Tokyo time | Wednesday, May 29 | Thursday, May 30 | Friday, May 31 |
09:00 ‑ 09:30 | O1: Opening | S5: Wireless Communications | K2: Keynote 2 |
09:30 ‑ 10:00 | K1: Keynote 1 | ||
10:00 ‑ 10:20 | C7: Networking Break | ||
10:20 ‑ 10:30 | S9: Simulation & Modeling 2 | ||
10:30 ‑ 10:50 | C1: Networking Break | C4: Networking Break | |
10:50 ‑ 12:10 | S1: Vehicular Network Applications | S6: In-vehicle Communications | |
12:10 ‑ 12:20 | L3: Lunch | ||
12:20 ‑ 13:15 | L1: Lunch | L2: Lunch | |
13:15 ‑ 13:25 | P1: Panel | ||
13:25 ‑ 14:15 | S2: Security in Vehicular Communications | S7: Cooperative Driving | |
14:15 ‑ 14:35 | C8: Networking Break | ||
14:35 ‑ 14:55 | S10: Vulnerable Road Users (VRUs) | ||
14:55 ‑ 15:15 | C5: Networking Break | ||
15:15 ‑ 15:35 | C2: Networking Break | S8: Short Papers 2 | |
15:35 ‑ 16:20 | S3: Simulation & Modeling 1 | ||
16:20 ‑ 16:25 | C6: Networking Break | ||
16:25 ‑ 16:30 | O2: Closing | ||
16:30 ‑ 16:40 | D1: Poster / Demo | ||
16:40 ‑ 16:45 | |||
16:45 ‑ 16:55 | C3: Networking Break | ||
16:55 ‑ 18:00 | S4: Short Papers 1 | ||
18:00 ‑ 18:30 | |||
18:30 ‑ 20:00 | Welcome Cocktail Hour | B1: Banquet | |
20:00 ‑ 21:00 |
Abstract: In the context of 5G and Beyond 5G, there is a strong emphasis on accommodating multi-use cases. Similarly, within the Vehicle-to-Everything (V2X) , the focus has shifted from standalone safety driving assistance services to a broader array of services. Notably, recent years have witnessed the implementation of features such as over-the-air (OTA) software updates for in-vehicle systems and data collection from onboard sensors to cloud servers, even in commercial vehicles. These developments are set to revolutionize the process of vehicle development. In this talk, we will spotlight our initiatives, emphasizing the impact of connected vehicles. We will delve into cooperative automated driving and remote driving as integral components of V2X's natural evolution. Additionally, we will discuss the future vehicle networkization necessary to support diverse multi-use cases.
Vehicle-to-everything (V2X) communication can enhance the capabilities of connected and automated vehicles (CAVs) to perform cooperative maneuver coordination (CMC), offering improvements in traffic safety, awareness, comfort and efficiency. Recent years have seen progress in the development of cooperative driving applications that leverage V2X maneuver coordination service and message (MCS and MCM), driven by both research and standardization efforts. However, the evaluation of such applications, along with defining relevant use cases and scenarios, lacks a common framework. This paper aims to fill this gap by proposing a framework for classifying use cases, scenarios and metrics for simulation-based performance evaluation of CMC applications. A simulation environment is employed to assess the effectiveness of the proposed metrics and evaluate a highway merging use case across various scenarios without coordination, with coordination, and in different communication conditions. Thus, this paper serves as a contribution to the development, testing and evaluation of CMC applications.
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.
The proliferation of rich automotive data contents keeps increasing the communication demand from / to connected vehicles, putting a strain on the limited bandwidth resources of cellular networks. Vehicle-to-Vehicle (V2V) communication holds promise to mitigate the load on the network infrastructure, as it enables vehicles to directly re-distribute the downloaded data contents to other vehicles, while gathering and aggregating sensor data from their neighbors before uploading the data to cloud computing platforms. However, the feasibility of V2V-assisted communications remains unclear in many aspects, with the biggest challenge being the lack of large-scale vehicle trace datasets, representing the mobility of privately owned vehicles, which can be used as input for network simulations. Although synthetic vehicle traces generated by a road traffic simulator could be a possible alternative to the real-world datasets, the plausibility of the network simulation results is largely attributed to the accuracy of the simulated road traffic volume along roads. In this paper, we develop a realistic vehicle mobility simulation model of the whole city of Nagoya, Japan. The road traffic volume is thoroughly calibrated on 1,618 road segments across the city, while traffic light cycles are also aligned with the historical data. The model can be executed in an open-sourced road traffic simulator SUMO, which can easily be interfaced with various network simulators like OMNeT++. 1.6 million vehicle trips are processed during a simulation for a 24-hour period, making it one of the largest SUMO traffic scenarios that are open to the research community.
The proliferation of Vehicle-to-Everything (V2X) networks and services is growing, driven by the increase of V2X communications on roadways. These services include navigation, entertainment, emergency assistance, and support for autonomous driving. Precise time accuracy forms the Foundation of the data generation in services like Intelligent Transport Systems (ITS). For this basis to be achieved, accurate time sources, and time synchronization must be achieved throughout the entire network. The high mobility of the networks results in high packet loss, asymmetrical latency, and jitter, that can influence traditional network protocols like Network Time Protocol (NTP). This paper proposes a set of practical time synchronization approaches in V2X scenarios, exploring the use of NTP, Precision Time Protocol (PTP), and Global Navigation Satellite System (GNSS) to synchronize the various network nodes to a given reference clock. Additionally, we also take into consideration the impact of synchronization tools, and technologies like Pulse-per-second (PPS) in GNSS. These approaches were tested in a city-scale platform, the Aveiro Tech City Living Lab (ATCLL), in static and mobility scenarios. The obtained results show that it is possible to achieve sub-Microsecond accuracy at all times in a vehicular network, even with high mobility and no network connection.
Inclusion of Vulnerable Road Users (VRUs), such as pedestrians and cyclists, to Vehicle-to-Everything (V2X) communication system introduces a new dimension to it by improving the road safety of these non-vehicle road users and reducing accidents on the road. At the same time, integrating VRU devices to the vehicular communication poses unique issues and challenges. Specifically from security perspective, enrollment and issuance of digital certificates to consumer devices such as smartphones to be a legitimate members of the V2X communication require consideration that may not be applicable to other entities such as vehicles. This paper presents an approach and a solution to address these challenges. Based on our observation, this is the first proposal to address the issues associated with security and privacy aspects of VRUs in the Intelligent Transportation System (ITS) system. We recommend that relevant standard defining organizations (SDOs) such as ETSI ITS and IEEE 1609 Wireless Access in Vehicular Environments (WAVE) to consider our solution toward standardization.
Hierarchical Federated Learning (HFL) faces the significant challenge of adversarial or unreliable vehicles in vehicular networks, which can compromise the model's integrity through misleading updates. Addressing this, our study introduces a novel framework that integrates dynamic vehicle selection and robust anomaly detection mechanisms, aiming to optimize participant selection and mitigate risks associated with malicious contributions. Our approach involves a comprehensive vehicle reliability assessment, considering historical accuracy, contribution frequency, and anomaly records. An anomaly detection algorithm is utilized to identify anomalous behavior by analyzing the cosine similarity of local or model parameters during the federated learning (FL) process. These anomaly records are then registered and combined with past performance for accuracy and contribution frequency to identify the most suitable vehicles for each learning round. Dynamic client selection and anomaly detection algorithms are deployed at different levels, including cluster heads (CHs), cluster members (CMs), and the Evolving Packet Core (EPC), to detect and filter out spurious updates. Through simulation-based performance evaluation, our proposed algorithm demonstrates remarkable resilience even under intense attack conditions. Even in the worst-case scenarios, it achieves convergence times at 63% as effective as those in scenarios without any attacks. Conversely, in scenarios without utilizing our proposed algorithm, there is a high likelihood of non-convergence in the FL process.
