Keywords: Image Sensor, Vehicular Sensor
Abstract: Estimating camera intrinsic parameters without prior scene knowledge is a fundamental challenge in computer vision. This capability is particularly important for applications such as autonomous driving and vehicle platooning, where precalibrated setups are impractical and real-time adaptability is necessary. To advance the state-of-the-art, we present a set of equations based on the calibrated trifocal tensor, enabling projective camera self-calibration from minimal image data. Our method, termed TrifocalCalib, significantly improves accuracy and robustness compared to both recent learning-based and classical approaches. Unlike many existing techniques, our approach requires no calibration target, imposes no constraints on camera motion, and simultaneously estimates both focal length and principal point. Evaluations in both procedurally generated synthetic environments and structured dataset-based scenarios demonstrate the effectiveness of our approach. To support reproducibility, we make the code publicly available.

Keywords: Road Accident Investigation and Risk Assessment, Active and Passive Safety Systems, Driver Assistance Systems
Abstract: Ensuring safety is the primary objective of automated driving, which necessitates a comprehensive and accurate perception of the environment. While numerous performance evaluation metrics exist for assessing perception capabilities, incorporating safety-specific metrics is essential to reliably evaluate object detection systems. A key component for safety evaluation is the ability to distinguish between relevant and non-relevant objects — a challenge addressed by criticality or relevance metrics. This paper presents the first in-depth analysis of criticality metrics for safety evaluation of object detection systems. Through a comprehensive review of existing literature, we identify and assess a range of applicable metrics. Their effectiveness is empirically validated using the DeepAccident dataset, which features a variety of safety-critical scenarios. To enhance evaluation accuracy, we propose two novel application strategies: bidirectional criticality rating and multi-metric aggregation. Our approach demonstrates up to a 100% improvement in terms of criticality classification accuracy, highlighting its potential to significantly advance the safety evaluation of object detection systems in automated vehicles.

Keywords: Vehicular Sensor, Resilient and Robust Sensing, On-Vehicle Sensor Networks
Abstract: Supervised 2D detection methods rely heavily on large amounts of labeled data to enable robust model generalization. However, manually annotating such datasets is inherently unscalable and prohibitively expensive. We propose a fully automated, training-free approach for 2D object annotation, built upon the capabilities of pre-trained zero-shot models. Our approach integrates a Vision-Language Model (VLM) for class extraction, Grounding DINO for zero-shot object detection, and the Segment Anything Model (SAM) for mask generation. A key contribution is the use of the VLM not to query against a full class list, but to first identify the specific object categories present in an image, drastically improving detection quality. Furthermore, we introduce a refinement step using Non-Maximum Suppression (NMS), eliminating duplicate predictions. We evaluate our pipeline on diverse benchmarks, including PASCAL VOC and COCO, demonstrating its effectiveness and generalization capabilities without any task-specific training. Our method offers a practical path towards accelerating dataset creation with minimal human supervision.

Keywords: Vehicular Sensor, Vehicular Signal Processing and Pattern Recognition, Driver Assistance Systems
Abstract: Indirect-time-of-flight (iToF) LiDAR sensors based on Amplitude-Modulated Continuous Wave (AMCW) technology and CMOS technology offer a cost-efficient solution for generating dense, accurate point clouds in near-field applications such as parking and navigating traffic jams. However, due to technology limitations, iToF sensors face critical challenges such as depth range limitations and ambiguities, motion blur noise, and multipath interference. In this dataset, we provide the recorded data for a single front view near field iToF LiDAR, with point cloud frames containing limitations like depth ambiguity, motion blur, and multipath interference, the corrected ground truth point cloud, and additional inputs, including grayscale images and signal-to-noise ratio frames for defective frames. This dataset aims to help researchers find solutions to overcome the technology limitations of the iToF sensors and encourage their use in autonomous mobile robots and automotive-grade applications.

