In machine learning (ML), federated learning (FL) has emerged as a solution to privacy concerns by enabling collaborative model training without sharing local data. This study explores the application of FL in the Internet of Things (IoT) and addresses challenges in data distribution and model selection across heterogeneous environments. We propose an FL framework that integrates ensemble learning with stacking and knowledge distillation, and we conduct experiments to evaluate its performance. Our approach involves a stacking ensemble of weak learners on IoT devices, aggregating results with a convolutional neural network (CNN) model on the server, and applying knowledge distillation to improve model performance and defense against malicious attacks. We tested our framework on the N-BaIoT, IoT-23, and ToN-IoT datasets, comparing it with FL algorithms such as FedAvg, FedProx, FedSGD, FedOpt, and FedDyn, as well as the centralized CNN algorithm. The results show competitive accuracy, precision, recall, and F1-score, with notable advantages in model size, inference time, and handling of heterogeneous data among resource-constrained IoT devices. Knowledge distillation further enhances generalization and performance in distributed data scenarios. With heterogeneous IoT devices, our proposed design outperforms the FL algorithms mentioned above. For example, on the ToN-IoT dataset, compared to other FL methods, our method demonstrates that using separate ensemble learning models on IoT devices reduces IoT resource requirements by over 50% while achieving an F1-score that is over 14% higher.

Ultra-Reliable low-latency Communication (URLLC) fundamentally differs from traditional wireless communication and networking theory. Its requirements are contradictory because closed-loop protocols that address reliability violate the low-latency requirements. URLLC remains a daunting challenge for emerging machine-to-machine (M2M) communication. Suppose we optimize for reliability as opposed to only throughput. In that case, proactive open-loop communications can address the ultra-low-latency requirements. Meanwhile, simultaneous association with multiple Access Points (APs) via a virtual cell can address ultra-reliable requirements. In this paper, we expand upon proactive multi-AP communication via low-complexity URLLC for bursty network traffic by analyzing its heterogeneous dynamic nature and impact. We design a strategy-proof proactive cell association reliability mechanism where a Smart Machine (SM) can utilize a Single Frequency Network (SFN) to achieve URLLC communications. Without additional control overhead, our communication system implicitly provides cell association reliability information to each SM via an AP’s cell breathing (cell radius conveys SM association reliability to each SM). Our analysis uses game theory to study the theoretical and fundamental AP-SM pairing behaviors of proactive open-loop explicit feedback-free protocols in a proactive virtual cell network. We confirm the theoretical results via numerical evaluations and present our insights. Our method can achieve reliable SM connectivity that is superior to non-proactive cell association methods.

Vehicle-to-everything (V2X) is a core 5G technology. V2X and its enabler, Device-to-Device (D2D), are essential for the Internet of Things (IoT) and the Internet of Vehicles (IoV). V2X enables vehicles to communicate with other vehicles (V2V), networks (V2N), and infrastructure (V2I). While V2X enables ubiquitous vehicular connectivity, the impact of bursty data on the network’s overall Quality of Service (QoS), such as when a vehicle accident occurs, is often ignored. In this work, we study both 4G and 5G V2X utilizing Evolved Universal Terrestrial Radio Access New Radio (E-UTRA-NR) and propose the use of socially aware 5G NR Dual Connectivity (en-DC1) for traffic differentiation. We also propose localized QoS, wherein high-priority QoS flows traverse 5G road side units (RSUs) and normal-priority QoS flows traverse 4G Base Station (BS). We formulate a max-min fair QoS-aware Non-Orthogonal Multiple Access (NOMA) resource allocation scheme, QoS reclassify. QoS reclassify enables localized QoS and traffic steering to mitigate bursty network traffic’s impact on the network’s overall QoS. We then solve QoS reclassify via Integer Linear Programming (ILP) and derive its approximation. We demonstrate that both optimal and approximation QoS reclassify resource allocation schemes in our socially aware QoS management methodology outperform socially unaware legacy 4G V2X algorithms (no localized QoS support, no traffic steering) and socially aware 5G V2X (no localized QoS support, yet utilizes traffic steering). Our proposed QoS reclassify scheme’s QoS flow end-to-end latency requires only ≈15% of the time legacy 4G V2X requires.

Ultra-Reliable Low Latency Communication (URLLC) remains a daunting challenge toward 6G, of particular interest in use cases involving machine-to-machine communications and Extended Reality (XR) applications. Proactive communication was proposed to effectively establish URLLC for minimal latency. In this paper, we further innovate opportunistic multi-access of proactive virtual cell communications for smart machines. We employ game-theoretic analysis to explore the fundamental behavior of such a decentralized mechanism. Instead of optimizing throughput in traditional wireless access, maximizing ultra-reliability by a strategy-proof proactive mechanism of multi-cell networking architecture disruptively achieves minimal latency without the need for any additional control overhead, whose impressive performance is further confirmed by numerical evaluations.

Modern society is increasingly dependent on numerous Internet of Things (IoT) devices to assist in a variety of scenarios, such as smart homes and cities, healthcare systems, and cyber–physical systems. Despite IoT’s increasing popularity, IoT security remains a challenge due to the multitude of attack vectors. Existing cyber-attack defense methods attempt to protect the network from both within and outside the network. Network intrusion detection systems (NIDSs) act as device borders within network security and offer a potential defense methodology. This research analyzes the performance of an Artificial Intelligent Internet of Things (AIoT) lightweight botnet attack detection model by deploying meta-learning ensemble botnet detection models and evaluates the capability of a single-board system in addressing cyber-attack threats. The Aposemat IoT-23[1], UC Irvine KDD99[2], and UNSW TON[3]data sets provide IoT and network traffic network flow captures which are used to evaluate the proposed meta-learning methodologies. Experiments show that deployment of our proposed methodologies on edge devices exhibits similar results to PC-based Desktop CPU-trained models. Over the three data sets, when considering a binary classifier (benign versus malignant), our models can consistently achieve above 97.9% accuracy with a false positive rate (FPR) less than 3.8% and an inference time less than 3.95 s. In this work, we show that for binary classification our meta-learners provide consistently stable high accuracy low FPR performance across all three data sets, while maintaining reasonable inference times.