Intrusion detection systems (IDSs) have become a crucial component in ensuring the security of in-vehicle networks. With the emergence of machine learning (ML), more flexible and efficient statistical methods have been introduced to the field of IDSs. However, the current metrics used to evaluate IDSs, including those based on the Common Vulnerability Scoring System (CVSS), need to adequately capture the unique characteristics of in-vehicle networks. To address this issue, we propose a new metric based on CVSS, designed explicitly for in-vehicle IDSs. Our metric considers cybersecurity information, including the severity of attacks encountered, to provide a more informed representation of performance. By prioritizing high performance on severe attacks, our metric provides a more accurate representation of an IDS's effectiveness in the context of in-vehicle networks. This new metric bridges the gap between ML and the unique cybersecurity challenges of in-vehicle networks, enhancing the overall security posture of the automotive industry.
Connected cars are susceptible to cyberattacks. Security and safety of future vehicles highly depend on a holistic protection of automotive components, of which the time-sensitive backbone network takes a significant role. These onboard Time-Sensitive Networks (TSNs) require monitoring for safety and - as versatile platforms to host Network Anomaly Detection Systems (NADSs) - for security. Still a thorough evaluation of anomaly detection methods in the context of hard real-time operations, automotive protocol stacks, and domain specific attack vectors is missing along with appropriate input datasets. In this paper, we present an assessment framework that allows for reproducible, comparable, and rapid evaluation of detection algorithms. It is based on a simulation toolchain, which contributes configurable topologies, traffic streams, anomalies, attacks, and detectors. We demonstrate the assessment of NADSs in a comprehensive in-vehicular network with its communication flows, on which we model traffic anomalies. We evaluate exemplary detection mechanisms and reveal how the detection performance is influenced by different combinations of TSN traffic flows and anomaly types. Our approach translates to other real-time Ethernet domains, such as industrial facilities, airplanes, and UAVs.
The Federal Communications Commission (FCC) in the United States has reallocated 45 MHz from the 75 MHz spectrum previously reserved for ITS, for the use of unlicensed devices such as Wi-Fi. On the other hand, as V2X deployment and the number of V2X applications are expected to gradually increase, the remaining 30 MHz may not be sufficient for ITS communication, resulting in options seeking to share unlicensed spectrum with Wi-Fi. This paper introduces an analytical model and decentralized spectrum sharing protocol for V2X sidelink to coexist with standardized IEEE 802.11 Wi-Fi in the 5.9 GHz unlicensed spectrum. The protocol uses both theoretical projection and historical channel usage information to ensure fair coexistence between Wi-Fi and V2X sidelink. Simulation based evaluation indicates that this protocol improves the performance of both types of communications when sharing unlicensed bands.
In the vehicle-to-cloud (V2C) interactions using 3GPP sidelink, vehicles transmit data to a server on the Internet and receive responses in return. This paper shows through real-life measurement that bi-directional end-to-end communication latency in V2C can be dominated by that caused by sidelink communication if the application data arrivals and the sidelink transmit resources are not synchronized. It is counter to our expectation that the delay on the non-sidelink leg that spans the 5G RAN, 5G core network, and the Internet will account for the majority of the latency. For reducing inflated latency on the sidelink, this paper explores several solution approaches: reduced selection window size, decreased packet delay budget, and adaptive selection window postponement. Through real-life measurements, we show that such measures can slash sidelink latency inflation by more than 50% so that 5G V2X sidelink can be more easily applied to real-time applications built on V2C communication.
Vehicular communication is a key enabler in making Automated Vehicles (AVs) collaborate by sharing information, which complements on-board sensor information and facilitates precise vehicle control. This paper presents a tailored measurement campaign aimed at analyzing the performance of two vehicular communication technologies, namely IEEE 802.11p and LTE-V2X. Our study focuses on key metrics for cooperating AVs, such as end-to-end latency and packet delivery ratios. Additionally, we investigate the feasibility of channel coexistence, assessing the challenges associated with concurrent channel access. The results derived from field tests are correlated with simulations conducted on PLEXE and OpenCV2X, i.e., platforms used for simulating IEEE 802.11p and LTE-V2X, respectively. This combined methodology, comprising field tests and simulations, enables the attainment of replicable conclusions, which in turn enables better design choices.
Autonomous vehicles and driver assistance systems depend heavily on the quality of their perception algorithms and sensors. Although the advances in hardware and software have increased the confidence in perception data, the field-of-view limitation cannot be overcome by single-vehicle perception. Cooperative Perception aims to solve the field-of-view limitations by employing perception data transmitted over vehicular networks. The vehicular network is still in its early stages of real-world implantation. Therefore, the effectiveness of cooperative perception relating to the market penetration of connected vehicles and infrastructure remains to be seen. In addition to the number of vehicles transmitting perception data, the drive to reach a Cooperative, Connected, and Automated Mobility paradigm must contend with the inherent lack of confidence in network data that cannot be blindly trusted. In this work, we propose a Collective Perception Aggregation Pipeline as a basis for collective perception data verification. The pipeline's inputs accept any data source that produces high-level perception data, such as sensors, perception algorithms, and cooperative perception messages. We investigate network latency's effects on the pipeline's data fusion component through simulated vehicular scenarios. Although the increase in message volume caused a higher channel busy ratio, the increased latency did not affect the prediction error of the data fusion algorithm. In a simulated scenario of 90 connected vehicles transmitting collective perception data, with a channel busy ratio of more than 20% and more than 50% of perception messages arriving out-of-sequence, the fusion algorithm maintained a standard deviation of the prediction error of 0.5609 for the X position, 0.5627 for the Y position, and 0.6267 for the yaw of the object.
he quest for safer and more efficient transportation through cooperative, connected and automated mobility (CCAM) calls for realistic performance analysis tools, especially with respect to wireless communications. While the simulation of existing and emerging communication technologies is an option, the most realistic results can be obtained by employing real hardware, as done for example in field operational tests (FOTs). For CCAM, however, performing FOTs requires vehicles, which are generally expensive. and performing such tests can be very demanding in terms of manpower, let alone considering safety issues. Mobility simulation with hardware-in-the-loop (HIL) serves as a middle ground, but current solutions lack flexibility and reconfigurability. This work thus proposes ColosSUMO as a way to couple Colosseum, the world's largest wireless network emulator, with the SUMO mobility simulator, showing its design concept, how it can be exploited to simulate realistic vehicular environments, and its flexibility in terms of communication technologies.
Base station densification is one of the key approaches for delivering high capacity in radio access networks. However, current static deployments are often impractical and financially unsustainable, as they increase both capital and operational expenditures of the network. An alternative paradigm is the moving base stations (MBSs) approach, by which part of base stations are installed on vehicles. However, to the best of our knowledge, it is still unclear if and up to which point MBSs allow decreasing the number of static base stations (BSs) deployed in urban settings. In this work, we start tackling this issue by proposing a modeling approach for a first-order evaluation of potential infrastructure savings enabled by the MBS paradigm. Starting from a set of stochastic geometry results, and a traffic demand profile over time, we formulate an optimization problem for the derivation of the optimal combination of moving and static BSs which minimizes the overall amount of BSs deployed, while guaranteeing a target mean QoS for users. Initial results on a two-district scenario with measurement-based network traffic profiles suggest that substantial infrastructure savings are achievable. We show that these results are robust against different values of user density.
Relaying is an emerging strategy to improve reliability in millimeter-wave (mmWave) vehicular networks, which are susceptible to link outages caused by blockage. The benefits of relaying, however, may be limited by the time overhead and undesired beam directions resulting from the commonly used exhaustive beam sweeping and phased arrays. In this paper, we propose a beam training scheme with true time delay (TTD) arrays based on deep reinforcement learning (DRL). The algorithm leverages frequency dependent beam patterns, akin to a rainbow beam, to track relay vehicles with negligible time overhead and point toward the desired direction within the wide bandwidth. Numerical simulations shows that the proposed method outperforms state-of-the-art DRL-based relay selection algorithm using phased arrays, motivating further investigation.
