined Vehicles
Matthias Weiß, Anish Navalgund, Johannes Stümpfle, Falk Dettinger and Michael Weyrich · 2025
Read An OpenSource CI/CD Pipeline for Variant-Rich Software-Defined Vehicles by Matthias Weiß, Anish Navalgund, Johannes Stümpfle, Falk Dettinger and Michael Weyrich (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-19446.txt.
Introduction
Matthias Weiß and co-authors build an open pipeline for software defined vehicle updates in a 2025 paper. Cars now ship features by wireless update, and variants multiply fast. Isolated build setups make that sprawl hard to manage. Their pipeline automates build, test, and release with container tools.
A custom updater pushes variants matched to each car hardware set. Rollback support guards failed releases. AI model updates for driving features ride the same rails. A valet parking demo with small robots proved variant picks and updates end to end. The open stack gives makers a shared base for fleet scale software care.
Catalog record
| CREATOR | Matthias Weiß, Anish Navalgund, Johannes Stümpfle, Falk Dettinger and Michael Weyrich |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 26 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-19446
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
An OpenSource CI/CD Pipeline for Variant-Rich Software-Defined Vehicles Thanks: Identify applicable funding agency here. If none, delete this. Matthias Weiß, Anish Navalgund, Johannes Stümpfle, Falk Dettinger and Michael Weyrich Abstract Software-defined vehicles (SDVs) offer a wide range of connected functionalities, including enhanced driving behavior and fleet management. These features are continuously updated via over-the-air (OTA) mechanisms, resulting in a growing number of software versions and variants due to the diversity of vehicles, cloud/edge environments, and stakeholders involved. The lack of a unified integration environment further complicates development, as connected mobility solutions are often built in isolation. To ensure reliable operations across heterogeneous systems, a dynamic orchestration of functions that considers hardware and software variability is essential. This paper presents an open-source CI/CD pipeline tailored for SDVs. It automates the build, test, and deployment phases using a combination of containerized open-source tools, creating a standardized, portable, and scalable ecosystem accessible to all stakeholders. Additionally, a custom OTA middleware distributes software updates and supports rollbacks across vehicles and backend services. Update variants are derived based on deployment target dependencies and hardware configurations. The pipeline also supports continuous development and deployment of AI models for autonomous driving features. Its effectiveness is evaluated using an automated valet parking (AVP) scenario involving TurtleBots and a coordinating backend server. Two object detection variants are developed and deployed to match hardware-specific requirements. Results demonstrate seamless OTA updates, correct variant selection, and successful orchestration across all targets. Overall, the proposed pipeline provides a scalable and efficient solution for managing software variants and OTA updates in SDVs, contributing to the advancement of future mobility technologies. I Introduction Software-defined vehicles (SDVs) represent a transformative shift in the automotive industry, enabling dynamic, software-driven functionalities that enhance vehicle behavior, safety, and user experience [ 1 ] [ 2 ] . Unlike traditional vehicles with fixed capabilities, SDVs leverage continuous connectivity to cloud and edge infrastructures to enable advanced features such as real-time traffic-based routing, predictive maintenance, and remote fleet management [ 3 ] [ 4 ] . This software-centric evolution necessitates frequent and often automated software updates via over-the-air (OTA) mechanisms, making vehicles not only transport devices but also continuously evolving cyber-physical systems [ 5 ] . A central challenge emerging from this evolution is the growing complexity in managing software variants. Even conventional vehicles exhibit significant variability due to regional regulations, hardware configurations, and feature bundles. With the increasing reliance on software, this variability expands dramatically: different vehicle generations, sensor configurations, compute platforms, and regional software stacks must be considered [ 6 ] . Consequently, each update potentially requires multiple tailored variants to ensure compatibility and functionality across a heterogeneous fleet. To address this complexity, continuous integration and continuous deployment (CI/CD) pipelines have become a critical tool in modern software engineering. These pipelines automate the build, testing, and deployment of software, supporting rapid and reliable iterations. In the context of SDVs, CI/CD practices promise to streamline the development and delivery of both onboard and backend services, ensuring consistent quality