Keywords: Vehicular Signal Processing and Pattern Recognition, Resilient and Robust Sensing, Inter-Vehicular Communication
Abstract: Collective Perception (CP) has emerged as a promising approach to overcome the limitations of individual perception in the context of autonomous driving. Various approaches have been proposed to realize collective perception; however, the Sensor2Sensor domain gap that arises from the utilization of different sensor systems in Connected and Automated Vehicles (CAVs) remains mostly unaddressed. This is primarily due to the paucity of datasets containing heterogeneous sensor setups among the CAVs. The recently released SCOPE datasets address this issue by providing data from three different LiDAR sensors for each CAV. This study is the first to address the Sensor2Sensor domain gap in vehicle-to-vehicle (V2V) collective perception. First, we present our sensor-domain robust architecture S2S-Net. Then an in-depth analysis of the Sensor2Sensor domain adaptation capabilities of state-of-the-art CP methods and S2S-Net is conducted on the SCOPE dataset. This study shows that, all evaluated state-of-the-art mehtods for collective perception highly suffer from the Sensor2Sensor domain gap, while S2S-Net demonstrates the capability to maintain very high performance in unseen sensor domains and outperforms the evaluated state-of-the-art methods by up to 44 percentage points.

Keywords: ICT in Road Safety and Infrastructure, Vehicular Signal Processing and Pattern Recognition, Resilient and Robust Sensing
Abstract: We present a fully interpretable and flexible statistical method for background subtraction in roadside LiDAR data, aimed at enhancing infrastructure-based perception in automated driving. Our approach introduces both a Gaussian distribution grid (GDG), which models the spatial statistics of the background using background-only scans, and a filtering algorithm that uses this representation to classify LiDAR points as foreground or background. The method supports diverse LiDAR types, including multiline 360-degree and micro-electro-mechanical systems (MEMS) sensors, and adapts to various configurations. Evaluated on the publicly available RCooper dataset, it outperforms state-of-the-art techniques in accuracy and flexibility, even with minimal background data. Its efficient implementation ensures reliable performance on low-resource hardware, enabling scalable real-world deployment.

Keywords: Resilient and Robust Sensing, Image Sensor, Vehicular Signal Processing and Pattern Recognition
Abstract: Automotive cameras are widely used to provide adequate data for perception tasks in Automated Driving Systems (ADS). The massive in-vehicle deployment has made it challenging to transmit real-time generated visual data on wired vehicular communication networks. Recent studies have shown Bayer compression to be a promising cost-effective solution to real-time reduce multi-camera data and transmit it. However, there is a lack of research investigating, in the Bayer domain, the correlation between compression and non-ideal data, e.g. noisy data. This paper conducts a comprehensive investigation of camera noise factors (both internal and external) on Bayer images, particularly focussing on relationship with downstream tasks such as compression and then perception, namely object detection. We assembled a modified JPEG codec with a lightweight Bayer adaptation technique suitable for real-time implementation. An ad hoc synthetic noisy dataset is constructed, including different key camera noise models. We used state-of-the-art detection models to investigate the impact on object detection-based perception. The results demonstrate that the different detector architectures show different levels of robustness to different types of noise. It was also observed that low quality compression in noisy image data does not degrade but actually improves the IQA and detection performances. This work presents a strong basis to optimise the efficiency camera data pipeline.

Keywords: Driver Assistance Systems, Image Sensor, Vehicular Sensor
Abstract: Autonomous vehicle (AV) researchers are increasingly exploring the use of monocular depth estimation (MDE) models for real-time perception, in combination with other depth-sensing technologies such as LIDAR and radar. However, MDE models remain vulnerable to adversarial attacks, which can degrade their accuracy and lead to potentially hazardous misinterpretations. This paper investigates the susceptibility of MDE models to 2D adversarial patch attacks and proposes methods to improve model robustness. In this work, we test the effects of an optimized 2D adversarial patch derived from the 3D2Fool algorithm on state-of-the-art MDE models such as Monodepth2, DenseDepth, DepthAnythingV2, and ZoeDepth. Our experiments reveal varying levels of vulnerability among the models, with significant disruptions in the depth predictions. To counter these attacks, we implemented adversarial training, which significantly improved model resilience. Our results demonstrate that adversarial training can substantially enhance the robustness of MDE systems, thus contributing to the safety and cybersecurity of AV technologies.