The support of Cooperative Intelligent Transport Systems (C-ITS) services requires seamless interoperability between involved stakeholders. To this aim, the 5G Automotive Association has recently endorsed a Vehicle-to-Network-to-Everything (V2N2X) architecture trialed at national initiatives to support road traffic management V2X services. The architecture enables interoperability at the application level through a cloud-federated Information Sharing Domain (ISD) that supports data sharing and interoperability among stakeholders. This study analyses the possibility to support critical and latency-sensitive V2X services using 5G-based Vehicle-to-Network-to-Vehicle (V2N2V) communications over the federated cloud-based V2N2X architecture. The analysis considers the intersection collision avoidance (ICA) service as a case study and scenarios involving multiple Mobile Network Operators (MNOs) and Original Equipment Manufacturer (OEM) clouds. We show that the ICA requirements can be supported, provided connections with controlled latencies (under Service Level Agreements or SLAs) are established between the OEM clouds and the ISD. However, the small tolerance to latency variations can compromise the support of the critical and latency-sensitive V2X services over the federated cloud-based V2N2X architecture, and solutions are necessary to ensure the scalability of the system.
The advances in the automotive industry with the ever-increasing request for Connected and Autonomous Vehicles (CAVs) are pushing for a new epoch of networked wireless systems. Vehicular communications, or Vehicle-to-Everything (V2X), are expected to be among the main actors of the future beyond 5G and 6G networks. However, the challenging application requirements, the fast variability of the vehicular environment, and the harsh propagation conditions of high frequencies call for sophisticated control mechanisms to ensure the success of such a disruptive technology. While traditional Radio Access Networks (RAN) lack the flexibility to support the required control primitives, the emergent concept of Open RAN (O-RAN) appears as an ideal enabler of V2X communication orchestration. However, how to effectively integrate the two ecosystems is still an open issue. In this paper, we discuss possible integration strategies, highlighting the challenges and opportunities of leveraging O-RAN to enable real-time V2X control. Additionally, we enrich our discussion with potential research directions stemming from the current state-of-the-art and we provide preliminary simulation results that validate the effectiveness of the proposed integration
This research introduces a groundbreaking methodology leveraging power-domain Non-Orthogonal Multiple Access (NOMA) for enhancing Vehicle-to-Everything (V2X) communications in 6G networks. By prioritizing safety-critical Decentralized Event-Triggering Notification Messages (DENM) over Cooperative Awareness Messages (CAM) through differential power allocation, we ensure the integrity and timely delivery of vital data. The technique of superimposing different message types within a single subframe, coupled with Successive Interference Cancellation (SIC) at the receiver, substantially improves spectral efficiency, bandwidth utilization, and communication reliability. Our simulations reveal significant enhancements in Packet Delivery Rate (PDR), End-to-End Latency (E2E), and Spectral Efficiency over traditional Orthogonal Multiple Access (OMA) approaches, marking a pivotal advancement toward the realization of ultra-reliable and low-latency communication (URLLC) in smart transportation systems. This work not only paves the way for smarter, safer, and more efficient mobility solutions but also aligns with the evolving 6G standards, emphasizing our commitment to addressing the fundamental challenges in V2X communications.
In the context of 5G Vehicle-to-Everything (V2X) autonomous mode, 3GPP introduced a resource re-evaluation mechanism to reduce collision occurrences. This mechanism requires stations to perform an additional check for available resources before transmission, to detect possible late-arriving reservations from other stations. The findings of this paper confirm that, while mandatory, the resource re-evaluation brings minimal performance enhancements to the overall system. Building upon this observation, the paper proposes a standard-compliant resource allocation mechanism named Double Initial Transmission (DIT), which leverages resource re-evaluation more effectively. Focusing on scenarios with prevalent periodic traffic, the paper introduces an innovative approach: the first packet within a sequence of periodic packets is transmitted twice, utilizing two different resources. The first transmission reserves a resource for the subsequent one. After transmitting the first packet, the resource used for its initial transmission is released, while the second resource is reserved periodically for the following packets. This strategy significantly reduces the probability of periodic collisions and has a minimal impact on the overall channel occupation. Furthermore, in instances of potential collisions on the second resource, the re-evaluation mechanism conducted before transmission serves as a proactive measure, detecting and addressing conflicts before they occur. The paper aims to provide an analysis of this modified resource allocation strategy and its implications for improving the performance of 5G NR-V2X, particularly in the context of periodic traffic scenarios.
This work proposes a solution for software-defined vehicular networks that, combined with Cooperative-ITS messages, aims to manage a multihoming vehicular network with heterogeneous wireless technologies. The work starts with the integration of SDN (through the Open vSwitch (OVS)) in the different interfaces present in the On-Board Units (OBUs) and Road Side Units (RSUs) - ITS-G5, 5G and Ethernet. Then, several mechanisms for the network management, such as the uplink decision logic, are researched and integrated into the OBU's SDN controller. The proposed approach is tested in both laboratory and city environments to assess the correct behavior and robustness of the solution. The results show the advantage of SDN-based multihoming vehicular networks, where different OBU applications are distributed across the different available communication technologies. The overhead introduced by this solution can be considered negligible, and the duration of horizontal and vertical handovers managed by the SDN controller is lower than 20 ms.
In this paper, we explore the impact of different antenna positions and radiation patterns on the channel performance of an Integrated Communication and Sensing (ICAS) extended model in a highway environment. We have considered a 3-dB omnidirectional Transmitter (Tx) and a directive Receiver (Rx) for our analysis. The antenna radiation patterns of patch and horn antennas are used for the directive receiver. In our system, we have considered three different setups. In the first setup, Rx has a directive antenna (either horn or patch) installed on the front bumper of the car. In the second setup, the directive antenna is installed on the back bumper of the car. For the third setup, the radiation patterns for the front and back bumper of the car are combined, thus giving rise to Combined Gain Pattern (CGP). The channel responses for the front and back bumper-placed antenna setups are evaluated and compared with the respective CGP setup channel response. The performance of an ICAS system depends on the successful detection of targets. We have considered two types of scatterers in our system. The moving vehicles on the road represent Dynamic Targets (DT) whereas, stationary point reflection sources along the sides of the highway represent Diffuse (DI) scattering components. We have shown in our results that the CGP setup outperforms the front and back bumper-located antenna setups in terms of channel magnitude and accurate target detection capabilities. Moreover, CGP setup helps in modeling the environment better as compared to front and back-bumper-installed setups.
Automotive Electrical and Electronic (E/E) architectures are rapidly evolving. With the transition towards automotive Ethernet, service-oriented communication, and zonal architectures, secure in-vehicle communication becomes even more critical. In this paper, we systematically analyze the four most prominent security protocols considered for in-vehicle communication in E/E architectures: Secure Onboard Communication (SecOC), Media Access Control Security (MACsec), Internet Protocol Security (IPsec), and Transport Layer Security (TLS). In addition, we consider a security extension specifically proposed for an automotive communication middleware. Our analysis includes a formal security analysis of our combination of MACsec and access control for the MACsec-based architecture. We compare the protocols and give recommendations for their usage.
The In-Vehicle Infotainment (IVI) system of a modern connected vehicle is a critical attack point and must be thoroughly tested. In this paper, we introduce a comprehensive methodology for penetration testing specifically tailored to IVI systems in connected vehicles and demonstrate its applicability with the case study of testing the IVI Wi-Fi interface of a Volkswagen (VW) ID.3. We define a generic IVI model as the basis for specifying the test scope. The type, depth, and complexity of the tests to be performed can be determined using our proposed attacker model. This ensures that all relevant aspects are tested. Our methodology consists of seven phases and contains detailed descriptions of the tasks to be carried out in each phase. Following our methodology in the case study, we identified vulnerabilities in the VW IVI system that could allow attackers with Wi-Fi access to crash the system or manipulate volume settings.
