Abstract: A survey of the literature shows that the number of IoT attacks are gradually growing over the years due to the growing trend of Internet-enabled devices. Software defined networking (SDN) is a promising advanced computer network technology that supports IoT. A network intrusion detection system is an essential component in the SDN-IoT network environment to detect attacks and classify the attacks into their categories. Following, this work proposes a deep-learning-based approach that detects attacks and classifies them into their attack categories. The model extracts the internal feature representations from the gated recurrent unit (GRU) deep learning layers; further, the optimal features were extracted using kernel principal component analysis (kernel-PCA). Next, features were fused together, and attack detection and its classification is done using the fully connected network. The proposed feature fused GRU network has achieved better performance than the GRU model and other well-known classical machine-learning-based models. The proposed method can be used in real time to effectively monitor the network traffic in the SDN-IoT environment to proactively alert about possible attacks and classify them into their attack categories.

Abstract: Internet of Things (IoT) devices will play an important role in emerging applications, since their sensing, actuation, processing, and wireless communication capabilities stimulate data collection, transmission and decision processes of smart applications. However, new challenges arise from the widespread popularity of IoT devices, including the need for processing more complicated data structures and high dimensional data/signals. The unprecedented volume, heterogeneity, and velocity of IoT data calls for a communication paradigm shift from a search for accuracy or fidelity to semantics extraction and goal accomplishment. In this paper, we provide a partial but insightful overview of recent research efforts in this newly formed area of goal-oriented (GO) and semantic communications, focusing on the problem of GO data compression for IoT applications.

Abstract: Internet of Underwater Things (IoUT) have gained rapid momentum over the past decade with applications spanning from environmental monitoring and exploration, defence applications, etc. The traditional IoUT systems use machine learning (ML) approaches which cater the needs of reliability, efficiency and timeliness. However, an extensive review of the various studies conducted highlight the significance of data privacy and security in IoUT frameworks as a predominant factor in achieving desired outcomes in mission critical applications. Federated learning (FL) is a secured, decentralized framework which is a recent development in ML, that can help in fulfilling the challenges faced by conventional ML approaches in IoUT. This article presents an overview of the various applications of FL in IoUT, its challenges, open issues and indicates direction of future research prospects.

Abstract: Urban computing (UC) is an interdisciplinary field that seeks to improve people's lives in urban areas. To achieve this objective, UC collects and analyzes data from several sources. In recent years, the Internet of Drones (IoD) has received significant attention from the academic community and has emerged as a potential data source for UC applications. The goal of this work is to examine how IoD can leverage UC. To the best of our knowledge, this work is the first to connect UC and IoD. Data acquired by IoD can fill gaps in data collected from other sources and provide new data for UC considering the aerial view of drones. Thus, this work first introduces the relationship between the concepts of UC and IoD. Second, it proposes a general framework considering the perspective of IoD for UC. Third, it presents how IoD applications can leverage UC in some categories: public safety and security, environment, traffic improvement, drone-assisted networks, and other potential scenarios. Finally, it discusses some key challenges in this area.

Abstract: The rapid increase in heterogeneous data traffic with the ongoing development of self-organizing and self-sustaining networks exposes the limitations of the fifth generation (5G) system, which was originally aimed at enabling the realization of the Internet of Everything. This study presents flexible design agreements of beyond 5G (B5G) from the current 3GPP study and proposes an intelligent network architecture for the 5G and |B5G paradigm to ensure that the network is self-sustained and self-organized. The key idea is to use machine learning (ML) to dynamically schedule flexible transmission time intervals at the slot level to optimize network performance. This study also provides an overview of the queuing model of the medium access control layer and presents how ML-enabled scheduling plays an important role in reducing queuing latency and providing reliable services of the B5G network.

Abstract: Urban computing (UC) is an interdisciplinary field that seeks to improve people's lives in urban areas. To achieve this objective, UC collects and analyzes data from several sources. In recent years, the Internet of Drones (IoD) has received significant attention from the academic community and has emerged as a potential data source for UC applications. The goal of this work is to examine how IoD can leverage UC. To the best of our knowledge, this work is the first to connect UC and IoD. Data acquired by IoD can fill gaps in data collected from other sources and provide new data for UC considering the aerial view of drones. Thus, this work first introduces the relationship between the concepts of UC and IoD. Second, it proposes a general framework considering the perspective of IoD for UC. Third, it presents how IoD applications can leverage UC in some categories: public safety and security, environment, traffic improvement, drone-assisted networks, and other potential scenarios. Finally, it discusses some key challenges in this area.

