Accurate prediction of end-to-end latency is essential for latency-sensitive 5G applications in autonomous driving and railway environments. In dynamic urban environments with non-line-of-sight (NLOS) conditions, handovers, and interference, tail latency events can significantly impact service quality, even with low average delays. This paper investigates whether passive layer-1 physical (PHY) key performance indicators (KPI), augmented by recent round-trip time (RTT) history in hybrid models contain sufficient predictive information to anticipate short-term RTT degradations. Using a large-scale, real-world 5G drive-test dataset comprising over 836,000 timestamped samples, we compare classical time-series methods with variational mode decomposition (VMD)-augmented deep learning models. Models are evaluated as probabilistic predictors of RTT exceeding application-relevant thresholds.

In this paper, we provide a comprehensive evaluation of 5G technology within the railway communication systems which holds immense promise for revolutionizing the efficiency, safety, and connectivity of the rail networks. In this study, we establish a test field to examine the behavior of 5G signals along railway tracks. Our study outlines the technical requirements for the implementation of 5G within future railway mobile communication system (FRMCS) infrastructures, encompassing critical factors such as physical layer (PHY) key performance indicators (KPI), reliability, coverage, and specific usage scenarios. Through testbed outdoor measurements, we compare empirical findings with predefined models to discern the efficacy of 5G in railway environments. This research offers valuable insights into the viability and potential challenges of integrating 5G within the railway industry, paving the way for enhanced operational efficiency and passenger experience in the future of rail transport.

Very few available individual bandwidth reservation schemes provide efficient and cost-effective bandwidth reservation that is required for safety-critical and time-sensitive vehicular networked applications. These schemes allow vehicles to make reservation requests for the required resources. Accordingly, a Mobile Network Operator (MNO) can allocate and guarantee bandwidth resources based on these requests. However, due to uncertainty in future reservation time and bandwidth costs, the design of an optimized reservation strategy is challenging. In this article, we propose a novel multi-objective bandwidth reservation update approach with an optimal strategy based on Double Deep Q-Network (DDQN). The key design objectives are to minimize the reservation cost with multiple MNOs and to ensure reliable resource provisioning in uncertain situations by solving scenarios such as underbooked and overbooked reservations along the driving path. The enhancements and advantages of our proposed strategy have been demonstrated through extensive experimental results when compared to other methods like greedy update or other deep reinforcement learning approaches. Our strategy demonstrates a 40

Ensuring the security and resilience of 5G networks requires comprehensive datasets that capture both control and data plane traffic. However, publicly available datasets remain limited, particularly those covering real-world attack scenarios and resource allocation. To address this gap, we introduce a dataset that includes diverse attack vectors targeting both planes, such as flooding, fuzzing, and Packet Forwarding Control Protocol (PFCP)-based Denial-of-Service (DoS) attacks. The dataset is collected from open-source and commercial testbeds, as well as a MATLAB-based simulation of 5 G resource allocation patterns. We provide statistical and correlation analyses to highlight key attack indicators, demonstrating that features such as src2dst_mean_piat_ms for ICMP floods and RequestMessages for Deregistration floods are highly effective in distinguishing between benign and malicious traffic. Furthermore, SHAP-based feature importance analysis validates the dataset’s applicability for Artificial Intelligence (AI)-driven anomaly detection. By bridging this gap, our dataset enables researchers to advance 5G security mechanisms and optimize resource management strategies.

In the rapidly evolving landscape of Industry 4.0 (I4.0), the convergence of information and operational technologies necessitates real-time communication and collaboration across cyber-physical systems and the Internet of Things (IoT). Rapid data transmission is particularly critical within enterprises (vertically) and among stakeholders (horizontally) in this complex, heterogeneous ecosystem. While current research has focused on data application, processing, and storage within the cloud-edge-device continuum, cross-edge transmission has received less attention, resulting in challenges such as suboptimal routing and excessive delays in horizontal communications. To address the above issues, this paper introduces a Connection-As-Required Scheme (CARS) specifically designed for delay-sensitive IoT and Cyber-Physical System (CPS) applications, where low-latency communication is essential for operational efficiency. CARS leverages Lyapunov optimization and backpressure algorithms to optimize traffic scheduling and routing, minimizing communication delay between entities. Benchmarking against state-of-the-art solutions, CARS reduces Round-Trip Time (RTT) to approximately 47.0% of conventional methods and decreases delay by 24.5% in TCP-based and 26.0% in UDP-based applications. These results highlight the potential of CARS to facilitate effective, low-latency collaboration in diverse I4.0 environments.