Keywords: Image Sensor, Vehicle Testing, Vehicular Signal Processing and Pattern Recognition
Abstract: Vision systems in automated driving deliver information on which subsequent processing stages rely to perform safety-critical decisions. With the application of deep-learning models, more effort is needed to validate the sensor models used in virtual environments, as it is not known which features of the image data affect the prediction. This work proposes a validation method based on a standardized image-quality target according to ISO 12233:2023 and ISO 14524, together with a high-precision geometrical digital twin for direct comparison between real and simulated images taken of a driving scenario. The proposed approach is applied to the default camera sensor model in CARLA and targets the geometric and the radiometric parts separately. Thereby, this work proposes an enhanced model that fits the characteristics of the real camera used during experiments. Images generated with the proposed simulation model reduce the sim-to-real gap and are 112.73% better than the default model in terms of the Michelson-contrast. Evaluating the OECF-curves yields an 85.16% improvement. To demonstrate the applicability, we assess the performance of a deep neural network detecting a EuroNCAP pedestrian target at distances ranging from 5m to 100m. The results show that the default simulation model leads to overconfident predictions compared to the proposed model and the real camera. Furthermore, we will use the enhanced model in X-in-the-Loop tests of automated driving systems.

Keywords: Driver Assistance Systems
Abstract: This study investigated the long-term effects of a voice-based stepwise driver-assistance system providing either real-time feedback alone or feedback plus advice. Ten experienced drivers completed six weekly drives on an urban route with two accident-prone areas. Weekly acceptance ratings and behavioral readiness metrics (stepwise score, mean speed, throttle-off count, foot-on-brake count, brake-on count, and lateral checks) were collected. Acceptance was analyzed via two-way ANOVA (Week × Group); behavioral retention in a pre feedback zone (Area 1) was assessed using linear mixed-effects models. Although both systems showed good acceptance, feedback with advice yielded significantly higher and more stable usefulness ratings. Advice recipients showed larger, sustained speed reductions (3.5 - 4.4 km/h from Week 4, p < 0.01) and persistent readiness-behavior improvements, unlike the feedback-only group, which showed limited improvement. Notably, in Area 1, advice-trained drivers maintained these improved behaviors without active feedback, indicating habit internalization. Thus, stepwise feedback and advice tailored to driver readiness may lower learning difficulty and psychological reactance to enhancing readiness behavior. Overall, this spaced-learning approach using feedback with advice appears effective for attenuating alert fatigue, maintaining acceptance, and transforming momentary corrections into lasting safe driving habits, even in experienced drivers.

Keywords: Active and Passive Safety Systems, Driver Assistance Systems
Abstract: Driving under the influence of alcohol (DUI) remains a major contributor to traffic accidents worldwide, with risk increasing exponentially as blood alcohol concentration (BAC) rises. This study investigated 120 participants (80 men, 40 women; aged 18–65) in a controlled driving simulation. Each completed two sessions, one sober and one with a BAC between 0.6 ‰ and 1.0 ‰, to assess alcohol-related impairment. Subjective self-assessment was measured using pre- and post-driving questionnaires, while objective impairment was evaluated through baseline tests including reaction time, walk-and-turn, and time estimation. Breath alcohol concentration (BrAC) was recorded as a reference. Statistical analyses (t-tests, MANOVA, multiple regression) examined performance differences and associations between BrAC, test results, and self-assessments. Results revealed significant impairments in motor and cognitive performance under alcohol, with the walk-and-turn test showing the strongest correlation with BrAC. Self-assessments also reflected intoxication, suggesting awareness of impairment. Unsupervised k-means clustering identified three distinct impairment profiles, showing that functional impairment cannot be determined by BrAC alone. The findings provide a validated ground truth for developing detection algorithms based on behavioral indicators such as driving and gaze behavior. This study lays the foundation for data-driven approaches to DUI prevention and context-aware driver support.