With the advent of next-generation networks, such as 5G, and their support for spectrum hungry applications, industry has become more ambitious in finding new methods of exploiting existing spectrum. Methods to exploit existing spectrum include network densification and spectrum reuse, both of which attempt to increase spectral efficiency. Reuse of underutilized spectrum offers many opportunities for Service Providers (SPs) and a means of alleviating network congestion. In this paper we present green Non-Orthogonal Multiple Ac-cess (NOMA) Machine-to-Machine (M2M), where underutilized NOMA power-domain resources are re-purposed for non-QoS (no guaranteed Quality of Service) delay tolerant network flows in a non-performance impacting manner. Based on the set of users allocated NOMA resources, small cells determine the minimum Signal-to-Interference-Noise-Ratio (SINR) required, and then al-locate the remaining power for green M2M. Using game theory we design an incentive compatible resource determination and allocation mechanism. In the mechanism, delay tolerant flows compete for resources via a non-QoS Variable Resource Block (VRB) auction. An inherent benefit of this model is that it enables information dissemination using existing standardized methods and does not adversely impact system performance. We show that our auction approximates the profit and al-locative efficiency performance characteristics of the Vickrey Clarke Groves (VCG) auction, yet runs in P-time. In addition, we show that via NOMA and statistical SINR the underutilized spectrum can be estimated and exploited without sacrificing the SP's available spectrum, while also enabling the SP the ability to increase their revenue without impacting the QoS guarantees of primary cellular users.

Ultra-Reliable Low Latency Communications (URLLC), a key area of 5G, is designed to support dependable and time-sensitive communications. Yet when addressing Quality of Service (QoS) requirements, traditional 5G does not consider delay spread an interference metric reported back to the base station (as delay spread approaches the duration of the cyclic prefix, the channel capacity decreases due to inter-symbol interference). As IoT equipment may lack the processing capability to mitigate delay spread’s effects, not accounting for delay spread in reported channel condition metrics can result in resources not being assigned in an efficient manner.For out-of-coverage remote devices, such as in smart factories and Industrial Internet of Things [1] (IIoT), Non-Orthogonal Multiple Access (NOMA) small cells or Layer 3 User Equipment (UE) to Network (UE-to-Network) relays must be relied on to extend the network to localized areas. Due to power domain NOMA’s Resource Block (RB) multiplexing, in a multi-level NOMA system the impact of delay spread on the set of viable resource allocations becomes more evident as the interference generated can render 2nd level RBs unusable. Consequently, the likelihood of achieving URLLC’s low End-to-End Radio Latency requirements becomes more difficult in multi-level NOMA systems.In this paper we investigate a tractable methodology to mitigate the impact of delay spread and improve the End-to-End latency of IIoT devices via 5G NOMA multicast flow numerology migration for URLLC NOMA relays used in a non-public network. We derive a theoretically optimal integer linear programming (ILP) resource allocation algorithm and present a tractable low-complexity approximation algorithm, 5G Multicast Flow Migration over NOMA (5GMFMN). In terms of end-to-end radio latency, our evaluation shows 5GMFMN yields results similar to ILP, and both algorithms can yield over 4× improvement over traditional 5G at a max delay spread of 2.1µs.

Vehicle-to-everything (V2X) is one of the core technologies associated with 5G. It enables vehicles to communicate with other vehicles, networks, and infrastructure. Despite the promise V2X offers, i.e. ubiquitous vehicular connectivity, an often overlooked aspect of V2X is the impact bursty data, such as that encountered when a vehicle accident occurs, has on the network's Quality of Service (QoS). In this work, we study both 4G V2X and 5G V2X utilizing Evolved Universal Terrestrial Radio Access New Radio (E-UTRA-NR) Dual Connectivity (en-DC) Road Side Units (RSUs). By using social groups to determine QoS, versus statically defining QoS, we find that we can leverage 5G RSUs to minimize the impact of localized bursty traffic events on the overall network's QoS, while at the same time enabling regional awareness of localized road traffic conditions via a socially-aware Road Condition Warning Systems (RCWS). We demonstrate that our proposed socially-aware QoS management methodology outperforms legacy socially-unaware 4G V2X and can prove beneficial for 5G V2X and beyond.

Scalable Video Coding (SVC) is an extension of H.264/Advanced Video Coding (AVC), which has good characteristics for video transmission over networks. SVC encodes a video into a base layer (BL) and multiple enhancement layers (ELs). With the degree of importance of each layer, in addition to different modulation and coding schemes (MCSs), we can assign different retry attempt limits to each layer to improve quality of experience (QoE) metrics, such as average playback bitrate and buffering ratio. For example, we can assign a slower MCS and a higher retry limit to the BL, to reduce the loss rate and maintain the playback smoothness; whereas, at the same time, we can also assign faster MCSs and lower retry limits to each of the ELs, to reduce the buffering ratio. In this paper, we present a QoE-based link adaptation (QLA) scheme for H.264/SVC video streaming over IEEE 802.11 b/g wireless LANs. We then present a multicast extension of the QLA scheme, multicast QLA (MQLA). We implemented our QLA and MQLA schemes on a Linux-based Wi-Fi protocol driver, mac80211, and built a testbed to conduct our experiments. Experiment results show that both our schemes exhibit an improved QoE performance over the default link adaptation scheme provided by the Linux wireless driver, minstrel.