Recently the in-vehicle networks have more and more become one of the most important components in vehicles, especially electric automobiles. Considering the high complexity and stringent requirements for reliability in automotive software and hardware, it is highly desirable to test new applications in an emulated environment before deploying them in real vehicles for both Original Equipment Manufacturers (OEMs) and researchers. By doing so, hidden problems with high impacts may be identified before production.
In this paper, we demonstrate a Systematic Automotive Network Emulation (SANE), which integrates functionalities from the operating system layer to the network application layers. Recognizing the complexity of in-vehicle network environments, we place great importance on the scalability of the framework and make it easy to deploy and modify even on a large scale. We present and analyze a series of experimental results with different configurations to find the impacts of various factors like in-vehicle Electrical/Electronic (E/E) architecture on network performance and the trade-offs between different Quality of Service (QoS) strategies, while also confirming the feasibility of the framework. Our evaluations validate that zone-based architecture has better performance than domain-based architecture in the same configuration.
We tackle the challenges of accurately replicating modern automotive network architectures, particularly those reliant on automotive Ethernet and Time-Sensitive Networking (TSN), in simulation environments. We describe an open-source TSN simulation model tailored to match a real-world TSN testbed built using off-the-shelf hardware. We address key challenges such as hardware and software imperfections as well as varying data traffic patterns from automotive sensors. By incorporating real data traces from LIDAR and camera sensors, we improve simulation accuracy. Our findings underscore the importance of accounting for hardware and software nuances to ensure faithful simulation results, thus advancing the reliability of automotive network simulations.
Research on platooning has tackled many different facets of the topic, from lateral and longitudinal control to algorithms for optimization of platoon size and coordination. However, very few works started addressing the following two foundational questions. The first: Given a stretch of road where platooning is enabled and some form of V2X communication is feasible, will vehicles be able to organize locally in platoons without the need of central intervention? The second: Is a coordination protocol for platoon formation possible, or will traffic dynamics prevent the efficient formation of platoons? This paper gives a first, positive answer to these questions: A platoon formation protocol is proposed and tested upon DSRC communications, though it can be implemented on top of any transmission technology. We assess the performance in terms of platoon formation efficiency, on a multi-lane highway, as a function of the penetration rate of platooning-enabled vehicles and traffic characteristics. Realistic simulation results highlight the properties of the protocol as well as the impact of different traffic parameters, foremost the penetration rate and maximum distance between vehicles considered to start a platoon negotiation. These initial results pave the road for more sophisticated analysis, enhancements of the protocol and evaluation of advanced -possibly centralized- approaches to improve platoon management to achieve safer roads with increased capacity.
This paper investigates the effectiveness of incorporating intent sharing messages into Cooperative Adaptive Cruise Control (CACC) systems. In contrast to traditional information exchange limited to current state information, intent sharing involves providing details about the future trajectory of connected vehicles. This work employs a concise representation of the intent of a connected vehicle, encompassing anticipated speed and acceleration bounds. The proposed approach, referred to as intent sharing-based CACC (I-CACC), utilizes a reinforcement learning-based controller that leverages this additional intent information. Through an extensive simulation study using experimental datasets, we compare the performance of I-CACC to conventional CACC. The results reveal the superior performance of I-CACC across various metrics, encompassing safety, comfort, string stability, and gap-keeping.
Connected and automated vehicles (CAVs) have demonstrated their capacity to address the safety and sustainability challenges inherent in our existing transportation systems. Cooperative adaptive cruise control (CACC) is an example of cooperative driving among CAVs, leveraging vehicle-to-vehicle (V2V) communication to enhance the car-following capabilities. In general, CACC utilizes preceding vehicle's status data, e.g., most recent position, velocity and acceleration, as a feedforward signal to its controller to maintain a desired car following distance from the preceding vehicle. In this work we extend the car following framework from status-sharing to intent-sharing, where the preceding vehicle's future trajectory is shared in addition to the latest state information. In fact, intent-sharing communication is currently being standardized in the United States (SAE) and Europe (ETSI). Then, we present two receding horizon control frameworks that utilize intent sharing to minimize spacing errors while considering string stability and promoting smooth driving behavior for the ego vehicle. We then quantify the benefits of intent sharing and discuss the improvements of our frameworks over traditional methods. The results are demonstrated through microscopic simulations using real world highway data.
Traffic efficiency and air pollutants are a major issue in urban transport systems, as they are critical points for traffic flows and people living in urban areas, respectively. In response to this challenge, connected and autonomous vehicles (CAVs) have garnered increasing attention, with the potential to address the above challenges through cooperative manoeuvring and real-time communication. In this paper, we consider an intersection served by a multi-access edge computing (MEC) where CAVs communicate with a controller to pass the intersection without the need of signals and without stopping unless in presence of heavy congestion. For this purpose, we propose a first-in-first-scheduled (FIFS) algorithm and we specify the communication protocol for its practical implementation. The solution is validated through simulations with varying traffic densities, showing that it provides significant gains in travel time and fuel consumption compared to both uncontrolled intersections, with or without traffic lights, and a benchmark solution derived from the literature.
The softwarization of vehicles and the evolution towards autonomous driving is imposing increasing flexibility and reliability demands to future in-vehicle networks (IVN). Research on 6G advocates for a seamless integration of vehicles with cellular networks for a deep edge-edge-cloud continuum that facilitates the opportunistic offloading of in-vehicle processing to the edge/cloud. Realizing this vision requires a seamless connection of IVNs with the cellular networks, which can be facilitated through the gradual adoption of in-vehicle wireless subnetworks. These subnetworks can support increasing dependable and deterministic service levels using predictive schedulers that can anticipate in-vehicle traffic flows and patterns to schedule communication resources and computing workloads. This requires an accurate characterization of IVN traffic, and this study progresses the state-of-the-art with a first characterization of IVN traffic in autonomous vehicles. The study characterizes the data captured by a full suite of sensors as well as the processed data for supporting automated driving. We also derive spatial and time correlations between the IVN data that can serve to anticipate network demands and predict traffic flows for the support of deterministic IVN services.
The use of Reinforcement Learning (RL) in au- tonomous vehicles (AVs) is expected to enhance the safe ma- neuvering of AVs to improve road safety. However, existing literature on AVs focuses on the impacts of image perturbations as adversarial examples (AEs) during the testing phase. Limited attention has been given to more intrusive type of AEs such as vehicles with adversarial intent to proactively induce accidents on the road. Without addressing this type of AEs, the learned policy remains vulnerable to different types of AEs, making the RL not usable in AVs in practice given the nature of cyber- physical systems (CPS), for which negative consequences includes accidents, property losses, and injuries. To address this gap, we focus on the training phase and fortify the learned policy using our expanded AE definitions. In this paper, we present our approach to realize this training model to build a more robust policy against adversaries.
Smart autonomous vehicles can cooperatively drive as platoons that offers benefits like enhanced safety, traffic efficiency and fuel conservation. While traditionally platoons have followed a single-lane, train-like structure they face challenges when scaling that include communication range limitations and lane-change difficulties. In this article, we propose a new paradigm of multi-lane platoons that spreads platoons across multiple lanes. We explore the characteristics of multi-lane platoons particularly focusing on communication parameters. Additionally, we propose a cross-layer mechanism to seamlessly integrate this concept within the existing communication standard, ETSI. Our work significantly enhances platoon communication performance in mixed traffic scenarios and we propose optimizations to improve its effectiveness.