Abstract: Wireless Industry 4.0 applications typically have stringent latency and reliability requirements. Even though state-of-the-art Wi-Fi networks can reliably achieve single digit milliseconds latency, new emerging time-critical applications have requirements that current Wi-Fi cannot meet. In this paper, we present next generation Wi-Fi technologies and describe how they can be leveraged to enable three time-critical Industry 4.0 use cases: wireless industrial automation control, remote rendering in extended reality applications and cooperative simultaneous localization and mapping using autonomous mobile robots in a factory plant.

Abstract: The global food supply chain needs to evolve to meet a 50 percent increase in food demand by 2050. While food production grows every year, according to a report by the Food and Agriculture Organization (FAO), around 25 percent of roots and tubers, 20 percent of fruits and vegetables, 8 percent of grains and pulses, and 13 percent of animal products are lost before distribution. The majority of such losses are attributed to inadequate monitoring and poor handling during storage and transport. Moreover, the handling of produce in the supply chain also impacts its nutritional content and shelf life. Such losses, when coupled with the rising frequency of epidemics and climate events, further exacerbate the problem of food security for the global population. In addressing the problem of food loss in the supply chain, the biggest hurdle is the lack of traceability and information. On one hand, precision farming helps improve food production efficiency. On the other hand, useful insights post-harvest are not measured due to cost limitations. The ones that are measured often end up in silos or are lost. To overcome these challenges, we propose a framework that leverages low-power IoT sensing networks, smart edges, and data-driven optimization to re-invent the supply chain. In this work, we derive from lessons learned while working with various agricultural supply chain partners and share insights based on some technology solutions that we have explored. We take a bottom-up approach in analyzing the major challenges faced by today's food supply chain. Starting with individual food pallets, we propose ways to develop an agile and low-cost data pipeline that can sense and track the food as it moves through the global supply chain. Further, we propose a dedicated optimization framework that can leverage cloud analytics to boost sustainability and efficiency in the global food chain to meet the growing demand.

Abstract: A survey of the literature shows that the number of IoT attacks are gradually growing over the years due to the growing trend of Internet-enabled devices. Software defined networking (SDN) is a promising advanced computer network technology that supports IoT. A network intrusion detection system is an essential component in the SDN-IoT network environment to detect attacks and classify the attacks into their categories. Following, this work proposes a deep-learning-based approach that detects attacks and classifies them into their attack categories. The model extracts the internal feature representations from the gated recurrent unit (GRU) deep learning layers; further, the optimal features were extracted using kernel principal component analysis (kernel-PCA). Next, features were fused together, and attack detection and its classification is done using the fully connected network. The proposed feature fused GRU network has achieved better performance than the GRU model and other well-known classical machine-learning-based models. The proposed method can be used in real time to effectively monitor the network traffic in the SDN-IoT environment to proactively alert about possible attacks and classify them into their attack categories.

Abstract: While the onset of the COVID-19 pandemic has increased the popularity of home-based consultations, worries over privacy, high consultations costs, slow response times, and the burden on doctors due to the overwhelming number of COVID-19 cases have made current in-person and online models ineffective. In this study, we present an advanced, privacy-protected, artificial intelligence and blockchain-based consultation framework for minor medical conditions. Patients can post their medical queries anonymously on the blockchain network, which may be answered by any available medical professionals. The queries are sorted into their respective domains using naive Bayes and logistic regression. The consultations provided by medical specialists are evaluated based on their reputation, expertise, detail orientation, and the use of supporting documents, and rewards are given in accordance with the evaluation scheme. This fair and incentivized system provides cheaper and more accessible healthcare to patients, which is the need of the hour.

