Connected Automated Vehicles (CAVs) utilize their onboard sensors to perceive the environment. The perception range and accuracy can be affected by adverse weather or non-line-of-sight conditions. Cooperative perception or sensor sharing can overcome these limitations by enabling CAVs to exchange sensor data, thus collectively enhancing their perception capabilities. Previous studies have shown the potential of cooperative perception, but limited attention has been given to the fusion of V2X data received through cooperative perception messages with onboard sensor information. The fusion process can be influenced by the quantity and quality of the V2X data. An increased volume of V2X data can reduce uncertainty in the perceived environment; however, when the data is noisy, it may compromise the accuracy of the fusion results. This study investigates the fusion of onboard sensor and V2X data in cooperative perception, and demonstrates that while perception can significantly improve as the V2X penetration rate increases, it can introduce a significant number of false positives if V2X data is not highly accurate. False positives result in the detection of ghost objects that do not actually exist. These ghost objects can, in turn, compromise safety and driving efficiency. Our analysis found that false positives or ghost objects can appear even with accurate V2X data. These findings highlight the challenges in cooperative perception and the importance of developing robust data fusion methods to enhance the reliability of cooperative perception. This is particularly relevant in light of ongoing standardization efforts, such as ETSI TS 103 324 on collective perception.
In the context of the Internet of Things (IoT), an appropriate communication mechanism is crucial to preventing attacks, particularly in IoT applications. The inherent constraints of these devices, such as limited duty cycles, reduced energy capacity, or the inability to listen before transmitting, increase their susceptibility to interference. Low-power wide-area Networks (LPWANs), particularly LoRa, are increasingly used in IoT due to their long-range and low-power capabilities. However, their decentralized structure and the well-known limitations of IoT devices make them vulnerable to attacks, especially jamming. To address this, we propose a ‘mitigation mechanism’ to optimize network performance in LoRa communications. We evaluated the mechanism’s resilience against reactive jamming attacks in simulated environments, showing a clear improvement in network performance over conventional LoRa schemes.
The design of positioning solutions for the Industrial Internet of Things (IIoT) faces several open challenges, particularly in complex indoor environments typical of industrial settings: first and foremost, the limited availability of experimental datasets from realistic scenarios, in particular in relation to private 5G networks.This paper investigates range-free positioning for IIoT in indoor environments using a combination of public and private 5G networks operating at both midband (3.7 GHz) and mmWave (26 GHz) frequencies. The study introduces one of the first publicly available datasets collected in a realistic industrial setting, utilizing synchronization signals from 5G networks. A Weighted k-Nearest Neighbors (WkNN) algorithm is used to perform fingerprinting-based positioning using the experimental data, comparing the achievable accuracy in public vs. private deployments. The results show that private mmWave deployments significantly outperform public networks in terms of positioning accuracy, achieving errors as low as 2.16 meters. Additionally, the study highlights the discriminative power of time-based features such as Time of Arrival (ToA) in enhancing range-free positioning accuracy.
The Service-Based Architecture (SBA) has been successfully employed in the 5G Core since Release 15. Its modular and loosely coupled approach to network function interaction, based on the producer/consumer paradigm, perfectly pairs with the softwarization trend mobile networks have undergone over the last decade. Still, the Radio Access Network (RAN) did not leverage this approach, retaining the traditional Protocol-Based Architecture (PBA). In this paper, we make the case for adopting the SBA in the RAN, a transition we argue is both advantageous and feasible. We first present a comprehensive analysis of the benefits, such as enhanced flexibility and innovation, while also considering potential implementation challenges. Then, we detail how the SBA can be used to provide two exemplary procedures currently performed using the PBA, showing the feasibility of a fully service-based 5G network architecture.
The expansion of 5G networks has led to remarkable data volume and complexity, introducing significant security challenges that require the implementation of robust and scalable anomaly detection mechanisms. Traditional centralized approaches pose privacy risks and scalability challenges due to the distributed nature of 5G infrastructures. Federated Learning (FL) offers a decentralized solution but often overlooks the importance of feature relevance and privacy preservation during model aggregation. This paper introduces a novel Feature-Aware Federated framework that integrates feature importance into the aggregation process while ensuring differential privacy. We employ integrated gradients to compute the importance of the features for each client, aggregate them globally with differential privacy noise, and use these insights to weight model parameters during aggregation. Additionally, we propose Dynamic Feature Importance Adaptation (DFIA) to update feature importance occasionally, enhancing the model’s adaptability to evolving data distributions. Experimental results demonstrate that our framework outperforms traditional federated approaches such as FedAvg, and FedProx in unsupervised anomaly detection tasks within 5G networks, achieving higher accuracy and robustness while preserving data privacy.