Cooperative adaptive cruise control systems are utilized to decrease inter-vehicle distances (IVDs) in vehicle platoons by using wireless communication. Leader-follower architecture is designed to compensate for the communication time delay by using the Smith predictor (SP) and further decrease the IVDs. However, the leader-follower architecture creates a need of adding feedback communication, which has drawbacks such as higher communication traffic. This study proposes a novel one-vehicle look-ahead architecture without feedback communication. The proposed method introduces an artificial delay and the SP. Simulation results show that the proposed method provides the same acceleration and IVD responses as the leader-follower architecture, other than the additional artificial delay in all the vehicles in the platoon.
Advanced Driver Assistance Systems (ADAS) and Autonomous Driving (AD) offer a range of economic, social, and environmental benefits. However, the European Union prioritizes the enhanced safety it brings to Intelligent Transport Systems (ITS). With the increasing complexity of ADAS/AD functionalities, platforms capable of handling extensive validation tests are needed. Conventional testing methods become impractical due to the exponential growth in integration and validation tests. Model-based approaches offer a pathway for digital disruption, enabling more efficient validation processes. Hence, the emphasis lies on executing extensive test sets and minimizing them through constructive means using correct-by-construction methodologies. To address these problems, we focus on two aspects that we identified as essential and at two distinct levels: A strategic platform that supports ADAS/AD system development through testing and validation methods, including co-simulation of AD scenarios using CARLA and Artery V2X simulators connected via a ROS network, continuous integration, and verification usage for streamlined validation processes; Establishing a highly connected and intelligent ecosystem with V2X communication infrastructure aiming at accelerating technologies and advancing infrastructure for Autonomous Driving Systems (ADS).
This paper presents a system that detects and identifies vehicles with three-dementional (3D) markers from data acquired by LIDAR sensors.This paper shows the feasibility of our system through an experiment.
Platooning has been researched for decades but debate about its lasting impact is still ongoing. Meanwhile, adaptive cruise control (ACC) became de facto standard for all new cars as well as for automated driving on the freeway. An evaluation of the personal benefit these systems offer remains difficult. To this end, we propose to assess driving systems by looking at the overall trip cost to incentivize drivers accordingly. For this, we define a novel metric to quantify the total trip cost, combining fuel cost and travel time within a monetary unit. We show the application of our new metric in a case study, comparing human driving, ACC, and platooning. Our results indicate that human driving always loses against ACC and that platooning has a significant advantage in mid to high traffic densities.
The technological maturity of autonomous driving has improved with the development of hardware, software, and element technologies. There is a demand for coping with unexpected (emergency) situations and diversification of driving environments for safer autonomous driving. However, it still recognizes the surrounding environment by relying on sensors mounted on the vehicle. Demand for wireless communication technology is increasing in order to secure the safety of autonomous vehicles. This is because it is necessary to apply V2X technology to overcome the physical limitations of sensors and to provide additional services. In this paper, we presented the implementation results for scenario-based functional evaluation using connected autonomous vehicles with C-V2X technology. A message set for C-V2X-based data transmission was presented, and the received objects data was overlapped with the autonomous vehicle system. According to the scenarios for verification of unit functions, the autonomous vehicle system was tested on the proving ground. As a result of the experiment, it was confirmed that it could be used for scenario-based functional evaluation of the autonomous vehicle. In addition, by using virtual objects data rather than actual vehicle functional evaluation, it can contribute to improving safety.
Since the last one decade, researchers are continuously trying to implement and evaluate the performance of DAVN's (Drone Assisted Vehicular network). DAVN efficiently integrates the networking and communication technologies of drones with connected vehicles (CV). In this poster, we have demonstrated the DAVN communication scenarios; Drone to vehicle and infrastructure communication using road weather data exchange. We have used the real-time road-weather and road-traffic data that we obtained during our pilot measurements in Northern Finland to carry out these pilot scenarios. Later, the executed and generated test scenarios are added to Wireshark and NS-2 (Network Simulator) to evaluate the performance of 5G network and ITS-G5. The performance evaluation for DAVN was performed by considering the following parameters: End-To-End delay, Packet Delivery Ratio (PDR) and Average Throughput.
Large-scale VANET simulations are computationally intensive. Disolv is a simulation architecture proposed to support city-scale VANET studies. This paper describes software decisions taken to realize a concrete implementation of Disolv. We describe a common workflow and provide guidelines on best utilizing Disolv. Finally, a small experiment demonstrates the performance gains by comparing the execution times with state-of-the-art VANET simulators.
New Radio (NR) Vehicle-to-Everything (V2X) technology promises safer and more efficient transportation by enabling direct communication between vehicles. Resource allocation in NR V2X Sidelink (SL) mode 2(a) faces challenges like the hidden node problem, prompting the proposal of cooperative resource scheduling, such as in mode 2(b). However, mode 2(b) definition is still in its infancy and lacks performance evaluation. In this paper, we present a cooperative resource scheduling scheme inspired by mode 2(b) to address these challenges. Through simulation-based evaluation, we demonstrate its potential, especially in virtual train coupling scenarios, offering insights to improve NR V2X communication.
Reconfigurable intelligent surface (RIS)-assisted non-orthogonal multiple access (NOMA) system is a key technology for next-generation wireless communication systems to enhance coverage, optimize spectrum utilization, and increase throughput. In this poster, we examine Infrastructure-to-Vehicle (I2V) communication through RIS, employing a hybrid orthogonal multiple access (OMA)-NOMA system with imperfect phase compensation. We consider lower and upper bounds on the power allocation factor and formulate an optimization problem to achieve (\alpha)-fairness among the vehicles. Additionally, we propose a vehicle pairing and power allocation strategy that ensures the achievable data rate is greater than its OMA rate. Through extensive simulations, we compare and show that the proposed algorithm outperforms state-of-the-art algorithms.
Reducing traffic accidents and congestion remains a significant challenge, especially at intersections crucial for safety and efficiency. Virtual Platooning has been proposed to improve traffic efficiency. However, existing research has focused on Connected Autonomous Vehicles (CAVs), without taking into account human-driven vehicles, which lead to issues such as delay in reaction to stimulus onset and difficulty in controlling acceleration. In this paper, we propose method accounts for Connected human-driven vehicles (CHVs) and address these challenges through control interval adjustments, a target speed determination, and larger minimum following distances. Simulation using a traffic simulator confirms that our approach improves traffic safety and efficiency.
Congestion on expressways occurs mainly due to concentration of traffic. Deceleration control, in which vehicles following the congested vehicle group intentionally slow down, is known to effectively work to mitigate this type of congestion. Previous studies on deceleration control such as variable speed limit (VSL) have been focused on environments with negligible system processing delay, that is wireless communication delay and server processing delay. This delay, however, could pose negative impact on performance in practical use. This paper proposes a novel method of deceleration control in which each vehicle independently determines its target speed by referring to both its current cruise information and surrounding traffic flow information provided by the server with the help of reinforced learning. Performance evaluation results show the proposed method outperforms existing approach in terms of average speed performance when the delay increases up to several seconds.
Connected and Automated Vehicles (CAVs) utilize Vehicle-to-Everything (V2X) communication to exchange messages and share information with surrounding vehicles. Collective Perception Messages (CPMs) contain object information detected by the onboard sensors. Through the exchange with nearby CAVs, CAVs can perceive the surrounding objects. Higher vehicle traffic densities, however, can lead to a shortage of communication resources and degradation of communication quality. This, as a result, leads to a decreased object recognition frequency. To mitigate such degradation, it is necessary to reduce the traffic volume of CPMs. Therefore, it is important to decrease the message size of CPMs. Various existing studies aimed at it have reduced the redundancy of CPMs. Most of the studies, however, do not prioritize the inclusion of high-accident risk objects in CPMs, which leads to a higher Age of Information (AoI) for such objects. Furthermore, in the studies, CPMs are often set to be variable length and aperiodic, which can degrade communication quality in Cellular-V2X (C-V2X). In this paper, we propose a risk and redundancy-based object selection method, called RRS, which selects a fixed number of objects for CPMs based on the accident risks and the redundancy of CPMs in C-V2X. In this proposal, we define an accident avoidance acceleration a as an indicator of the accident risks. Based on a and the number of CPM receptions indicating the redundancy, CAVs calculate the priority of objects for CPMs. CAVs select a fixed number of objects based on the priority to generate CPMs with fixed length periodically. We evaluate RRS through simulation experiments in an intersection scenario. The results show that RRS decreases the AoI for objects with high accident risks and the number of the highest risks.
