Automated vehicular dashboard validation and verification play a key role in the rapid integration of high-precision dashboards. Vision-based methods for accurately estimating tachometer and speedometer readings, validated against internal vehicle signals, are met with two major challenges: limited training data and the selection of the most effective gauge estimation method. To address these challenges, we leverage automated segmentation and tracking via the Segment Anything Model and DeAOT (SAM-track) to streamline the gauge labeling process required for training machine learning models, facilitating the rapid creation of a training dataset. This dataset is then used to train a Detectron 2 model for segmenting speedometer and tachometer gauges. For angle estimation, we train three different machine learning models on augmented data to predict gauge angles based on segmented masks. Additionally, our pipeline includes conventional rule-based angle estimation methods. Preliminary benchmarking results highlight ResNet as the most accurate method, achieving a Mean Absolute Error (MAE) of $0.86 \text{km} / \mathrm{h}$ for speed estimation and 49.74 rounds per minute (RPM) for tachometer readings. The study further examines how different resolution levels and visual distortions impact performance, two common issue in real-world applications. When gauge masks are resized between $1 / 4$ and 4 times their original resolution, traditional rule-based methods, such as ellipse fitting, demonstrate greater robustness in maintaining cross-resolution accuracy compared to machine learning models, highlighting their reliability for real-world applications.

This paper addresses the critical challenge of capturing and preserving knowledge about Internet of Things (IoT) infrastructure in increasingly complex technological environments. We highlight key limitations in the use of plain technical documentation and even detailed Standard Operating Procedures (SOPs) to describe how IoT devices are used in operational settings. These limitations include inconsistent instructions, lack of machine-readability, and undefined or ambiguous terminology - all of which hinder effective knowledge sharing and reuse. To overcome these issues, we propose the use of Knowledge Graphs (KGs) as a structured and systematic method to describe the contextual environments in which IoT instruments operate, collect data, and act upon information. While our broader objective is to explore how KGs can support the understanding, reuse, and management of IoT-generated data, this paper specifically focuses on the capture of contextual IoT knowledge necessary for building such graphs. We introduce the INS(trument) metadata template, which facilitates the description of IoT components, the relationships among them, and the influence of these components and interactions on data interpretation and reuse. To demonstrate the practical application of our approach, we present a case study from the Arrowhead fPVN Project in the context of automotive battery innovation. In this case, the INS metadata template is used to describe IoT components involved in two end-of-life battery tests-one conducted by AVL List GmbH, and the other by Sandia National Laboratories. The metadata was compiled from publicly available documentation provided by both organizations.

The automated generation of diversified training scenarios has been an important ingredient in many complex learning tasks, especially in real-world application domains such as autonomous driving, where auto-curriculum generation is considered vital for obtaining robust and general policies. However, crafting traffic scenarios with multiple, heterogeneous agents is typically considered a tedious and time-consuming task, especially in more complex simulation environments. To this end, we introduce MATS-Gym, a multi-agent training framework for autonomous driving that uses partial-scenario specifications to generate traffic scenarios with a variable number of agents which are executed in CARLA, a high-fidelity driving simulator. MATS-Gym reconciles scenario execution engines, such as Scenic and ScenarioRunner, with established multi-agent training frameworks where the interaction between the environment and the agents is modeled as a partially observable stochastic game. Furthermore, we integrate MATS-Gym with techniques from unsupervised environment design to automate the generation of adaptive auto-curricula, which is the first application of such algorithms to the domain of autonomous driving. The code is available at https://github. com/AutonomousDrivingExaminer/mats-gym.

Video moment retrieval aims to locate the timestamps best matching the query description within an untrimmed video. However, existing video moment retrieval approaches typically suffer from two major limitations: (1)Utilize only negative moment-sentence pairs sampled from intra-videos, which may overfit the bias of the dataset and not have an excellent understanding of the video and query due to the dataset size and annotation biases. (2)Decouple the video and the query, perform unimodal learning separately, and then concatenate them together as multimodal fusion features. In this paper, we propose a novel approach named Momentum Contrastive Matching Network(MCMN). Inspired by MoCo, we propose the Momentum Cross-modal Contrast for cross-modal learning to enable large-scale negative sample interactions, which contributes to the generation of more precise and discriminative representations, and use temporal decay to model key attenuation in the memory queue when computing the contrastive loss. In addition, we use an attention module to adaptively generate clip-specific word embeddings to achieve semantic alignment from a temporal perspective, which are considered to be more important for finding relevant video contents with large boundary ambiguities. Experimental results on the three major video moment retrieval benchmark datasets, including TACoS, Charades-STA, and ActivityNet Captions demonstrate that MCMN surpasses previous methods and reaches state-of-the-art with disparate visual features.