This paper presents a nationwide academic testbed that integrates private 5G mobile networks, known as "local 5G" in Japan, with SINET6, the country’s ultra-high-speed and low-latency academic backbone network. Unlike public 5G networks, local 5G systems offer advanced features such as flexible uplink/downlink resource allocation and independent deployment of core and radio access functions. We developed a novel architecture in which both the 5G core and RAN are integrated into a compact, portable system connected to SINET via campus LANs. Deployed across multiple universities in Japan, this testbed enables high-performance mobile data communication tailored for academic research. We report on its design, implementation, performance evaluations, and a range of practical use case demonstrations conducted by participating institutions.
The rapid growth of the Industrial Internet of Things (IIoT) has brought new challenges in protecting industrial networks, especially when Intrusion Detection Systems (IDS) must run on low-resource edge devices. Traditional deep learning-based IDS models often require too much memory and processing power, making them unsuitable for real-time IIoT environments. In this paper, we propose IIoT-TinyDNN, a lightweight deep neural network IDS designed specifically for edge-based IIoT deployment. TinyDNN uses a simplified dense architecture combined with magnitude-based unstructured pruning to significantly reduce model size, memory usage, and computation, while still providing strong detection accuracy. We evaluated TinyDNN on two well-known IIoT security datasets: Edge-IIoTset and TON_IoT. In the TON_IoT dataset, TinyDNN achieved an accuracy of 99.37% using only 1,537 parameters and a fast inference time of 8.24s. Compared to traditional Convolutional Neural Network (CNN) and Deep Neural Network (DNN) models, TinyDNN reduces the number of Floating-Point Operations (FLOPS) by up to 6.8× and achieves over 3× faster inference compared to CNN, while maintaining comparable accuracy. Across various sparsity pruning levels (20%, 40%, 60%), TinyDNN consistently maintains a good balance between speed, accuracy, and memory usage. These results confirm that TinyDNN is a practical and efficient solution for time-sensitive IIoT intrusion detection on resource-constrained edge devices. To support further research, we have made our implementation, datasets, and evaluation tools publicly available.
This paper addresses grant-free, asynchronous control-to-control (C2C) communications over a shared wireless channel. Controllers transmit commands of variable length, encoded using low-density parity-check codes and encapsulated in self-contained packets that include a preamble and a tail sequence. In the absence of a global time reference, each controller operates independently and sends replicas of its packet within a locally defined virtual frame to improve reliability. The receiver has no access to metadata and must detect message boundaries (i.e., preamble and tail sequences) directly from the received signal. We propose a multi-branch, multi-label convolutional neural network that jointly detects preamble and tail sequences, enabling fully data-driven boundary identification. Simulation results show that the proposed receiver enables robust boundary detection under high traffic load, supporting dense and uncoordinated C2C deployments.
The rapid growth of Internet of Things (IoT) devices and applications demands efficient computing closer to the source. Edge computing is crucial for reducing latency, improving bandwidth, and enhancing privacy, addressing the limitations of centralized cloud architectures. However, realizing the full potential of IoT and edge computing requires standardized interworking. Without a common framework, fragmented and proprietary IoT deployments hinder scalability and interoperability. Thus, a standard-based approach is essential for future-proof, vendor-agnostic, and globally interoperable IoT-edge solutions. In this context, the convergence of standardized frameworks for edge computing and IoT is a key enabler for scalable, low-latency, and interoperable systems across domains such as smart cities, industry, and maritime logistics. This paper presents ongoing work within the ESTIMED project that investigates the integration of ETSI Multi-access Edge Computing (MEC) and the oneM2M Partnership Project—two complementary standards for edge computing and IoT service abstraction. By analyzing strategic use cases, the study highlights the technical and operational benefits of this interworking and extracts key insights to identify integration opportunities.
Modern artificial intelligence (AI) techniques, combined with powerful far-edge computing capabilities, enable vehicles to cooperate, generate, and share knowledge with other vehicles over ad-hoc wireless technologies. In such decentralized vehicular knowledge networks, Named Data Networking (NDN) offers a promising paradigm to link knowledge consumers and producers without reliance on centralized infrastructure. One challenge in sharing knowledge over NDN lies in its redundancy-based multi-hop dissemination mechanism, which can introduce inefficiencies and delays. Given the size and time sensitivity of knowledge, this necessitates fast and reliable dissemination strategies. This paper presents Mobility-Aware Knowledge Sharing (MAKS), a scheme that considers vehicular dynamics and trajectory information to develop a mobility-aware forwarding information base (MaFIB), ensuring reliable knowledge sharing and reverse path stability in continuously varying network conditions. Simulation results show that MAKS achieves a knowledge delivery ratio above 90%, reduces the path partition rate by over 40%, and lowers the number of retransmissions by more than threefold compared to other approaches.