The growth of autonomous driving technology is accelerating. However, complete autonomous driving has not been implemented yet. This paper proposes a multi-camera interoperable emulation framework for developing autonomous vehicle driving. We implement two components of the Advanced Driver Assistance System (ADAS). The vehicle adjusts its speed based on the distance from the object and stays in its lane. Smart Cruise Control (SCC) and Lane Keeping Assist (LKA) are. These two systems are remotely controlled in our framework. As a result, developing, applying, and simulating algorithms will be more convenient, and this can protect drivers from accidents caused by incomplete algorithms during simulations. Moreover, these systems can relieve the drivers' burden and fatigue during real driving and prevent dangerous situations that can occur due to other vehicles or pedestrians.
With the development of autonomous driving, multi-agent pathfinding is necessary. Because a little delay can cause a collision with a median strip or other vehicles, the pathfinding process must be fast and immediate. Many autonomous vehicle companies use the cloud for central processing because of complicated calculations, but using the cloud can cause delays and disconnection. So, edge-level pathfinding must also be executed together. But if every edge-level pathfinding goes greedy, traffic is paralyzed, and the overall travel time will be increased. One of the suitable methods is using the A search algorithm. Compared to the Dijkstra algorithm, A can't always find the shortest path, but it is much faster than Dijkstra. However, there will still be more ways to reduce calculation time, and there is a tendency for the shortest path to be near the wall. This paper lets the algorithm explore more nodes near the wall and get the results.
Edge computing has to potential to deliver vital low-latency services for ITS. However, due to the limited resources available on the edge, carefully allocating such services and adapting to changes in client distribution is necessary. In this poster, we show that a MAT reinforcement learning approach can perform better than heuristics.
In an effort to improve road safety, the Finnish Meteorological Institute (FMI) worked on a project to develop a cutting-edge wireless traffic network connecting vehicles and infrastructure. The FMI team conducted extensive field measurments on Petäjämaa and Sodankylä Airport test tracks by meticulously designing, refining, and piloting the measurements and related techniques. In this paper, our field measurements ranging from simple vehicle-to-vehicle (V2V) interactions to complex multi-hop scenarios involving multiple cars and roadside units using cellular based 5G networks are reported. The paper examines important metrics such as goodput time, latency, packet loss, and average throughput using state-of-the-art equipment such as Sunit vehicle PCs, laptops, modern smart phones and antennas. Despite facing challenges such as GPS inaccuracies and varying vehicle speeds, this study provides valuable insights into the network's performance under different real-time conditions. The results have significant implications for understanding connectivity ranges and achievable data rates in real-world settings using 5G network. The findings from this paper have the potential to significantly impact road safety standards, offering a glimpse into the future of transportation and paving the way for safer roads worldwide.
The basic idea of Heterogeneous Data for Enhanced Traffic Services (HDETS) is to facilitate different traffic services that utilize heterogeneous data to develop road safety services and to improve the resource utilization efficiency i.e., electric energy and Road Weather Services (RWS). The potential for smart traffic services is substantial, as the amount of heterogeneous data generated by real-time traffic environments to exploit situational awareness is much larger than the data generated today. In this paper we have discussed the unique aspect of HDETS by combining state-of-the-art cloud and edge architectures, novel 5G and 6G communication technologies, Artificial Intelligence (AI) methods and deep knowledge from this application area. The first objective of HDETS is to analyze the data sources which together enable recognizing emerging traffic situations and physical conditions affecting traffic. The second objective is to develop AI methods that are required for the service scenarios. The third objective is to develop a combined cloud and edge computing architecture for the AI methods. The fourth objective is to study wireless communication technologies for this architecture including 5G and 6G, short-range Vehicle Area Networking (VANET) and Visible Light Communication (VLC). Dealing with the above-mentioned objectives, this paper paves the way to respond to our research questions on the data sources and AI methods. It enables novel services, as well as the cloud and edge architectures, and communication technologies that fulfill the requirements set by these AI methods.
Conversational assistants based on large language models (LLMs) have spread widely across many domains, and the automotive industry is keen to follow suit.
However, current LLMs lack sufficient understanding of geospatial data; in addition, timely information, such as weather and traffic conditions, is inaccessible to LLMs.
In this demo, we present an in-car assistant capable of verbally communicating with the driver, and by utilizing external APIs, it can answer questions related to routing, finding points of interest, and is aware of the local weather and traffic conditions.
The assistant, including a customizable speech synthesizer, is accessible through a graphical user interface that facilitates experimentation by simulating the change in time, origin, destination, and location of the car.
This paper presents a real-time visualization system for LiDAR data across networks. Evaluation experiments show that the data reduction method of our system reduces the latency.
5G communication networks have shown to be a promising technology for novel application domains in the context of connected mobility, industrial networks, as well as in the aviation domain. To evaluate the impact of the networking characteristics on the application performance, the state-of-the-art currently relies on simulation-based performance evaluation or real-world field tests where many approaches focus only on individual aspects of the 5G network. However, to the best of our knowledge, a system to measure the overall networking characteristics of private 5G campus networks to gather a holistic and reproducible view of the network is still missing. In this paper we introduce a distributed and scalable framework to measure the performance of 5G campus networks in a spatio-temporal domain. A first feasibility study indicates that our system enables practical and easy-to-use field tests for 5G systems as being used in the vehicular networking domain for teleoperated driving.
Modern automotive networks based on Time- Sensitive Networking (TSN) are becoming increasingly complex. While hands-on experience is critical to understanding these concepts, the complexity and cost associated with TSN technologies often make practical training inaccessible. As an alternative, network simulation tools have been widely adopted, but they lack interactivity and immediate feedback. To bridge this gap, we propose an interactive and affordable TSN testbed built using off-the-shelf hardware. Our solution provides a user-friendly interface for configuring the testbed and experiencing real-time interactions, such as assessing the impact of background noise traffic on automotive LiDAR sensor data. We demonstrate the functionality of our testbed and provide open-source access to the source code, aiming to improve the quality of TSN training and live experimentation.
With the continuous development of Autonomous Vehicles (AVs), Intrusion Detection Systems (IDSs) became essential to ensure the security of in-vehicle (IV) networks. In the literature, classic machine learning (ML) metrics used to evaluate AI-based IV-IDSs present significant limitations and fail to assess their robustness fully. To address this, our study proposes a set of cyber resiliency metrics adapted from MITRE's Cyber Resiliency Metrics Catalog, tailored for AI-based IV-IDSs. We introduce specific calculation methods for each metric and validate their effectiveness through a simulated intrusion detection scenario. This approach aims to enhance the evaluation and resilience of IV-IDSs against advanced cyber threats and contribute to safer autonomous transportation.
Multipath is a common phenomenon that influences the accuracy of phase-based measurements. Distance estimation and positioning features are being increasingly integrated in many wireless standards to meet the growing need for indoor localization mechanisms. Previously used methods, such as RSSI ranging are not accurate enough, especially in multipath environments. The Bluetooth SIG is currently working towards enabling high-accuracy distance estimation between Bluetooth devices. Although this approach is more advanced it is still highly affected by multipath fading which impacts accuracy. In this work, a method to detect multipath based on IQ data samples and machine learning is presented. The aim is to improve accuracy, especially for measurements inside a vehicle, for phase- based distance algorithms, or for labeling IQ data samples for supervised and semi-supervised machine learning tasks. The data used for these evaluations takes into account three different environments which have different multipath characteristics.