Keywords: Road Accident Investigation and Risk Assessment
Abstract: Safe autonomous navigation in pedestrian-rich environments demands more than collision avoidance—it requires the prioritization of pedestrian and other vulnerable traffic participants. In this paper, we present a real-time framework that enables vehicles to make decisions based not only on predicted Time-To-Collision (TTC), but also on the relevance of each pedestrian’s intention within a multi-hypothesis predictive model. The system builds upon a multi-pedestrian, multi-intention tracker that employs a Probability Hypothesis Density filter enhanced with a Generalized Potential Field Approach (PHD-GPFA) to estimate multiple future trajectories and intentions per pedestrian. Each trajectory hypothesis is associated with a belief weight that captures intention confidence. Unlike traditional risk assessment methods that rely solely on TTC, we propose a joint relevance metric that fuses TTC with belief-based intention likelihoods, enabling the vehicle to identify and respond to the most critical interactions in the scene. At every simulation timestep, vehicle speed is dynamically adjusted based on the most relevant pedestrian hypothesis. We evaluate the system in a SUMO-based simulation across multi-agent pedestrian scenarios, where it demonstrates proactive and context-sensitive responses to dynamic pedestrian behaviors. Our results highlight the advantages of integrating intention confidence into risk-aware planning.

Keywords: Driver Assistance Systems
Abstract: Rainy weather significantly increases the risk of road accidents due to reduced visibility and vehicle traction. Understanding how experienced drivers adapt their visual perception through gaze behavior under such conditions is critical for designing robust driver monitoring systems (DMS) and for informing advanced driver assistance systems (ADAS). This case study investigates the eye gaze behavior of a driver operating the same highway route under both clear and rainy conditions. To this end, gaze behavior was analyzed by a two-step clustering approach: first, clustering gaze points within 10-second intervals, and then aggregating cluster centroids into meta-clusters. This, along with Markov transition matrices and metrics such as fixation duration, gaze elevation, and azimuth distributions, reveals meaningful behavioral shifts. While the overall gaze behavior focused on the road with occasional mirror checks remains consistent, rainy conditions lead to more frequent dashboard glances, longer fixation durations, and higher gaze elevation, indicating increased cognitive focus. These findings offer valuable insight into visual attention patterns under adverse conditions and highlight the potential of leveraging gaze modeling to aid design of more robust ADAS and DMS.

Keywords: Vehicular Signal Processing and Pattern Recognition, On-Vehicle Sensor Networks, Driver Assistance Systems
Abstract: Hands-on-wheel detection plays a critical role in improving safety within Autonomous Driving Systems, as it helps verify the driver's readiness to retake control when necessary. This becomes especially important in intermediate levels of automation (such as SAE Levels 2 and 3), where the vehicle is not fully autonomous and cannot manage all driving situations. In such cases, timely driver intervention is required to ensure safe operation during transitions between automated and manual control. This paper presents and compares two different methodologies for automatic Hands-on-wheel detection. The first leverages wearable sensors, specifically, wrist-mounted devices to capture motion data, which is then processed using a Long Short-Term Memory neural network for binary classification. To validate the performance of the proposed system, it was compared with a camera-based computer vision system trained with convolutional models to detect hand presence from visual input. Experimental results showed that wearable-based systems provide a viable alternative to vision-based solutions, offering benefits such as enhanced user privacy. These findings position wearable devices as a viable alternative for integration into real-time driver monitoring frameworks.