Abstract: The wide proliferation of the Internet of Things (IoT) and the unimaginable rapid advance of artificial intelligence (AI) jointly facilitate the Internet of Artificially Intelligent Things (A-IoT). Artificially Intelligent things (AI-T) run machine learning (ML) models locally while interacting and exchanging data with other AI-Ts through A-IoT. However, sensitive data may be abused during transmission by malicious or compromised AI-Ts. Federated learning is thereby proposed to achieve secure communication, where AI-Ts maintain the same ML model by exchanging model parameters instead of raw data. However, there are three significant issues for FL being applied in A-IoT. First, the context of model parameters of each AI-T may leak privacy, resulting from inference attacks. Second, falsified-data-based poisoning attacks may lead to a failure of ML model convergence. Third, exchanging model parameters cost more communication resources than expected in this resource-limited scenario. To address these issues, we propose privacy-preserving decentralized FL for secure and efficient communication (FL-SEC) over A-IoT. A novel blockchain structure is devised to achieve decentralized FL, which prevents single point of failure and poisoning attacks and also provides a personalized incentive mechanism. In addition, improved sign gradient descent is used to replace traditional gradient descent, which preserves the privacy of model parameters while significantly reducing the communication resource consumption. Experiments on real-world benchmark datasets show the superior performance of the proposed model.

Abstract: This tutorial discusses the problem of angle of arrival (AoA) estimation from small Internet of Things devices. It reviews the limitations and challenges of existing miniaturization strategies, which involve classical antenna arrays. As larger distance between antennas is typically needed for increased precision, the miniaturization without loss of accuracy is a significant problem. The article demonstrates that the use of electrically small antennas does not directly solve the problem, since smaller antennas may require increased distance to avoid coupling between elements. Therefore, this work proposes a new technique to perform AoA estimation in small platforms: a multiple-input multi-ple-output (MIMO) antenna. MIMO antennas are commonly used to increase the communication throughput; however, few works have studied their usage for AoA estimation in line-of-sight environments. The proposed solution offers performance similar to state-of-the-art arrays, but at a substantially reduced size. A performance study is carried out to confirm and validate the proposed technique's accuracy and miniaturization efficacy. Size reduction up to 75 percent compared to linear arrays is achieved with mean absolute errors smaller than 0.11°. The method also demonstrates a resolution of 5° with peak mean absolute error of 0.3° for two simultaneous impinging signals.

Abstract: Marine farms must employ innovative solutions and new technologies to increase productivity and supply the seafood demand sustainably. In this paper, we propose a novel four-layers architecture for smart marine farms. The proposed architecture relies on recent advancements and proposes the integration of Internet of Things (IoT) and Internet of Underwater Things (loUT), edge and cloud computing, and machine learning (ML) for efficient and reliable integration of underwater sensors, data transfer among the components in a smart aquaculture farm, and real-time data processing and inference for monitoring and control of smart marine farms. The designed architecture will tackle many fundamental challenges aimed at the autonomous, intelligent, and real-time monitoring and control of smart marine farms. For each component, the issues it tackles and the challenges and guidelines to implementing it are presented. Finally, we shed light on open challenges that still prevent innovative features in smart aquaculture.

Abstract: The rapid increase in heterogeneous data traffic with the ongoing development of self-organizing and self-sustaining networks exposes the limitations of the fifth generation (5G) system, which was originally aimed at enabling the realization of the Internet of Everything. This study presents flexible design agreements of beyond 5G (B5G) from the current 3GPP study and proposes an intelligent network architecture for the 5G and |B5G paradigm to ensure that the network is self-sustained and self-organized. The key idea is to use machine learning (ML) to dynamically schedule flexible transmission time intervals at the slot level to optimize network performance. This study also provides an overview of the queuing model of the medium access control layer and presents how ML-enabled scheduling plays an important role in reducing queuing latency and providing reliable services of the B5G network.

Abstract: The Internet of Drones (IoD) is an emerging paradigm generated over the Internet of Things (IoT) framework, where things are replaced with drones. IoD facilitates inter-drone communication and provides a mechanism for automatically controlling the drones from a remote location, even in non-line-of-sight conditions. IoD also consists of the onboard controller for making smart decisions by using artificial intelligence. This article highlights the role of IoD in precision agriculture, which helps in achieving qualitative and quantitative improvement in agricultural products, improving the financial benefits of farmers and exact climate predictions. Next, we describe the different challenges and corresponding solutions that may be encountered while using IoD in precision agriculture. We also depict a taxonomy of these challenges and solutions. Finally, we discuss different research opportunities in IoD-based precision agriculture.