Hardware-in-the-Loop (HiL) testing is a test methodology that can be used throughout the development of real-time embedded controllers to reduce development time and improve test effectiveness. HiL testing provides a way to simulate sensors, actuators, and mechanical components so that all I/O of the electronic control units under test are connected long before the final system is integrated. With electronic control units becoming increasingly complex, the number of test combinations required to ensure correct functionality and response increases exponentially. The presented demonstration integrates two physical On-Board Units (OBU) into an Artery V2X simulation using the Robot Operating System (ROS). This allows the exchange of ETSI Cooperative Awareness Messages (CAM) and Decentralized Environmental Notification Messages (DENM) via the V2X simulation environment.
Timely and accurate detection of speed bumps poses a significant challenge, particularly in large-scale vehicular networks where maintaining a low false positive detection rate is crucial to minimize manual verification efforts. In this paper, we propose an accurate speed bump detection algorithm with a low false positive rate, which is essential for large-scale crowd-sourcing systems like connected vehicular networks. Our demonstration showcases the algorithm's pipeline, encompassing the data collection and processing of the in-vehicle system, data fusion and store in the network backend, and visualization of speed-bumps. The dataset utilized in this demonstration comprises real-world driving data stored in Robot Operating System (ROS) bag files. Our demonstration illustrates that the proposed speed bump detection and sharing system can reliably detect speed bumps and irregular road surfaces with zero false positive detections across all recorded data.
By bridging the gap between the physical and cyber domains, the Digital Twin (DT) technology will contribute to revolutionize mobility. So far, the research focus has been on pairing connected and automated vehicles with their digital counterparts in order to monitor them in real-time and to predict and optimize their operation. On the contrary, capturing the driver behaviour through the associated DT is still at an early research stage. The benefits of the interplay between the vehicle DT and the driver DT (DDT) to achieve personalized and safe automated driving have to be practically unveiled. In this work, we fill this gap by showcasing the advantages derived from the interaction of the two types of DTs, with the aim of personalizing vision-based models to detect the driver drowsiness. Fine tuning is applied to the Machine Learning (ML) algorithm available on board, based on the data collected by the DDT. This results in an accuracy improvement up to 3%, at the cost of the ML model exchange between DTs.
Modern vehicles can detect events and react with other cars by communicating over vehicular networks. For example, the ego vehicle could detect the unsafe maneuvers of approaching vehicles and respond in coordination with other cars to prevent accidents. We focus on this application in this paper. The ego vehicle detects the unsafe cut-the-corner maneuver of turning vehicles. It forms a Vehicular Micro Cloud (VMC) and communicates with other cars over vehicular networks, and they collaboratively move backward to avoid collisions. We demonstrate the feasibility of the proposed application through driving simulator experiments. The experiments showed that the proposed application could prevent the potential collision caused by unsafe cut-the-corner maneuvers of turning vehicles.
To implement the V2X systems, reliable channel estimation is a major critical challenge due to the rapid time-varying characteristic of the vehicular channels. In this recent result paper, we propose a denoising autoencoder (DAE) based channel estimation scheme. The proposed scheme has a simple neural network structure and significantly improves the channel estimation accuracy by training the autoencoder such that it can remove the noise and distortions generated during the data pilot aided channel estimation process. Simulation results verify that the proposed DAE outperforms the conventional deep learning based channel estimation schemes.
Abstract: In this talk I will focus on conflict analysis in cooperative driving. I will present the concept of conflict charts which enable connected automated vehicles to maintain conflict-free maneuvers while improving their time efficiency. The detrimental effects of time delays arising in wireless communication and in vehicle dynamics will be highlighted. Moreover, to compensate for the negative effects of the delays, I will introduce the concept of intent and demonstrate its benefits using numerical simulations. The presented theoretical results have been validated experimentally at the Mcity test facility using real vehicles equipped with wireless vehicle-to-everything (V2X) communication devices.
In the evolving landscape of vehicular networks, the need for robust, scalable, and decentralized learning mechanisms is paramount. This paper introduces a novel Decentralized Federated Learning (DFL) framework for wireless technology recognition in vehicular networks, essential for intelligently allocating spectrum resources in Multi-Radio Access Technology (Multi-RAT) scenarios. In contrast with centralized learning at the base station level, our approach leverages Roadside Units (RSUs) for model training and aggregation, eliminating central server dependency and enhancing resilience to single points of failure. Each vehicle trains a Convolutional Neural Network (CNN) for wireless technology recognition using the Fourier transform of In-phase and Quadrature (IQ) samples collected from a specific combination of technologies. The proposed framework is comprised of two steps. First, Centralized Federated Learning (CFL) is employed at the RSU level to create an aggregated model, considering the users' connectivity status. Second, DFL is utilized to establish a global model at each RSU by sharing models with neighboring RSUs. This approach not only preserves data privacy and security but also optimizes learning by leveraging local computations and minimizing the need for extensive data transmission. Our experimental analysis validates the viability of this approach in providing a scalable and resilient solution for technology recognition in vehicular networks. Our results indicate that DFL surpasses its centralized counterpart by 30% in sparse deployments with low connectivity rates.
In Vehicle-to-Everything (V2X) applications, vehicles pass through network cells at high speeds, with ultra-dense deployments of base stations, frequent handover becomes a prominent issue to be addressed. Especially in V2X road-safety use cases, maintaining a long-term stable connection to the cell is more important than having high immediate throughput. Therefore, the traditional handover scheme that selects the base station with the highest signal strength without considering the remaining dwell time may lead to even more frequent handovers. In this paper, we propose a new handover scheme that aims to select the target cell with maximal remaining dwell time, such that after handover the user equipment will camp on the new cell without the need to handover again soon. The remaining dwell time for the potential target cells is predicted based on the user equipment's local measurement of received signal strength, driving speed and direction using 1D-CNN deep learning networks. The new handover scheme does not require any private user information to be exchanged with the BS. Validation results show that our 1D-CNN-based deep learning network provides high-accuracy predictions of remaining dwell time achieving an R2 score of 0.96. Network level simulation results demonstrate that the deep learning network based handover can reduce 74% of unnecessary handovers in ultra-dense scenarios.
Vehicle-to-everything represents a major step in the evolution of Intelligent Transportation Systems (ITS), by allowing connected and automated vehicles, and the infrastructure, to share information seamlessly. Vehicular use cases with reduced latency in data processing and transmission require the utilisation of Vehicular Edge Computing (VEC). VEC needs to integrate ITS services comprising communication architectures and message formats that are accessible and efficient, and ensure timing constraints. For this purpose, this paper discusses some of the issues present in current vehicular protocol stack, and proposes Vanetza-NAP, a microservice architecture for ITS-G5 communications. Vanetza-NAP adds new features to Vanetza and support for: (1) new ITS message types, (2) runtime configuration mechanisms that facilitate orchestration, (3) parallelised design to handle multiple messages simultaneously, and (4) integrated messaging technologies to connect with applications (MQTT and DDS). This approach is evaluated to compare the performance of the new mechanisms and processing features, the delay for the different messaging technologies, the impact of the parallelised design, and the overhead of the microservice-oriented approach and new message types. The results demonstrate the benefit of the parallelised implementation and priority-based message queueing, and that the enhanced interoperability and extended capabilities justifies the small delay increase to support microservice-based edge architectures.
Future cooperative mobile systems all commonly share three characteristics: High reliance on stable and performant network connectivity, high mobility of involved nodes, and operation in both cities and remote/rural areas, where uninterrupted availability of the required infrastructure cannot be guaranteed. Researchers and industry alike are thus looking towards Vehicle-to-Satellite (V2S) communication with Low Earth Orbit (LEO) satellites to bridge connectivity gaps. Yet, the interplay of many parameters impacting system performance is frequently overlooked. In this paper, we present an extensive simulation study investigating the impact of all of: ground station location on Earth, small-scale ground station position in the overall city layout (regarding both neighboring building locations and heights), and properties of LEO satellite constellations (in terms of both density and inclination). We found that each of these many parameters substantially impacts the performance of V2S communication. At the same time we could confirm that street and building geometry in the overall city layout give rise to systematic patterns in V2S connectivity on the ground.