Understanding attention is crucial for improving safety in driving scenarios. Detected and classified objects, along with their observation by the driver, are used as a measure of attention. This paper investigates the differences between human and artificial attention in real-world and replay driving scenarios. By analyzing attention patterns from drivers and a vision-language model agent, we identify a number of differences. The results highlight the limitations of current AI attention models and suggest the way forward for developing more context-aware systems.

Connected cooperative and automated mobility (CCAM) benefits from reliable wireless vehicle-to-everything (V2X) communication links in safety-critical and time-sensitive situations. The ego vehicle's perception, primarily derived from LIDAR, RADAR, and camera data, is limited by the line-of-sight (LOS). Sensor information beyond the LOS can be acquired by reliable V2X communication links from other cooperative vehicles or infrastructure elements. We identify CCAM use cases for both real-world applications and test phases, which stand to gain from understanding spatial reliability regions for communication links. Frame error rate (FER) classes for these regions, from the perspective of the ego vehicle, are provided to aid decision-making for autonomous vehicles. We propose a testbed architecture for system validation, verification, and test scenario generation, which integrates FER prediction through a high-performance open-source computing reference framework (HOPE). Our study demonstrates that the measured FER within a city scenario closely aligns with the FER obtained via a hardware-in-the-loop (HiL) framework and a non-stationary geometry-based stochastic channel model (GSCM) that utilizes OpenStreetMap data enriched with event-specific static objects. We use the GSCM and the HiL framework to overcome the fundamental limits of estimating the FER in non-stationary scenarios. As a final demonstration of the HOPE framework, we achieve an 80 % accuracy in predicting the FER class.

AbstractVehicles are on the verge building highly networked and interconnected systems with each other. This requires open architectures with standardized interfaces. These interfaces provide huge surfaces for potential threats from cyber attacks. Regulators therefore demand to mitigate these risks using structured security engineering processes. Testing the effectiveness of this measures, on the other hand, is less standardized. To fill this gap, this book chapter contains an approach for structured and comprehensive cybersecurity testing of contemporary vehicular systems. It gives an overview of how to define secure systems and contains specific approaches for (semi-)automated cybersecurity testing of vehicular systems, including model-based testing and the description of an automated platform for executing tests.

In this paper we present a hardware-in-the-loop (HiL) framework for testing wireless vehicle-to-everything (V2X) communication hardware, i.e., modems under realistic channel conditions. The framework includes a wireless channel emulator, which is capable of emulating non-stationary wireless channels in real-time. We validate the HiL framework by comparing the frame error rate (FER) obtained via emulation with data obtained during a V2X measurement campaign using the same IEEE 802.11p based modems. To do this we acquire measured time-variant channel transfer function and FER measurements simultaneously. The results show that our HiL approach is feasible and that we can obtain FER measurements in the laboratory that closely match the measurement results obtained on the road, giving the maximal distance of 0.099 between their cumulative distribution functions.

The NimbleAI Horizon Europe project leverages key principles of energy-efficient visual sensing and processing in biological eyes and brains, and harnesses the latest advances in $\mathbf{33D}$ stacked silicon integration, to create an integral sensing-processing neuromorphic architecture that efficiently and accurately runs computer vision algorithms in area-constrained endpoint chips. The rationale behind the NimbleAI architecture is: sense data only with high information value and discard data as soon as they are found not to be useful for the application (in a given context). The NimbleAI sensing-processing architecture is to be specialized after-deployment by tunning system-level trade-offs for each particular computer vision algorithm and deployment environment. The objectives of NimbleAI are: (1) $\mathbf{100x}$ performance per mW gains compared to state-of-the-practice solutions (i.e., CPU/GPUs processing frame-based video); (2) $\mathbf{50x}$ processing latency reduction compared to CPU/GPUs; (3) energy consumption in the order of tens of mWs; and (4) silicon area of approx. 50 mm 2 .