Onsite bandwidth reservation requests often face challenges such as price fluctuations and fairness issues due to unpredictable bandwidth availability and stringent latency requirements. Requesting bandwidth in advance can mitigate the impact of these fluctuations and ensure timely access to critical resources. In a multi-Mobile Network Operator (MNO) environment, vehicles need to select cost-effective and reliable resources for their safety-critical applications. This research aims to minimize resource costs by finding the best price among multiple MNOs. It formulates multi-operator scenarios as a Markov Decision Process (MDP), utilizing a Deep Reinforcement Learning (DRL) algorithm, specifically Dueling Deep Q-Learning. For efficient and stable learning, we propose a novel area-wise approach and an adaptive MDP synthetic close to the real environment. The Temporal Fusion Transformer (TFT) is used to handle time-dependent data and model training. Furthermore, the research leverages Amazon spot price data and adopts a multi-phase training approach, involving initial training on synthetic data, followed by real-world data. These phases enable the DRL agent to make informed decisions using insights from historical data and real-time observations. The results show that our model leads to significant cost reductions, up to 40%, compared to scenarios without a policy model in such a complex environment.

Resource trading between vehicles, which involves the buying and selling of computing and bandwidth resources, is a promising approach for cost-effectively provisioning services in safety-critical applications such as autonomous driving. These applications require a guarantee and the timely receipt of resources through efficient advance reservations. However, due to uncertainties in future reservation duration and resource costs, vehicles exhibit two distinct patterns: some may have reserved insufficient resources and need to purchase more (acting as vehicle requesters), while others may have overbooked resources and need to sell (acting as vehicle providers). In this paper, we formulate the resource trading problem from both the requester and provider perspectives and propose a resource trading architecture to optimize bandwidth reservation. It utilizes blockchain smart contracts for secure and efficient resource exchange within a mobile network operator (MNO) environment. Two algorithms are introduced: a provider selection algorithm to enhance system efficiency by selecting cost-effective providers, and a decision-making algorithm to assist providers in choosing between selling overbooked bandwidth or canceling it. Through simulations, the results show that these algorithms lead to significant cost reductions for requesters and profit gains for providers, up to 59% and 19%, respectively, compared to reservation schemes without resource trading in such a dynamic environment.

Edge computing resources are crucial for time-sensitive and safety-critical vehicular (TSSCV) applications, such as autonomous and remote driving. These applications require guaranteed resource availability for real-time communication and computation to ensure safety and efficiency. With the emergence of integrated sensing and communication (ISAC) in 6G communication systems, extensive sensing data will be added to the communication data. To meet these challenges, resource reservation becomes a necessity. Resource reservation approaches can be either network-side or vehicle-side. Most existing approaches are network-side. However, these approaches are insufficient for these applications because they only guarantee resources to individual vehicles with a certain probability. Vehicle-side reservation is challenging because vehicles usually do not have sufficient information to reason about future available resources and their costs. This comprehensive survey aims to define the characteristics of TSSCV applications and to analyze existing reservation approaches and cost-effective schemes, considering different reservation scenarios in single and multiple mobile network operator (MNO) environments. In addition, it explores strategies to revise or update reservations when faced with uncertain reservation times and costs. To this end, we propose to enable resource trading and reservation exchange facilitated by blockchain smart contracts. This approach promises a secure and efficient method for resource management in future smart vehicles.