Abstract: The Internet of Things (IoT) can be described as a considerable number of sensors and physical devices connected to different applications, supported with networking technologies to communicate with other devices and the Internet. With the growing number of IoT users, emerging services, and the need for high availability and data exchange, cyberattacks on those applications have increased in recent years. Therefore, securing IoT applications has allured particular consideration from the industry and research fields. This article illustrates and comprehensively analyzes the effectiveness of using FL/TL trending techniques used with different Machine Learning (ML) and Deep Learning (DL) algorithms to drive the Intrusion Detection Systems (IDS) to secure the IoT applications. The Internet of Medical Things (IoMT) is considered in this article as a use case in which we have demonstrated that using federated and transfer learning can improve model performance, increase learning process speed, reduce the amount of data needed to be trained, and preserve the user's data privacy compared with the traditional learning approaches.

Abstract: Space-air-ground-sea integrated networks (SAGSINs) are promising to offer ubiquitous Internet services across the globe while confronting research challenges such as security vulnerabilities, privacy leakage concerns, and difficulty in resource sharing. On one hand, emerging network slicing and network softwarization technologies can fulfill diverse requirements with the provision of various services on top of heterogeneous SAGSIN hardware and software resources. On the other hand, blockchain and smart contracts can compensate for network slicing and softwarization to offer secure and automatic network services. This article presents an investigation on the convergence of blockchains with network slicing and network softwarization technologies for SAGSINs from the perspectives of network management and brokerage services of SAGSINs. In contrast to existing studies, this article is the first to incorporate blockchains into network slicing and network softwarization dedicated for SAGSINs. This article starts with a summary of key characteristics and challenges of SAGSINs. Then a review of network slicing and network softwarization is given in the context of SAGSINs. This article next presents an integrated framework of network slicing, network softwarization, and blockchain for SAGSINs. Moreover, this article outlines a set of open issues and research challenges that would be useful to guide future research in this area.

Abstract: 6G envisions artificial intelligence (AI) powered solutions for enhancing the quality-of-service (QoS) in the network and to ensure optimal utilization of resources. In this work, we propose an architecture based on the combination of unmanned aerial vehicles (UAVs), AI and blockchain for agricultural supply-chain management with the purpose of ensuring traceability, transparency, tracking inventories and contracts. We propose a solution to facilitate on-device AI by generating a roadmap of models with various resource-accuracy trade-offs. A fully convolutional neural network (FCN) model is used for biomass estimation through images captured by the UAV. Instead of a single compressed FCN model for deployment on UAV, we motivate the idea of iterative pruning to provide multiple task-specific models with various complexities and accuracy. To alleviate the impact of flight failure in a 6G enabled dynamic UAV network, the proposed model selection strategy will assist UAVs to update the model based on the runtime resource requirements.

Abstract: In this article, we propose a machine learning (ML)-assisted beam selection framework that leverages the availability of digital twins to reduce beam training overheads and thus facilitate the efficient operation of time-sensitive IoT applications in dynamic industrial environments. Our approach employs a digital twin of the environment to create an accurate map-based channel model and train a beam predictor that narrows the beam search space to a set of candidate configurations. To verify the proposed concept, we perform shooting-and-bouncing ray modeling for a reconstructed 3D model of an industrial vehicle calibrated using the real-world millimeter-wave propagation data collected during a measurement campaign. We confirm that lightweight ML models are capable of predicting the optimal beam configuration while enjoying a considerably smaller size compared to the map-based channel model.

Abstract: Wireless Industry 4.0 applications typically have stringent latency and reliability requirements. Even though state-of-the-art Wi-Fi networks can reliably achieve single digit milliseconds latency, new emerging time-critical applications have requirements that current Wi-Fi cannot meet. In this paper, we present next generation Wi-Fi technologies and describe how they can be leveraged to enable three time-critical Industry 4.0 use cases: wireless industrial automation control, remote rendering in extended reality applications and cooperative simultaneous localization and mapping using autonomous mobile robots in a factory plant.