Keywords: On-Vehicle Sensor Networks, Vehicular Signal Processing and Pattern Recognition, Driver Assistance Systems
Abstract: In conditional autonomous driving systems, one of the primary challenges is ensuring that drivers respond in a timely and safe manner when their intervention is required, especially if they have lost situational awareness. This study investigates the variation in physiological indicators among individuals during autonomous driving. To this end, simulated scenarios were designed in which participants experienced two conditions: autonomous driving with active environmental observation, and autonomous driving while performing cognitive distraction tasks. Electrocardiogram signals and electrodermal activity were collected to identify activation patterns during the experiment's progression. These findings provide valuable information for the development of artificial intelligence models that can estimate a driver's attentional state in real-time, thereby improving safety and human-machine interaction. The results show significant differences in physiological responses among individuals, highlighting the importance of considering personalized profiles.

Keywords: Active and Passive Safety Systems
Abstract: This paper proposes a framework for integrating the subjective safety of vulnerable road users into autonomous driving (AD) systems' decision-making. Cognitive-psychological data on cyclists are collected in laboratory experiments and validated in an empirical field test. The identified subjective safety parameters ensure the optimal distance and velocity for AD systems, represented by a mobile robot. Approaching decision-making design from a human-centred perspective ensures safety, comfort, and high acceptance of AD systems.

Keywords: Road Accident Investigation and Risk Assessment, Telematics, Active and Passive Safety Systems
Abstract: Repeated driving along the same route is common in the real world and promotes familiarity, which significantly influences driving behavior and safety. While prior studies have characterized familiarity using aggregated route-level trends, little attention has been given to how it manifests across diverse scenarios and individuals. This study adopts a counterexample-driven approach to examine whether such global summaries adequately reflect nuanced behavioral adaptations across contexts and drivers. A preliminary study was conducted using a driving simulator with eye tracking. Participants repeatedly drove a virtual urban route containing intersections, curves, and regulatory signs. Gaze behavior was analyzed using stationary and dynamic metrics to capture changes in visual attention. Results show that while global route-level trends indicate increased driving speed and decreased gaze entropy, scenario- and individual-level analyses reveal substantial variability. Notably, the common assumption that fixation duration on traffic signs uniformly decreases with increasing familiarity is challenged. These findings highlight the limitations of relying solely on global analyses and underscore the value of incorporating scenario- and individual-level perspectives to better understand how driving familiarity develops. This shift may inform the design of adaptive, human-centric autonomous systems with improved safety and efficiency.

Keywords: Driver Assistance Systems, Road Accident Investigation and Risk Assessment, ICT in Road Safety and Infrastructure
Abstract: Mixed traffic in African cities creates unique safety challenges for motorcycles and cars. This paper presents the first large-scale study of how drivers respond to motorcycle encounters in African urban traffic. We analyzed 568 dashcam videos that contain 36,658 encounters in Kigali, Rwanda. Our analysis shows that six geometric encounter types reduce to three behavioral strategies: threat avoidance deceleration for crossing encounters (13. 8% speed reduction), traffic flow acceleration for stationary and overtaking scenarios (5.1% to 11.0% speed increase), and neutral maintenance for oncoming and following encounters (3.8% to 5.4% reduction). ANOVA confirms significant differences between types ((F (5, 36651) = 8.742, p < 0.001). These findings help ADAS systems prioritize alerts by identifying crossing encounters as primary threat scenarios requiring intervention. Reaction times (0.33-0.86 seconds) provide baselines for autonomous vehicle calibration in mixed traffic environments. The behavioral simplification of six types of encounters into three strategies significantly advances the design of safety systems for complex urban environments