The proliferation of the Internet of Things (IoT) is occurring swiftly and is all-encompassing. The cyber attack on Dyn in 2016 brought to light the notable susceptibilities of intelligent networks. The issue of security in the realm of the Internet of Things (IoT) has emerged as a significant concern. The security of the Internet of Things (IoT) is compromised by the potential danger posed by exploiting devices connected to the Internet. The susceptibility of Things to botnets poses a significant threat to the entire Internet ecosystem (smart devices). In recent years, there has been a simultaneous evolution in the complexity and variety of security attack vectors. Therefore, it is imperative to analyse IoT methodologies to detect and alleviate emerging security breaches. The present study analyses network datasets, distinguishing between those of the Internet of Things (IoT) and those that do not, and provides a thorough overview of the findings. Our primary focus is on IoT Network Intrusion Detection (NID) studies, wherein we examine the available datasets, tools, and machine learning (ML) techniques employed in the implementation of network intrusion detection (NID). Subsequently, an evaluation, assessment, and summary of the current state-of-the-art research on IoT-related Network Intrusion Detection (NID) conducted between 2018 and 2024 is presented. This includes an analysis of the publication year, dataset, attack types, experiment results, and the advantages, disadvantages, and classifiers employed in the studies. This review emphasises research related to IoT NID that employs Supervised Machine Learning classifiers, owing to the high success rate of such classifiers in security and privacy domains. Furthermore, this survey incorporates a comprehensive analysis of research endeavours on IoT NID. Furthermore, we have identified publicly available IoT datasets that can be utilised for NID experiments, which would benefit academic and industrial research purposes. Moreover, we analyse potential prospects and future advancements. The review’s findings indicate that the Internet of Things (IoT) has been substantiated by its swift proliferation in recent times, leading to even broader network coverage. This study presented conventional datasets gathered over a decade ago and current datasets published within the past decade and utilised in recent research. The survey provides a succinct overview of prevailing research trends in IoT NID for security professionals.

In this paper, we develop a machine learning model to detect active eavesdroppers in a Massive Multiple Input Multiple Output (MIMO) system. Massive MIMO systems are naturally immune to passive eavesdroppers, but this is dramatically degraded by active eavesdroppers. We propose two machine learning-based schemes, i.e. a Support Vector Machine (SVM) based scheme and a Naive-Bayes (NB) based scheme, to classify and detect the presence of an active eavesdropper. Then, we apply a Deep Neural Network (DNN) for detecting the presence of an active eavesdropper. We first build structured datasets based on the Received Signal Strength (RSS) and then apply SVM classifiers, NB classifiers, and DNN to those structured datasets. We built a machine learning model based on a realistic scenario where the Channel State Information (CSI) of the channels (legitimate users and eavesdroppers) is unknown. We exploit the massive MIMO technique features to improve the performance of the detection models. The work presented here provides insights into the design of DNN and new machine learning-based secure transmission schemes in Massive MIMO.

Time-sensitive and safety-critical networked vehicular applications, such as autonomous driving, require deterministic guaranteed resources. This is achieved through advanced individual bandwidth reservations. The efficient timing of a vehicle decision to place a cost-efficient reservation request is crucial, as vehicles typically lack sufficient information about future bandwidth resource availability and costs. Predicting bandwidth costs often using time-series machine learning models like Long Short-Term Memory (LSTM). However, standard LSTM models typically require longer durations of multiple input data sets to achieve high accuracy. In certain scenarios, quick decisions must be made, even if the vehicle means sacrificing some accuracy. We propose a batched LSTM model to assist vehicles in placing bandwidth reservation requests within a limited data for an upcoming driving path. The model divides data during training to enhance computational efficiency and model performance. We validated our model using historical Amazon price data, providing a real-world scenario for experiment. The results demonstrate that the batched LSTM model not only achieves higher accuracy within a short input data duration but also significantly reduces bandwidth costs by up to 27% compared to traditional time-series machine learning models.

Intrusion detection systems (IDSs) that continuously monitor data flow and take swift action when attacks are identified safeguard networks. Conventional IDS exhibit limitations, such as reduced detection rates and increased computational complexity, attributed to the redundancy and substantial correlation of network data. Ensemble learning (EL) is effective for detecting network attacks. Nonetheless, network traffic data and memory space requirements are typically significant. Therefore, deploying the EL approach on Internet-of-Things (IoT) devices with limited memory is challenging. In this paper, we use feature importance (FI), a filter-based feature selection technique for feature dimensionality reduction, to reduce the feature dimensions of an IoT/IIoT network traffic dataset. We also employ lightweight stacking ensemble learning (SEL) to appropriately identify network traffic records and analyse the reduced features after applying FI to the dataset. Extensive experiments use the Edge-IIoTset dataset containing IoT and IIoT network records. We show that FI reduces the storage space needed to store comprehensive network traffic data by 86.9%, leading to a significant decrease in training and testing time. Regarding accuracy, precision, recall, training and test time, our classifier that utilised the eight best dataset features recorded 87.37%, 90.65%, 77.73%, 80.88%, 16.18 s and 0.10 s for its overall performance. Despite the reduced features, our proposed SEL classifier shows insignificant accuracy compromise. Finally, we pioneered the explanation of SEL by using a decision tree to analyse its performance gain against single learners.