Vehicular communications face unique security issues in wireless communications. While new vehicles are equipped with a large set of communication technologies, product life cycles are long and software updates are not widespread. The result is a host of outdated and unpatched technologies being used on the street. This has especially severe security impacts because autonomous vehicles are pushing into the market, which will rely, at least partly, on the integrity of the provided information. We provide an overview of the currently deployed communication systems and their security weaknesses and features to collect and compare widely used security mechanisms. In this survey, we focus on technologies that work in an ad hoc manner. This includes Long-Term Evolution mode 4 (LTE-PC5), Wireless Access in Vehicular Environments (WAVE), Intelligent Transportation Systems at 5 Gigahertz (ITS-G5), and Bluetooth. First, we detail the underlying protocols and their architectural components. Then, we list security designs and concepts, as well as the currently known security flaws and exploits. Our overview shows the individual strengths and weaknesses of each protocol. This provides a path to interfacing separate protocols while being mindful of their respective limitations.

Wireless sensor networks are increasingly used for improving the operation of industrial machines and facilities. Such complex and distributed networks are composed of embedded devices. These devices are typically powered by batteries or energy harvesting, and can thus be classified as energy constrained devices. Especially, the embedded devices powered by energy harvesting rely on sufficient harvestable energy. The correct operation of the whole network may depend on the correct operation of a single device. Thus, relevant energy harvesting scenarios must be ideally covered by functional tests of such networks. To be able to emulate different energy harvesting scenarios in combination with the distributed character of such networks, we have developed a distributed testbed based on the robot operating system (ROS). This paper presents the architecture and describes relevant testing aspects for such devices. Furthermore, it introduces a use case for testing wireless sensor nodes supplied by energy harvesting and presents exemplary testing procedure results. Finally, it discusses the advantages of incorporating such testing procedures into a continuous integration process to verify the correct functionality of the embedded devices during development.

This paper presents a novel approach to improving wireless communications in harsh propagation environments to achieve higher overall reliability and durability of wireless battery powered sensor systems in the context of in-vehicle communication. The goal is to investigate the physical layer and establish an antenna recommendation system for a specific harsh environment, i.e., an engine compartment of a vehicle. We propose the usage of electromagnetic (EM) and ray tracing simulations as a computationally cost-effective method to establish such a recommendation system, which we test by means of an experimental testbed-or test environment-that consists of both a physical, as well as its identical simulation, model. A pool of antennas is evaluated to identify and verify antenna behavior and properties at specified positions in the harsh environment. We use a vector network analyzer (VNA) for accurate measurements and a received signal strength indicator (RSSI) for a first estimation of system performance. Our analysis of the experimental measurements and its EM simulation counterparts shows that both types of data lead to equivalent antenna recommendations at each of the defined positions and experimental conditions. This evaluation and verification process by measurements on an experimental testbed is important to validate the antenna recommendation process. Our results indicate that-with properly characterized antennas-such measurements can be substituted with EM simulations on an accurate EM model, which can contribute to dramatically speeding up the antenna positioning and selection process.

This paper discusses a freely available and open dataset containing vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I) and vehicle-to-pedestrian (V2P) OFDM-based wireless channel measurement data including synchronised sensor data such as radar, LiDAR and high precision GPS. The wireless channel measurement is conducted at the carrier frequencies of 3.2 GHz and 5.81 GHz which are the most promising frequency bands in which future V2X communication systems will operate. The dataset contains the wireless channel measurement data of various V2X scenarios along with synchronized sensor information from a vehicle. In addition to the wireless channel measurement data, the dataset also includes frame error rate measurements from a IEEE 802.11p based communication system, synchronized to the other measurement data.

This paper describes the measurement setup for the proliferation of the vehicular communication systems in operating cars. The setup builds on a communication sniffer which detects cooperative awareness messages broadcast from ETSI ITS-G5 systems, decodes them and records the public information in these messages. The experimental characterization of the setup in an outdoor scenario is presented in order to verify the sniffer's performance with respect to transmit power and communication range. The setup is deployed for measurements on a busy road with slow moving traffic in the center of the city of Linz, Austria. The number of active users as well as their transmit power, duration of the transmission and packet length have been recorded.

Internet-of-Things (IoT) devices and other embedded devices are more and more used to measure different conditions inside of buildings and industrial facilities as well as to monitoring machines or industrial processes.IoT sensors communicate wirelessly and are typically supplied by batteries.Energy harvesting can be used to extend their operational time or enable self-sufficient supply of them.Energy from the environment is converted into electrical energy by energy harvesting devices (EHDs), for example solar cells or thermo-generators.However, the available output power of the EHDs is highly dependent on the mounting location as well as on environmental conditions and may vary greatly over time.Therefore, it is meaningful to evaluate the EHDs at the location of use over a certain period of time in order to characterize them in real world scenarios.This paper presents the evaluation setup and the results of the characterization of four different solar cells at different locations in an office building and at different weather conditions.Furthermore, a method is presented to estimate the possibility to supply embedded device using energy harvesting.The results can be used to simplify the selection of a suitable EHD and the design process of an energy management system.The method is applied on two different use cases to estimate the needed size of the solar cells to enable a continuous supply.