Keywords: Vehicle/Engine Control, Active and Passive Safety Systems, Driver Assistance Systems
Abstract: As autonomous driving and active safety systems advance, maintaining robust vehicle motion control near performance limits has emerged as an essential requirement. This paper presents an improved dynamic inversion-based control framework, in which Model Predictive Control is adopted in the top layer to serve as a motion controller explicitly considering coupled actuator boundaries.To accurately approximate the Attainable Force Subset, a novel method is proposed that incorporates realistic tire force constraints and approximates tire friction circles using polygonal representations, effectively capturing the coupling between longitudinal and lateral tire forces. The resulting Attainable Force Subset formulation is embedded into the MPC framework, enabling constraint-aware reference generation and flexible prioritization among competing control objectives. The proposed approach is validated through challenging maneuvers involving sinusoidal variations in reference speed and slip angle. Numerical results obtained with CasADi demonstrate significantly improved path tracking accuracy and yaw stability compared to the baseline scheme embedded with post-processing correction. Furthermore, compared to nonlinear MPC incorporating full vehicle dynamics, the proposed approach achieves comparable control performance while offering superior computational efficiency and solve-time consistency, thus making it well-suited for real-time implementation of vehicle motion control at handling lim

Keywords: Vehicular Sensor, Vehicular Signal Processing and Pattern Recognition, Driver Assistance Systems
Abstract: Modern vehicles rely more and more on sophisticated sensor systems for navigation and dynamics monitoring, however most approaches depend on centralized, expensive sensor suites. This paper presents a novel approach to vehicle state estimation using strain sensors mounted on wheel rims. We demonstrate that strain patterns captured at the wheel-road interface contain rich information about vehicle dynamics that can be extracted using machine learning (ML) techniques. We collected high-resolution telemetry data from strain sensors mounted on all four wheels of a test vehicle, synchronized with ground truth generated from GPS signal for supervised training. Using ML methods, we successfully predicted vehicle speed and turning direction directly from the strain sensor data. Our results show that wheel-mounted strain sensors can serve as an independent, redundant source of vehicle dynamics information, with potential applications in sensor fusion, fault detection, and low-cost navigation systems. This work is a first step for future research into vehicle driving scenarios identification and road condition monitoring using the same sensor configuration.

Keywords: Active and Passive Safety Systems, Multi-Vehicle Systems, Navigation and Localization Systems
Abstract: This paper presents a hybrid framework for a safe and autonomous Multi-Vehicle system coordination in dynamic, complex settings with priority-based tasks. The system combines A* path planning algorithm with Q-learning-based reinforcement learning to ensure an efficient and safe path. It also applies a Market-based task allocation mechanism that assigns tasks based on the highest bidding, considering the following: priority, urgency, and vehicle proximity. Each vehicle, once assigned to a task, operates independently and adapts to any environmental changes. The results show efficiency in training, collision reduction, and efficient task completion. This work contributes a communication-free approach for real-time fleet coordination with collision avoidance. Moreover, it may be applied to various multi-vehicle applications.

Keywords: Driver Assistance Systems, Vehicle Testing, Road Accident Investigation and Risk Assessment
Abstract: Research on lane change prediction has gained a lot of momentum in the last couple of years. However, most research is confined to simulation or results obtained from datasets, leaving a gap between algorithmic advances and on-road deployment. This work closes that gap by demonstrating, on real hardware, a lane-change prediction system based on Knowledge Graph Embeddings (KGEs) and Bayesian inference. Moreover, the ego-vehicle employs a longitudinal braking action to ensure the safety of both itself and the surrounding vehicles. Our architecture consists of two modules: (i) a perception module that senses the environment, derives input numerical features, and converts them into linguistic categories; and communicates them to the prediction module; (ii) a pretrained prediction module that executes a KGE and Bayesian inference model to anticipate the target vehicle's maneuver and transforms the prediction into longitudinal braking action. Real-world hardware experimental validation demonstrates that our prediction system anticipates the target vehicle's lane change three to four seconds in advance, providing the ego vehicle sufficient time to react and allowing the target vehicle to make the lane change safely.