LoRa technology, renowned for its low-power, long-range capabilities in IoT applications, faces challenges in real-world scenarios, including fading channels, interference, and environmental obstacles. This paper aims to study the reliability of LoRa in Non-Line-of-Sight (NLoS) conditions and in noisy and mobile environments for Industrial IoT (IIoT) applications. Experimental measurements consider factors like vegetation and infrastructure, introducing mobility to replicate NLoS conditions. Utilizing an open-source LoRa Physical Layer (PHY) Software-Defined Radio (SDR) prototype developed with GNU Radio, we assess communication reliability through metrics such as Block Error Rate (BLER), Signal-to-Noise-Interference-plus-Noise Ratio (SINR), and data rate. The study reveals the estimated overall reliability of the LoRa signal at 90.23%, emphasizing specific configuration details. This work contributes to the broader field of LoRa communication, encompassing hardware, software, protocols, and management, enhancing our understanding of LoRa’s dependability in challenging IIoT environments.

Open Radio Access Network (RAN) introduces a groundbreaking industry standard for Radio Access Networks, fostering vendor interoperability and network flexibility through open interfaces while leveraging network softwarization, Artificial, and Machine Learning Intelligence; however, it also poses significant security challenges due to its unique configuration, prompting stakeholders to cautiously approach its deployment and necessitating thorough analysis and implementation of security measures and standards. This paper systematically examines existing literature and case studies to underscore the indispensable role of Intrusion Detection Systems (IDS) in identifying and mitigating security breaches within Open RAN environments. We elucidate the distinct challenges that Open RAN’s disaggregated architecture introduced and classify them into technical and non-technical threats. Finally, we discussed a series of new advancements gaining momentum in the Open RAN security domain and provided insights for future research directions.

Connected and autonomous vehicles (CAVs) have emerged as a promising paradigm in the automotive industry, revolutionizing transportation systems with their potential to improve safety, efficiency, and overall user experience. The communication aspect of CAVs is vital to enable their advanced capabilities. In this article, we focus on analyzing the current landscape, challenges, and future vision of CAVs from a 5G perspective. Unlike previous theoretical studies, our approach is grounded in empirical evidence obtained through trials conducted over a commercial 5G deployment, considering two representative use cases: "Remote Driving" and "Quality of Service (QoS)-based Level-of-Automation Selection." Based on the quantitative findings, we identify and discuss several open challenges that need to be overcome to seamlessly integrate CAVs into the 5G ecosystem. Furthermore, we outline prospective enablers for solutions and enhancements that can address the identified shortcomings and provide a solid foundation for the future of CAVs in the context of 6G. Overall, this research contributes to the ongoing efforts to optimize communication systems for CAVs, facilitating their safe and efficient integration into the future smart transportation ecosystem.

In this paper we present and evaluate the performance of a routing and link scheduling algorithm for millimeter wave backhaul networks. The proposed algorithm models the access point behavior as being selfish by considering access points always aiming to maximize their individual utility, rather than the global optimization objective. Our system utilizes popular concepts from the economics and fairness literature. Specifically, in order to forward packets between the access points that comprise the backhaul network, the Shapley value method is applied, which is shown to induce solutions with reduced latency. The performance of the proposed algorithm is evaluated in terms of total delay and price of anarchy, which represents the inefficiency of a scheduling policy when users are allowed to adapt their rates in a selfish manner and reach an equilibrium. A relaxed version of the problem is also presented, providing a lower bound on the value of the optimal solution. According to simulation results, the system that employs the proposed algorithm outperforms in terms of delay and price of anarchy a system that considers a First-In-First-Out packet forwarding policy, and a system that employs local search global optimization, under which users aim at optimizing the overall network delay.