The collaborative activities of a diverse range of partners have resulted in a variety of assets directed towards trustworthy IoT and its integration into autonomous driving and Industry 4.0 applications. This paper strays away from technical development. Its motive is to establish a process and define an adequate set of high-level generic measures that could be implemented to support digital transformation beyond the project’s closing phase. The focus is placed on successful exploitation with a sustainable outlook for the project results in a quest to maximise benefits for a range of stakeholders. To that extent, the paper considers the realistic maximisation of benefits through the implementation of a strategy to improve the value proposition. These activities are prolonging and maximising the impact of the project.

In this paper we adopt and modify a well-known locality-aware hashing scheme to the problem of active stochastic scatterer selection in vehicular non-stationary geometry-based stochastic channel models (GSCM). We show, how under relaxed assumptions on the query set an efficient selection of active stochastic scatterers during simulation is computationally feasible. The proposed approach enables real-time simulation and emulation of large-scale GSCMs by restricting the active stochastic scatterer set to meet given resource constraints. We showcase our approach by introducing a GSCM that is bootstrapped via OpenStreetMap data. The stochastic scatterers are placed automatically along buildings, traffic signs and vegetation. We validate and investigate the impact of the proposed approach on the accuracy of a GSCM by means of second order statistics of the time- and frequency-varying fading process. For validation and performance evaluation we parameterize our GSCM using a vehicular wireless channel measurement campaign conducted in the inner city of Vienna. The impact of selecting only a subset of scatterers is then evaluated using the calibrated GSCM.

Historically, automotive test systems were designed for single core architectures on both the Electronic Control Unit (ECU) and the Verification and Validation (V&V). This, however, limited the utilization of shared resources, such as memory management or network interfaces. In this paper we present a redesign of an automotive test system that is based on a multi-core architecture and capable of managing mixed-criticality data that is exchanged between the ECU and V&V system. As part of the redesign, we implemented a Connectivity Manager (CM) that is in charge of multiplexing several data streams from multiple cores across a shared network. Due to the increased complexity of our system, a more flexible communication scheduling approach is required. Our solution to this problem is a novel dynamic prioritization approach that adapts to bandwidth changes on the shared communication network. Through simulations with realistic workloads on the CAN bus, we demonstrate the proper functioning of our algorithm with the result that higher critical data streams are favoured over less critical data streams in case of bus overloads or temporary bottlenecks.

In this paper, we present a novel equivalent circuit (EC) model that can be used to conduct a circuit-level interoperability analysis of a wireless power transfer (WPT) system and a near-field communication (NFC) system in close proximity to one another. Each system consists of two devices, which are respectively the WPT transmitter, WPT receiver, NFC reader, and NFC transponder (tag). Due to their close proximity, the WPT and the NFC systems interact with one another due to parasitic inductive and capacitive coupling between their coil antennas. Our investigations show that parasitic coupling causes the NFC system to produce a distorted communication signal. In particular, the detectability of the communication signal sent from the NFC tag to the NFC reader decreases rapidly, if the parasitic capacitive coupling between the coil antennas is too large. In addition, a custom-built prototype is presented here that can be used to take measurements and verify these results.

Through international regulations (most prominently the latest UNECE regulation) and standards, the already widely perceived higher need for cybersecurity in automotive systems has been recognized and will mandate higher efforts for cybersecurity engineering. The UNECE also demands the effectiveness of these engineering to be verified and validated through testing. This requires both a significantly higher rate and more comprehensiveness of cybersecurity testing that is not effectively to cope with using current, predominantly manual, automotive cybersecurity testing techniques. To allow for comprehensive and efficient testing at all stages of the automotive life cycle, including supply chain parts not at hand, and to facilitate efficient third party testing, as well as to test under real-world conditions, also methodologies for testing the cybersecurity of vehicular systems as a black box are necessary. This paper therefore presents a model and attack tree-based approach to (semi-)automate automotive cybersecurity testing, as well as considerations for automatically black box-deriving models for the use in attack modeling.