Abstract: In recent years, federated learning (FL) has emerged as a powerful paradigm for distributed learning thanks to its privacy-preserving capabilities. With the use of FL, a network of edge devices can make intelligent decisions without exposing their data to others. Despite its success, the traditional FL is not well suited to many practical applications such as those that involve the internet-of-things (IoT) or cyber-physical systems (CPS), where data access can be intermittent, and edge devices are semi-independent with device-specific dynamic behavior characteristics. Those devices are referred to here as semi-independent devices since they need to make decisions based on their own data and device characteristics, often independent of other devices and the information obtained from other devices in the network. Additionally, as new information becomes available, traditional FL must repeat the entire learning process and may not be able to provide timely and tailored solutions to participants. Personalized online FL, on the other hand, retains the collaborative and privacy-preserving aspects while learning in real time from intermittent data. It further enables devices to learn models customized to the device and the specific tasks it performs. In light of these reasons, personalized Online-FL is ideal for applications where the learning relies on heterogeneous data streams, and local optimization is beneficial. In this work, we want to bring attention to this new learning paradigm, present a few of the applications that could benefit from it, and highlight the principal challenges the research community faces in developing successful personalized Online-FL.

Abstract: Mobile multi-robot systems are an integral component of highly automated factories of the future. Since mobile robots have limited on-board computing capability and battery capacity, there is increasing interest in exploring approaches that enable robots to effectively leverage wireless communications and Edge Computing solutions for perception, navigation, planning, coordination, and control. It is, however, a major challenge achieving precision, high-speed, co-ordinated actions between robots due to tight end-to-end latency, and safety requirements, especially while enabling time-sensitive data exchange over wireless networks and execution of computing workloads distributed across robots and the Edge system. The traditional approach of designing compute, communications, and control components in an Edge system as independent components, limits the capacity and scalability of computing and wireless resources and is therefore unsuitable to meet performance guarantees for energy and resource-efficient time-sensitive robotic applications. In this article, we discuss technical challenges in the context of two Edge Robotics use cases such as conveyer object pick-up and robot navigation, which are representative of time-critical control in IoT applications. We propose research directions grounded in an end-to-end system co-design paradigm and describe technology components such as virtualized robot functions, compute-communications-control co-design, Edge system co-simulation, safety and security aspects that are core to Edge Robotics. We also briefly outline future research directions that are necessary to pave the path toward factory-scale Edge Robotics systems.

Abstract: Next-generation communication systems, including the sixth generation (6G) cellular systems, are expected to support a wide range of new ultra-reliable low-latency communications (URLLC). These emerging applications strictly demand high data rates and/or massive connectivity besides their strict reliability and latency requirements. To enable such URLLC applications alongside the enhanced mobile broadband (eMBB) and massive machine-type communication (mMTC), it is imperative to develop spectrally efficient methods, and non-conventional technologies and networking architectures. To this end, reconfigurable intelligent surfaces (RISs) have recently emerged as a key promising technology to enhance the capabilities of 6G wireless networks. The RIS technology offers a wide range of advantages, including offering a low-cost and energy-efficient solution for controlling the propagation environments. We study the potential of deploying RIS in 6G networks with an emphasis on accommodating the coexistence of URLLC and eMBB traffic. Specifically, we demonstrate through several case studies the potential of integrating RIS into 6G networks with the objective of meeting the requirements of URLLC without compromising the performance of eMBB users. Motivated by this and the fact that the RIS technology has not been fully investigated, we propose a number of relevant research directions and highlight what kind of challenges such directions may entail. We also elaborate on how such challenges, once tackled, may lead to realizing the full potential of integrating RIS into future wireless networks.

Abstract: While the onset of the COVID-19 pandemic has increased the popularity of home-based consultations, worries over privacy, high consultations costs, slow response times, and the burden on doctors due to the overwhelming number of COVID-19 cases have made current in-person and online models ineffective. In this study, we present an advanced, privacy-protected, artificial intelligence and blockchain-based consultation framework for minor medical conditions. Patients can post their medical queries anonymously on the blockchain network, which may be answered by any available medical professionals. The queries are sorted into their respective domains using naive Bayes and logistic regression. The consultations provided by medical specialists are evaluated based on their reputation, expertise, detail orientation, and the use of supporting documents, and rewards are given in accordance with the evaluation scheme. This fair and incentivized system provides cheaper and more accessible healthcare to patients, which is the need of the hour.