Keywords: Driver Assistance Systems, Vehicular Signal Processing and Pattern Recognition
Abstract: The lateral movement of vehicles is an important indicator for the prediction of cut-ins in automated driving (AD) functions, and a relevant factor for the effective perception range of AD sensors. With simulations being an integral part of the validation of AD functions, models to realistically reflect the lateral movement of vehicles are crucial in order to generate realistic inputs for the AD system’s sensors and algorithms. Earlier work therefore proposed a two-level stochastic model for the lateral movement of vehicles on highways, which was, however, limited to lane following maneuvers. Within this work, the model is extended towards a full lateral movement model for highway scenarios by extending it towards lane changes. The proposed complete model represents a consistent generalization of the previous lane following model, in sharing model components and parameters, and in maintaining a measurably high degree of realism and efficiency in simulation.

Keywords: Driver Assistance Systems, Navigation and Localization Systems
Abstract: On and off-ramps are understudied road sections even though they introduce a higher level of variation in highway interactions. Predicting vehicles’ behavior in these areas can decrease the impact of uncertainty and increase road safety. In this paper, the difference between this Area of Interest (AoI) and a straight highway section is studied. Multi-layered LSTM architecture to train the AoI model with ExiD drone dataset is utilized. In the process, different prediction horizons and different models’ workflow are tested. The results show great promise on horizons up to 4 seconds with prediction accuracy starting from about 76% for the AoI and 94% for the general highway scenarios on the maximum horizon.

Keywords: Inter-Vehicular Communication, ICT in Road Safety and Infrastructure, Road Accident Investigation and Risk Assessment
Abstract: Intelligent Transportation Systems (ITS) are increasingly leveraging Digital Twin (DT) technologies to address urban mobility challenges through advanced sensing and real time communication. This paper presents a comprehensive framework that integrates vehicle-based LiDAR and camera sensing, V2X communication compliant with ETSI CAM and CPM standards, and high-fidelity virtual reconstruction using the DT simulator. The proposed system enables real-time reproduction and analysis of driving environment by transmitting processed object detections from vehicles to a cloud-based DT platform. In a first step a simulation with CARLA is used for high-fidelity virtual representation. To assess the feasibility and scalability of this approach, a detailed simulation environment was developed using the Veins framework, combining SUMO for traffic modeling and OMNet++ for network simulation. A case study is conducted on a specific urban environment in Leganes, Spain, with the simulations performed across varying vehicle densities to evaluate the performance of V2X-based CAM and CPM message dissemination. Results demonstrate the effectiveness of CPM-based data exchange for dynamic situational awareness and highlight the key insights into channel load and packet loss under different traffic load.

Keywords: Vehicle Testing, Driver Assistance Systems
Abstract: Hardware-in-the-Loop (HiL) simulation is a well-established technique for system validation in automotive development. As autonomous vehicles become more prevalent, the ability to demonstrate the credibility of these HiL systems becomes increasingly important. A key factor influencing this credibility is the verification of the simulation models used as a subsystem within the HiL environment. The increasing complexity of vehicle software leads to more complex simulation models, which in turn significantly raises the effort required for their verification. At the same time, with the shift towards software-defined vehicles, the demand for faster and more frequent incremental software updates is growing. This trend presents a substantial challenge for providers of HiL simulation models, who must keep pace without impairing the ability to demonstrate and document the credibility of the simulation models for every release. In this context, regression testing plays a crucial role in incremental development processes of simulation models, as it ensures that newly introduced changes do not negatively impact existing system behavior. To address this, this paper introduces a regression-based test case selection method to increase the efficiency of HiL simulation model verification. The approach performs graph-based signal flow analysis of consecutive model releases to identify relevant vehicle test cases for the re-verification of a new model increment.

Keywords: Driver Assistance Systems, Telematics, Vehicle/Engine Control
Abstract: This paper presents the design and implementation of a robust and modular teleoperation system for Intelligent Vehicles (IVs), intended to serve as a fallback solution when autonomous vehicles (AVs) encounter operational limitations. Built using modern technologies such as WebRTC, LTE 4G and node-based architectures, the system enables real-time remote driving with low latency and high reliability. The proposed solution was deployed and validated on a real research vehicle, controlled from a remote teleoperation station. Both qualitative and quantitative evaluations were performed, including latency and usability analysis. The results demonstrate that the system is reliable, scalable and suitable for real-world teleoperation applications, offering a solid foundation for future research in AVs and remote driving systems.