Cell-free massive MIMO (CF-mMIMO) networks are considered the main candidate for supporting the next generation of wireless communications, by disrupting the constraining concept of cells and thereby eliminating its main shortcomings, such as cell-edge performance degradation and handover-related intelligence. Arguably, a critical challenge in a CF setup is the Access Point-User Equipment (AP-UE) clustering, as it directly affects both the offered spectral-efficiency (SE) and the energy consumption of the infrastructure. In this work, we introduce a sophisticated clustering problem with a dual objective of optimizing the achieved SE, while sustaining a small energy footprint. The SE objective accounts for user-specific channel conditions, fading characteristics, and inter-user interference, while energy consumption is modeled at both the activation (idle/on) and load-dependent operational levels at the APs. A weight-sensitive genetic algorithm is proposed to swiftly solve the optimization problem. Extensive simulations benchmark our method against established clustering schemes, highlighting our mechanism’s ability to discover a trade-off curve aligned with the dual goals of high throughput and green communications.
Future wireless networks will require architectures that can deliver uniformly high service quality, especially in dense deployments and dynamic environments. Cell-free massive multiple-input multiple-output (MIMO) systems have emerged as a promising solution by distributing many access points (APs) across a wide area to jointly serve users without the constraints of traditional cell boundaries. However, their practical deployment poses challenges in terms of cost, scalability, and coordination. To address these issues, the radio stripes architecture has been proposed, enabling low-complexity and scalable implementations through the serial interconnection of simple APs along a stripe. In this work, we address a key challenge in radio stripes: the design of efficient distributed precoding algorithms that can operate under the constraints of limited processing and communication capabilities. We propose a novel solution by applying the dynamic cooperation cluster (DCC) framework to the local team minimum mean-squared error (LTMMSE) precoder, resulting in a highly efficient and scalable design. Our results show that the DCC-enhanced LTMMSE achieves nearly the same spectral efficiency as the centralized baseline while reducing computation time by almost an order of magnitude. Moreover, it outperforms the only other known DCC-based method, the local partial MMSE (LP-MMSE), in both performance and computational complexity. A detailed computational cost analysis further reveals how the AP-to-UE ratio critically impacts system efficiency. These findings demonstrate the practical potential of DCC-enhanced LTMMSE for enabling real-world radio stripe deployments in future cell-free massive MIMO networks.
We study an information bulletin strategy for decentralized decision making in 6G multi-tenant systems. Queues periodically broadcast descriptor information as two Markov models (queue length inter-change dynamics and service-time distributions) that tenants use to decide whether to renege or jockey. The queues observe tenant responses and adapt their processing rates via a learning loop, with the goal of minimizing the aggregate delay and impatience while respecting service constraints.
In this paper, we address the challenge of energy-aware operation in unmanned aerial vehicle (UAV)-assisted communication systems equipped with active reconfigurable intelligent surface (RIS). We propose a dual-domain control model that jointly optimizes the energy harvesting (EH) duration, user transmit power, RIS-user scheduling, phase shifts, and amplification factors. To eliminate the need for conventional power splitters, the RIS is spatially split such that a subset of elements harvest energy while others forward signals, enabling simultaneous wireless information and power transfer (SWIPT) directly at the RIS level. The harvested energy assists the UAV’s onboard battery in powering the active RIS electronics, thereby enhancing system endurance and ensuring coverage. The joint optimization problem is formulated as a constrained Markov decision process (CMDP) with a hybrid discrete–continuous action space. To solve it, we develop a framework based on the softmax deep double-deterministic policy gradient (SD3) algorithm, incorporating critic ensemble learning and soft Bellman updates for improved training stability. The proposed algorithm enables real-time resource control while satisfying QoS, power, and scheduling constraints. Simulation results show that the proposed scheme consistently outperforms baseline DRL methods such as DDPG, PPO, and TD3. In particular, it achieves up to 65% EH efficiency under active RIS settings, with robust performance across different RIS scales and user densities.
The concept of Network Digital Twin (NDT) has emerged to deal with the increasing complexity of mobile network management and optimization. NDTs replicate the real network state and behavior to act as a sandbox detecting anomalies, optimizing the overall system performance, and automating processes. Data is the fuel of any DT. It notably helps in monitoring and building a future vision of the system. Beyond the volume of data, there is the question of governance, regulation, and ensuring data exchange between various stakeholders. The concept of Data Space (DS) arises as a potential governance solution for NDT data exchange – already used in several verticals, but poorly in networking. In this poster, we aim to foster discussion around DS and the related challenges and opportunities underlying NDT data exchange, in view of future 6G systems.