Vehicle-to-Everything (V2X) technologies are seeing dramatic growth as smart cities join the Internet of Things (IoT). Overseeing the adoption of this technology, governing bodies such as State Departments of Transportation (DOTs) are starting to contemplate Vehicle-to-Infrastructure (V2I) deployments, and they need to know which V2I devices will perform best in their specific jurisdictions. In this paper, we develop several methods to test commercial, off-the-shelf (COTS) Cellular-V2X (C-V2X) roadside units (RSUs) as a way to inform government deployment. These methods compare performance between different COTS RSUs side-by-side. We present these methods all together as an open source framework which evaluates an RSU multilaterally: at specification compliance, network management, and transmission range layers.
Abstract: Today's connected vehicles collect massive amounts of data from sensors, vehicle control systems and driver behavior; Advanced Driver Assistance Systems (ADAS), and automated vehicles (AV) generate even more data. In addition, customers have come to expect rich infotainment, convenience, and navigation experiences, which require data from outside sources. The quantity of data transfer needed could be on the order of 100 petabytes per month; connected vehicles typically depend on cellular connectivity for these telematics and infotainment functions, which is costly and can cause serious congestion for cellular networks. Wi-Fi offers a cost-effective solution that could reduce the cost and congestion of cellular connectivity, but generally doesn't have the mobility to provide real-time data exchange. Current Wi-Fi standards allow non-real time offload of vehicle data when in range of an access point but safety communications such as V2X will still require cellular or sidelink communication. Wi-Fi standards could address some of the mobility issues, which may make it more useful for data offload. This panel will address the use cases, technical requirements, standards, and deployment issues related to data offload using Wi-Fi.
Vehicle-to-X communication is increasingly focusing on vulnerable road users. A large group of vulnerable road users are pedelecs. In Vehicle-to-X communication, message generation rules play an important role. These rules can strongly influence the number of messages and the safety. This paper provides an overview of the current Vulnerable Road User Awareness Message generation rules and their impact based on a real pedelec data set. It also evaluates the importance of the rules and the frequency of message generation. All analyses are performed on data sets with different sample rates in order to get an overview of the impact of the sample rate. The results of this paper show that different sample rates lead to significant differences in the impact of the generation rules and also on the message generation frequency. Also a suggestion of a minimum sample rate for recording the sensor data of pedelecs is done. The importance of the different generation rules is measured with the help of an information loss metric. The loss of information caused by ignoring a rule varies significantly among the different rules. This loss also depends on the sample rate. Overall, this paper shows the current state of the generation rules, the different importance of each rule and a suggestion for a minimal sample rate.
According to the World Health Organization, the involvement of Vulnerable Road Users (VRUs) in traffic accidents remains a significant concern, with VRUs accounting for over half of traffic fatalities. The increase of automation and connectivity levels of vehicles has still an uncertain impact on VRU safety. By deploying the Collective Perception Service (CPS), vehicles can include information about VRUs in vehicle-to-Everything (V2X) messages, thus raising the general perception of the environment. Although an increased awareness is considered positive, one could argue that the awareness ratio, the metric used to measure perception, is only implicitly connected to the VRUs' safety. This paper introduces a tailored metric, the Risk Factor (RF), to measure the risk level for the interactions between Connected Automated Vehicles (CAVs) and VRUs. By evaluating the RF, we assess the impact of V2X communication on VRU risk mitigation. Our results show that high V2X penetration rates can reduce mean risk, quantified by our proposed metric, by up to 44%. Although the median risk value shows a significant decrease, suggesting a reduction in overall risk, the distribution of risk values reveals that CPS's mitigation effectiveness is overestimated, which is indicated by the divergence between RF and awareness ratio. Additionally, by analyzing a real-world traffic dataset, we pinpoint high-risk locations within a scenario, identifying areas near intersections and behind parked cars as especially dangerous. Our methodology can be ported and applied to other scenarios in order to identify high-risk areas. We value the proposed RF as an insightful metric for quantifying VRU safety in a highly automated and connected environment.
Electric bikes (e-bikes) are one of the major development assets towards green and inclusive urban transportation. Their high speed, up to 25 km/h, comparable to that of small motorcycles and pedelecs, make e-bikes more prone to traffic accidents than conventional bicycles. Hence, it is crucial to enhance the safety of these Vulnerable Road Users (VRUs) through Vehicle-to- Everything (V2X) communications, providing connectivity among road entities. The European Telecommunications Standards Institute (ETSI) has specified awareness messages, carrying position and kinematics information, to extend the horizon of vehicles and VRUs. While the generation patterns and relevant performance of such messages have been extensively investigated in the literature when considering vehicles, a few works have analyzed the so- called VRU Awareness Messages (VAMs) transmitted by VRUs, and to the best of our knowledge, none when considering e-bikes. In this work, we focus on e-bikes and, collecting mobility traces of participants with different riding skills in the city of Karlsruhe, Germany, we assess the dynamics of VAMs generated according to ETSI rules. The built prototype collects mobility data from the e-bike as well as other context information, thus providing helpful insights about VAM generation patterns and their impact on the network load.
According to the latest statistics, a consistent share of road accidents involves Vulnerable Road Users (VRUs). This calls for the adoption of urgent measures to guarantee their safety while sharing the road with other vehicles. Thanks to the technological advantages of wireless networks, Vehicle-to-Everything communication (V2X) have been proposed as a solution to improve not only the safety, but also the environmental impact and effectiveness of road transport systems. To this aim, the European Telecommunication Standards Institute (ETSI) has defined a set of standards dedicated to VRUs, with the definition of the so-called VRU Basic Service (VBS) defining a dedicated V2X message (VAM, VRU Awareness Message) and several rules for the transmission of such messages. This work proposes an open implementation of the VBS, aimed at evaluating V2X scenarios with connected VRUs. Our implementation targets both simulations and field tests, and has been used to extensively assess the effectiveness and the critical points of the VBS. Our work also explores advanced features like proximity triggering and redundancy mitigation, under different simulated and real-world scenarios, thanks to a setup with a connected vehicle and an equipped stroller.
Ensuring the safety of Vulnerable Road Users (VRUs) is a critical concern in transportation, demanding significant attention from researchers and engineers. Recent advancements in Vehicle-to-Everything (V2X) technology offer promising solutions to enhance VRU safety. Notably, VRUs often travel in groups, exhibiting similar movement patterns that facilitate the formation of clusters. The standardized Collective Perception Message (CPM) and VRU Awareness Message in ETSI's Release 2 consider this clustering behavior, allowing for the description of VRU clusters. Given the constraints of narrow channel bandwidth, the selection of an appropriate geometric shape for representing a VRU cluster becomes crucial for efficient data transmission. In our study we conduct a comprehensive evaluation of different geometric shapes used to describe VRU clusters. We introduce two metrics: Cluster Accuracy (CA) and Comprehensive Area Density Information (CADI), to assess the precision and efficiency of each shape. Beyond comparing predefined shapes, we propose an adaptive algorithm that selects the preferred shape for cluster description, prioritizing accuracy while maintaining a high level of efficiency. The study culminates by demonstrating the benefits of clustering on data transmission rates. We simulate VRU movement using real-world data and the transmission of CPMs by a roadside unit. The results reveal that broadcasting cluster information, as opposed to individual object data, can reduce the data transmission volume by two-thirds on average. This finding underscores the potential of clustering in V2X communications to enhance VRU safety while optimizing network resources.