Keywords: Vehicle Testing, Active and Passive Safety Systems, Driver Assistance Systems
Abstract: The development of Automated Driving Systems demands extensive testing to ensure passenger safety. This testing follows a stepwise approach, progressing from purely virtual environments to open real-world traffic. This work proposes an X-in-the-Loop architecture that enables a seamless transition between different testing modalities. The architecture relies on open standards and open-source algorithms, enhancing accessibility for developers. A proof of concept is implemented for a subset of the architecture, covering vehicle dynamics, data exchange between systems, and feedback of the vehicle motion into the simulation. System credibility within the Hardware-in-the-Loop setup is assessed against Vehicle-in-the-Loop tests. The results confirm the applicability of the proposed architecture. Moderate deviations in vehicle dynamics behaviour are observed, attributed to model simplifications. Nonetheless, the control algorithm parameters tuned using the simulation model perform effectively on the real vehicle. Future work will focus on integrating sensor models and perception systems into the test chain. This will support validation of the architecture in a fully closed-loop testing environment.

Keywords: Inter-Vehicular Communication, On-Vehicle Sensor Networks, Vehicular Sensor
Abstract: This paper presents a real-time, low-latency communication and energy- efficient framework for IoT-enabled micro-mobility systems, specifically targeting gated environments such as residential campuses, educational institutions and industrial zones. As a core application platform, an IoT-enabled Segway is deployed and integrated with the proposed framework, serving as a practical micro-mobility solution within gated communities. The proposed system enables both Vehicle-to-Vehicle (V2V) communication using ESP-NOW, and Vehicle-to-Infrastructure-to-Vehicle (V2I2V) communication using Wi-Fi-based UDP, implemented using ESP32-S3 and NodeMCU microcontrollers integrated with IMU, ToF, Ultrasonic, and GPS sensors. To evaluate performance, end-to-end latency is measured across varying distances under both Line-of-Sight (LoS) and Non-Line-of-Sight (NLoS) conditions. Along with the latency measurement, power profiling system is also implemented using the INA226 sensor, comparing the energy consumption of ESP-NOW and 4G-based communication. To extend this analysis, Simulation of Urban MObility (SUMO) simulation was performed. Results demonstrate the effectiveness of the proposed framework in supporting reliable micro-mobility communication, offering a practical foundation for intelligent transport in localized smart environments. This work introduces a novel, energy-efficient, and infrastructure-independent V2X communication solution tailored for smart gated communities.

Keywords: Driver Assistance Systems, Vehicular Signal Processing and Pattern Recognition, Active and Passive Safety Systems
Abstract: This paper validates a real-time system for mitigating motion sickness using readily accessible vehicle CAN-bus data and an interpretable CatBoost model. This work makes two primary contributions from a study in a vehicle simulator. The first is a framework for generating and delivering interpretable driving recommendations. The model's output is used to identify the specific root causes of discomfort and translate them into clear, actionable advice, with the most effective communication method established through a comprehensive Human-Machine Interface (HMI) study. The second contribution is a semi-autonomous, closed-loop control system that performs direct interventions to enhance comfort. Crucially, this system uses the very same recommendations as a transparency layer for its automated actions; when a correction is made by the vehicle, the interpretable reason for the intervention is simultaneously communicated to the driver. The system's efficacy is demonstrated in a comparative study, wherein a substantial reduction in high and moderate motion sickness scenarios and a significant increase in comfortable driving time are observed. Through this research, a practical framework is validated in which automated interventions are rendered transparent, proving that enhanced comfort can be achieved in synergy with driver understanding.