This poster explores an ongoing design effort to support low-latency, high-reliability communication for connected vehicles using standard 5G core functions. The approach integrates Low Latency, Low Loss, Scalable Throughput (L4S) with 3GPP-defined Network-as-a-Service (NaaS) APIs to allow real-time traffic control. The architecture leverages NEF (Network Exposure Function), PCF (Policy Exposure Function), and UPF (User Plane Function) to manage network behavior based on application-level triggers. Early simulation results suggest that L4S can maintain consistent delay performance for time-sensitive automotive scenarios such as cooperative driving and hazard signaling. The framework is under active evaluation and targets near-term deployment use cases.
Network Exposure Functions (NEF) are the programmable gateway of 5G Standalone (SA) cores, yet adoption is slowed by operational risk, scarce field data, and limited tooling for AI-driven automation. In Luxembourg, a twin-enabled approach is being explored, combining a campus-scale digital replica of a 5G SA network with field trials. By learning from real traffic, the system recommends NEF calls and enforces them autonomously, targeting use-cases like adaptive data rate for live events and ultra-reliable, low-latency communication for autonomous mobility. This poster outlines key challenges, our approach, and discussion points for CSCN.
LoRa technology operating in the 2.4 GHz Industrial, Scientific, and Medical (ISM) band presents an appealing solution for industrial and smart city deployments due to its global availability and compatibility. However, this frequency overlap raises concerns about cross-technology interference (CTI), particularly between LoRa transmissions and ubiquitous WiFi networks. This paper presents a comprehensive empirical study evaluating the interference effects between LoRa and WiFi under real-world conditions. Using off-the-shelf devices, we assess the impact of two WiFi modulations on LoRa packet delivery ratio across different LoRa settings. Our results reveal that while LoRa exhibits robustness against WiFi interference, certain configurations, especially high data rate modes, are vulnerable. Moreover, the results indicate that a LoRa packet can be successfully received if the WiFi received signal power is not higher than 25 dB than the LoRa one, on average. Finally, devoting more bits for error correction improves packet reception up to 23% in the conducted experiments.
This article examines the feasibility of employing converged core-metro-access optical networks to support Radio Access Network (RAN) fronthaul and mid-haul transport under realistic performance constraints. The analysis focuses on the profile of the bit error rate (BER) and the latency evaluation in the context of 5G functional splits. We used a commercially available Nokia ICE-X 400G multi-carrier coherent transceiver, leveraging its digital subcarrier multiplexing (DSCM) capability: Supporting DP-16QAM modulation at 4 GHz channel spacing and 64 GBaud symbol rate per subcarrier. The study is carried out on a representative metro-access topology, where all Radio Unit (RU)- Distributed Unit (DU) - Central Unit (CU) paths are evaluated against BER thresholds and latency bounds for front-haul and mid-haul segments, respectively. The results reveal critical trade-offs between optical reach, modulation format robustness, and latency compliance and demonstrate how mid-haul link distances and routing diversity significantly impact overall transport feasibility for disaggregated RAN deployments. This work emphasizes the potential of transparent optical metro-access infrastructures to serve as an efficient and scalable transport layer for 5G and beyond.
To leverage sub-terahertz (sub-THz) frequency bands for beyond the fifth generation (5G) era, it is essential to understand its transmission characteristics of the current 5G new radio (NR) signals. However, no prior study has conducted transmission performance evaluations of wideband 5G NR signals using actual transmission platform operating in the sub-THz band in indoor multipath-intensive environments with omnidirectional transmission/reception for wireless personal area network (WPAN) usages. To fill this void, this paper first develops a 5G NR signal transmission platform operating in the wideband sub-THz range with a nearly 1 GHz bandwidth for indoor WPAN usage scenarios and evaluates the transmission characteristics in indoor multipath environments. Through preliminary laboratory tests using a WPAN channel simulator, we propose an improvement of the symbol timing synchronization method to avoid underestimation of the transmission performance due to multipath effects. Consequently, we demonstrate an excellent agreement of the block error rate (BLER) performance with a computer simulation with an ideal symbol timing synchronization, even in multipath-intensive channel conditions, confirming that the developed platform allows a fair evaluation without underestimation. Finally, we conduct fundamental field experiments in an actual indoor meeting room environment for the WPAN usage. We demonstrate that short-range wireless transmission of up to 3.6 m and 1.7 m is feasible with carrier data rates of 1.7 Gbit/s and 5.1 Gbit/s, respectively, even with omnidirectional antennas. These results provide valuable reference data for drawing a design guideline of 5G NR-based sub-THz communication systems in WPAN scenarios.
