Falk Dettinger

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.

Abstract

Cloud-based offloading helps address energy consumption and performance challenges in executing resourceintensive vehicle algorithms. Utilizing 5G, with its low latency and high bandwidth, enables seamless vehicle-to-cloud integration. Currently, only non-standalone 5G is publicly available, and realworld applications remain underexplored compared to theoretical studies. This paper evaluates 5G non-standalone networks for cloud execution of vehicle functions, focusing on latency, Round Trip Time, and packet delivery. Tests used two AI-based algorithms—emotion recognition and object recognition—along an 8.8 km route in Baden-Württemberg, Germany, encompassing urban, rural, and forested areas. Two platforms were analyzed: a cloudlet in Frankfurt and a cloud in Mannheim, employing various deployment strategies like conventional applications and containerized and container-orchestrated setups. Key findings highlight an average signal quality of 84 %, with no connectivity interruptions despite minor drops in built-up areas. Packet analysis revealed a Packet Error Rate below 0.1 % for both algorithms. Transfer times varied significantly depending on the geographical location and the backend servers' network connections, while processing times were mainly influenced by the computation hardware in use. Additionally, cloud offloading seems only be a suitable option, when a round trip time of more than 150 ms is possible.

Abstract

Connected and software-defined vehicles promise to offer a broad range of services and advanced functions to customers, aiming to increase passenger comfort and support autonomous driving capabilities. Due to the high reliability and availability requirements of connected vehicles, it is crucial to resolve any occurring failures quickly. To achieve this however, a complex cloud/edge architecture with a mesh of dependencies must be navigated to diagnose the responsible root cause. As such, manual analyses become unfeasible since they would significantly delay the troubleshooting. To address this challenge, this paper presents SDVDiag, an extensible platform for the automated diagnosis of connected vehicle functions. The platform enables the creation of pipelines that cover all steps from initial data collection to the tracing of potential root causes. In addition, SDVDiag supports selfadaptive behavior by the ability to exchange modules at runtime. Dependencies between functions are detected and continuously updated, resulting in a dynamic graph view of the system. In addition, vital system metrics are monitored for anomalies. Whenever an incident is investigated, a snapshot of the graph is taken and augmented by relevant anomalies. Finally, the analysis is performed by traversing the graph and creating a ranking of the most likely causes. To evaluate the platform, it is deployed inside an 5G test fleet environment for connected vehicle functions. The results show that injected faults can be detected reliably. As such, the platform offers the potential to gain new insights and reduce downtime by identifying problems and their causes at an early stage.

Abstract

The limitations of on-board sensors and blind spots caused by occlusion cause the reduction of perception quality in autonomous vehicles. In such cases, cooperative perception provides additional data via Vehicle-to-Everything communication to enhance local perception, causing a large volume of data transmission. The vehicle can focus on acquiring and utilizing relevant data according to the prevailing road context by identifying the current traffic situation. To achieve this, we propose a concept for the situation identification of the vehicle using Bird's-Eye-View images. Firstly, the situation around the vehicle is identified using object detection with semantic segmentation, followed by understanding the context of the traffic using a situation identification module consisting of an open-source projective transformation network Cam2BEV and a situation identification neural network. The concept was evaluated and validated by running the software on the CARLA simulator using the in-built RGB camera and the semantic segmentation camera. Additionally, the portability of the situation identification module for real-world applications was verified on Cityscapes and nuScenes urban driving datasets. Overall, the proposed situation identification approach enables efficient sensor data management by prioritizing relevant data to the current traffic situation. The source code is available in the following link: https://github.com/akshaynarla/DySi_Select

Abstract

The automotive industry is undergoing an exten- sive transition from hardware-centric development to software- defined vehicles (SDVs). This paradigm shift requires innovative technical solutions and fundamentally different approaches for vehicle design. Together, these developments introduce substantial challenges for an entire industry, affecting both development processes and operations. Although SDVs have attracted con- siderable attention and widespread industry involvement, prior research has primarily emphasized systematic literature reviews identifying emerging academic trends. Consequently, little at- tention has been paid to industry and economic developments that influence how SDV technologies evolve in practice and are crucial for future advancement. In response, this paper provides a different, industry-oriented view of the current trends and associated challenges of SDVs. To achieve this, we conducted a comprehensive analysis of the SDV from an industry-based perspective. We analyzed three sources of information to identify the current trends in SDV development: expert interviews with automotive IT specialists, a study on major automotive trade fairs and workshops, and insights gained from the authors’ professional experience. Based upon, we derive specific challenges arising from these trends. Our findings offer a comprehensive

Abstract

Connecting the vehicle with cloud or edge systems enables new features for the customer. Additionally, existing vehicle functions or software components (SWCs) can also be offloaded to this network ecosystem. For example, it is envisaged that due to cost-intensive vehicle hardware, individual functions will not continue to run in the vehicle but will be executed in the cloud or edge. This means that the functions and their SWCs no longer remain statically deployed and executed on onboard hardware like an electronic control unit (ECU) but can be rede- ployed during the vehicle’s life cycle. To make this possible, application models that take into account the individual requirements of automotive functions are needed. This paper proposes five different application mod- els in addition to the standard in-vehicle application. The various fea- tures of the application models are predestined for the multiple-criteria decision-making (MCDM) approach. We propose OptiCAM, a MCDM model that helps software developers or electrical/electronic (E/E) archi- tects finding the best application model for SWCs based on technical and economical criteria. State-of-the-art vehicle functions are finally used to test OptiCAM and analyze the results.

Abstract

In semi-automated manufacturing, an increasing amount of intelligent mobile robots operate in close proximity to human workers. Considering future positions of humans allows to further improve the efficiency in terms of throughput of Autonomous Mobile Robots (AMR) and Automated Guided Vehicles (AGV). The longer the prediction horizon, i.e. the more position values of humans can be predicted in the near and distant future, the more a robot can adjust its route accordingly and optimize the process. This paper discusses the challenges of human motion trajectory prediction in manufacturing and presents a schedule-based approach that uses real-time schedule data obtained from Manufacturing Execution Systems (MES). Schedule-awareness in human motion trajectory prediction extends semantic mapping approaches and effectively reduces the number of probable destinations by considering which process steps are next for the currently produced goods. With a reduced set of destinations, the performance of forward-planning trajectory prediction can be improved. For evaluation, a commercial MES is used together with an Ultra-wideband-based Real-Time Locating System (RTLS) for obtaining position data of humans. On this basis, a naive Bayes classifier utilizes MES-schedule and real-time position data to predict human motion intentions. Abstract activity modeling ensures that only a few training data sets are required for deployment, thus making this approach suitable for rapidly changing manufacturing environments such as in flexible manufacturing.