Abstract: In this article, we propose a machine learning (ML)-assisted beam selection framework that leverages the availability of digital twins to reduce beam training overheads and thus facilitate the efficient operation of time-sensitive IoT applications in dynamic industrial environments. Our approach employs a digital twin of the environment to create an accurate map-based channel model and train a beam predictor that narrows the beam search space to a set of candidate configurations. To verify the proposed concept, we perform shooting-and-bouncing ray modeling for a reconstructed 3D model of an industrial vehicle calibrated using the real-world millimeter-wave propagation data collected during a measurement campaign. We confirm that lightweight ML models are capable of predicting the optimal beam configuration while enjoying a considerably smaller size compared to the map-based channel model.

Abstract: As Artificial Intelligence of Things (AIoT) has become increasingly important for modern AI applications, federated learning (FL) is envisioned to be the enabling technology for AIoT, especially for large-scale, data privacy-preserving scenarios. However, most existing FL is managed in a centralized manner (CFL), which confronts the limitations of scalability given the AioT device explosion. The key challenge faced by CFL is the communication bottleneck at the central model aggregation server, which leads to a high server-to-worker communication delay and thus severely slows down the model convergence. To address this challenge, this article introduces a generic decentralized FL (DFL) framework that can operate in either synchronous (Sync-DFL) mode or asynchronous (Async-DFL) mode to alleviate the high communication congestion around the central server. Moreover, Async-DFL is the first DFL in the literature to provide a generic FL framework that is fully asynchronous and able to completely avoid worker waiting, which leads to robust distributed model training in the inherently heterogeneous IoT environments, where stragglers (i.e., slow devices) are very common due to the largely varying computing/networking speeds of IoT devices. Our DFL framework is implemented, deployed, and experimented with in both simulation and physical testbeds. The results show that Async-DFL can accelerate the convergence speed of model training twice as fast as CFL, while maintaining convergence accuracy and effectively combating the impact of the stragglers.

Abstract: Space-air-ground-sea integrated networks (SAGSINs) are promising to offer ubiquitous Internet services across the globe while confronting research challenges such as security vulnerabilities, privacy leakage concerns, and difficulty in resource sharing. On one hand, emerging network slicing and network softwarization technologies can fulfill diverse requirements with the provision of various services on top of heterogeneous SAGSIN hardware and software resources. On the other hand, blockchain and smart contracts can compensate for network slicing and softwarization to offer secure and automatic network services. This article presents an investigation on the convergence of blockchains with network slicing and network softwarization technologies for SAGSINs from the perspectives of network management and brokerage services of SAGSINs. In contrast to existing studies, this article is the first to incorporate blockchains into network slicing and network softwarization dedicated for SAGSINs. This article starts with a summary of key characteristics and challenges of SAGSINs. Then a review of network slicing and network softwarization is given in the context of SAGSINs. This article next presents an integrated framework of network slicing, network softwarization, and blockchain for SAGSINs. Moreover, this article outlines a set of open issues and research challenges that would be useful to guide future research in this area.

Abstract: The expansion of Internet-of-Things (IoT) devices with a wealth of generated data opens up new possibilities for intelligent IoT applications (i.e., smart home and smart transportation), but the increasing concern about data privacy makes the enabling force of intelligent IoT, machine learning (ML), harder to deploy. Federated learning (FL), an emerging distributed ML paradigm that allows on-device ML model training without sharing private raw data, is becoming a promising solution to achieve collaborative intelligence in IoT. However, the privacy-preserving design of FL makes it vulnerable to Byzantine workers who behave arbitrarily and send poisonous model updates to the central server to corrupt the joint learning process. Moreover, traditional FL suffers from massive communications overhead due to the large number of training rounds to convergence with non-independent identically distributed (non-i.i.d.) data across clients. To address these problems, we propose a novel robust aggregation rule for FL, federated mutual information (FedMI) which leverages the mutual information between clients to build resilience to Byzantine workers and accelerate the convergence speed. We perform experiments over the non-i.i.d FEMNIST dataset under different adversarial settings. Experimental results demonstrate the effectiveness of the proposed FedMI: much faster convergence speed and higher defense capability when compared to the state-of-the-art robust aggregation rules for FL.