and faster rollout of innovation [ 7 ] . However, despite the availability of mature CI/CD tools, two key challenges remain unresolved in the automotive domain. First, SDVs require stable orchestration of updates across distributed targets —including in-vehicle platforms, edge nodes, and backend cloud services—each with unique requirements and availability constraints [ 8 ] . Second, the management of software variants —where a single codebase must be adapted to different hardware configurations and software environments—requires intelligent deployment logic and variant-aware tooling [ 9 ] . In light of these challenges, this paper addresses the following research question: How can an open-source CI/CD pipeline be designed to support stable, variant-aware deployment of connected vehicle software across heterogeneous targets including cloud, edge, and in-vehicle environments? The remainder of this paper is structured as follows: Section II reviews related work and existing CI/CD solutions in automotive and distributed systems. Section III presents the architecture and components of the proposed open-source CI/CD pipeline. Section IV evaluates the pipeline using an automated valet parking (AVP) scenario. Finally, Section V concludes the paper and outlines directions for future research. II Background II-A Variant-rich Software-Defined Vehicles The high variability in software-defined vehicles enables unprecedented configurability and individuality, allowing each vehicle to be tailored to specific market requirements and customer preferences. However, this same flexibility introduces substantial complexity in managing the resulting software ecosystem [ 10 ] . With SDVs designed to remain current through continuous over-the-air updates throughout their extended lifecycle—often spanning 10-15 years—the combination of high variance and frequent updates creates a significant risk of software erosion [ 11 ] . This risk is particularly concerning given that SDVs fundamentally rely on software for their core functionality. To address this challenge, the automotive industry is increasingly adopting software product line engineering (SPLE) approaches [ 12 ] . SPLE provides systematic methods for managing commonalities and variabilities across the vehicle software portfolio, enabling efficient development and maintenance of multiple product variants from a shared set of core assets. The challenge however lies in adopting such an SPL [ 13 , 14 ] . As vehicles receive regular updates to add features, improve performance, and address security vulnerabilities, the number of possible software configurations grows exponentially. Each update must be compatible not only with the base vehicle platform but also with the unique combination of features and variants present in individual vehicles. Modern middleware solutions like Adaptive AUTOSAR incorporate SPLE principles, providing service-oriented architecture for modular updates and Update and Configuration Management (UCM) services [ 15 ] . The challenge extends beyond technical deployment to encompass version control, dependency management, and regression testing across an ever-expanding matrix of software variants. II-B DevOps and CI/CD DevOps is a software engineering paradigm that emphasizes collaboration between development and operations teams, supported by extensive automation. Through continuous integration (CI) and continuous deployment (CD), DevOps enables rapid, reliable software updates. In the automotive domain, this approach is increasingly relevant as software-defined vehicles demand frequent OTA updates to both onboard functions and connected backend services [ 16 ] . CI/CD pipelines implement DevOps principles by automating build, test, and deployment processes. With the growing role of machine learning (ML) in SDVs—e.g., for perception or driver behavior—MLOps has emerged as an extension of CI/CD to support continuous training, validation, and deployment of ML models [ 17 ] . Together, CI/CD and MLOps provide a foundation for agile, data-driven vehicle software development. Recent industry initiatives such as COVESA [ 18 ] and SOAFEE [ 19 ] promote cloud-native architectures and runtime environments for vehicles, bringing containerization and orchestration tools to the edge. Academic efforts have demonstrated CI/CD integration with physical vehicles, enabling rapid iteration and feedback loops. However, two major gaps persist. First, variant management remains a challenge: software must be tailored to diverse vehicle configurations, yet current tools lack native support for variant-aware deployment [ 9 ] . Second, constructing closed ML data loops—where field data informs continuous model improvement and redeployment—is difficult due to bandwidth, validation, security and integration constraints [ 8 ] [ 20 ] . Overall, while the foundations for DevOps in SDVs exist, an end-to-end CI/CD solution that unifies backend orchestration, in-vehicle deployment, variant management, and ML lifecycle integration is still lacking.