Intrusion detection systems (IDSs) that continuously monitor data flow and take swift action when attacks are identified safeguard networks. Conventional IDS exhibit limitations, such as reduced detection rates and increased computational complexity, attributed to the redundancy and substantial correlation of network data. Ensemble learning (EL) is effective for detecting network attacks. Nonetheless, network traffic data and memory space requirements are typically significant. Therefore, deploying the EL approach on Internet-of-Things (IoT) devices with limited memory is challenging. In this paper, we use feature importance (FI), a filter-based feature selection technique for feature dimensionality reduction, to reduce the feature dimensions of an IoT/IIoT network traffic dataset. We also employ lightweight stacking ensemble learning (SEL) to appropriately identify network traffic records and analyse the reduced features after applying FI to the dataset. Extensive experiments use the Edge-IIoTset dataset containing IoT and IIoT network records. We show that FI reduces the storage space needed to store comprehensive network traffic data by 86.9%, leading to a significant decrease in training and testing time. Regarding accuracy, precision, recall, training and test time, our classifier that utilised the eight best dataset features recorded 87.37%, 90.65%, 77.73%, 80.88%, 16.18 s and 0.10 s for its overall performance. Despite the reduced features, our proposed SEL classifier shows insignificant accuracy compromise. Finally, we pioneered the explanation of SEL by using a decision tree to analyse its performance gain against single learners.

In this paper, we provide a theoretical study on the receiver operating characteristic (ROC) of frequency-modulated continuous-wave (FMCW)-based radars, orthogonal frequency division multiplexing (OFDM)-based radars and the superposition of FMCW waveform and OFDM symbols in form of IEEE 802.11bd frame structure. Considering a 2-dimensional discrete Fourier transform for the estimation of range and velocity within the delay-Doppler domain, we derive the binary hypothesis tests for individual detection of real and imaginary components as well as their superposition in form of complex signal power. Evaluating the probability of detection and probability of false-alarm, we provide a performance comparison when the detection is solely based on FMCW preamble, random OFDM symbols, and their hybrid form. Via extensive simulation results, we show the feasibility of joint communication and radar within the currently existing IEEE 802.11bd technology. Our theoretically calculated detection probabilities suggest 5–7 dB performance gain of hybrid radar with respect to the individual FMCW and OFDM radars.

Ultra-reliable low-latency communication (URLLC) has been introduced in the 5th Generation (5G) radios for mission-critical applications that demand strict reliability and latency traffic to guarantee the rapid delivery of short packets (up to 1 ms) with a success probability rate of 99.999%. The challenging reliability and latency requirements of URLLC have significant impact on the air-interface design, especially on the Hybrid Automatic Repeat reQuest (HARQ) mechanism. This study focuses on satisfying link latency requirements by reducing the delay that arises in the presence of the HARQ operation. To this end, we propose a Swift HARQ protocol empowered by machine learning techniques to estimate the decodability of a packet early enough within its maximum number of allowable retransmission attempts. This can allow the transmitter to react faster by dropping the non-decodable packets, or activating the repetition mode where parts of the HARQ feedback can be omitted. As shown through system-level simulations, the proposed model achieves a delay reduction of more than 50% compared to the traditional HARQ, and increases the system throughput by up to 40% when multiple HARQ retransmissions are required.

Massive MIMO is seen as a possible physical layer security technique to meet the sixth-generation (6G) security requirements. Massive multiple-input and multiple-output (MIMO) systems are naturally immune to passive eavesdroppers (Eves), but this is dramatically degraded by active Eves. In this chapter, we describe the use of massive MIMO enhanced through artificial intelligence (AI) to improve security on the physical layer. We describe a number of machine learning-based algorithms and a deep neural network (DNN) model capable of detecting the presence of an active Eav by exploiting the particular properties and features of massive MIMO.We describe a machine learning model and DNN applied to a realistic scenario where the channel state information (CSI) of the channels (i.e., legitimate user and Eves) is unknown.A set of different algorithms showing varying performance are compared. Moreover, the prediction complexity of the different algorithms is discussed. The chapter will explain the design issues for DNN and new machine learning-based secure transmission schemes in massive MIMO-based communication systems. Simulations proof the robustness of machine learning-based algorithms and DNN without need of feedback overhead and when the CSI of all channels is unknown. In this chapter, it will be shown that higher security of communication systems can already be achieved on the physical level.