Federated learning (FL) gained importance in sensitive Internet of Things (IoT) environments by creating a privacy-preserving ecosystem where participants share machine-learning models instead of raw data. However, FL shifts data control away from the server, exposing it to non-independent and identically distributed (non-IID) problems caused by biased clients (IoT devices). This hinders the learning process by increasing execution time and cost. Current solutions alter the FL structure or compromise privacy by offloading clients’ raw data to an external server. To mitigate these limitations, this article proposes a solution to the non-IID problem by introducing an initialization phase, orchestrated by the server, that constructs high-quality initial models. These models can boost FL accuracy and convergence, regardless of whether IoT participants exhibit non-IID properties. Our proposed initialization scheme involves clients training over the same model sequentially, lessening the impact of aggregation, a primary cause of model degradation in federated approaches. Furthermore, a regulator algorithm deployed on the server maintains model integrity and mitigates catastrophic forgetting, enhanced by a client selection process that emphasizes the compatibility of IoT clients to cooperate effectively. Moreover, we devise an optimization scheme based on clustering and genetic algorithms to reduce the selection time while ensuring optimal performance in IoT networks. Experiments on MNIST, KDD, and CIFAR10 data sets show promising results in terms of initial model resiliency against catastrophic forgetting and non-IID settings. Additionally, our findings suggest that our approach can significantly enhance FL training in IoT applications by achieving 40% higher initialization accuracy and a 20% average improvement in end results compared to conventional methods, all while reducing computation time by 80% compared to similar approaches.

In this article, we tackle the network delays in the Internet of Things (IoT) for an enhanced Quality of Service (QoS) through a stable and optimized federated fog computing infrastructure. Network delays contribute to a decline in QoS for IoT applications and may even disrupt time-critical functions. This article addresses the challenge of establishing fog federations, which are designed to enhance QoS. However, instabilities within these federations can lead to the withdrawal of providers, thereby diminishing federation profitability and expected QoS. Additionally, the techniques used to form federations could potentially pose data leakage risks to end-users whose data is involved in the process. In response, we propose a stable and comprehensive federated fog architecture that considers federated network profiling of the environment to enhance the QoS for IoT applications. This article introduces a decentralized evolutionary game theoretic algorithm built on the top of a genetic algorithm mechanism that addresses the fog federation formation issue. Furthermore, we present a decentralized federated learning algorithm that predicts the QoS between fog servers without the need to expose users’ location to external entities. Such a predictor module enhances the decision-making process when allocating resources during the federation formation phases without exposing the data privacy of the users/servers. Notably, our approach demonstrates superior stability and improved QoS when compared to other benchmark approaches.

As the Metaverse develops, it is becoming more crucial to prioritize the safety of users, especially regarding the potential risks, such as users experiencing dizziness or making incorrect movements that may lead to falls. With more virtual environments becoming increasingly available and immersive, detecting and preventing falls within the Metaverse is required. Given the constrained resources of wearable sensors, precise fall prediction models are critical to efficiently analyzing data gathered by these devices. Traditional fall detection systems require centralizing data collection, which raises privacy concerns over the collected data. Resource-aware Split Federated Learning (RSFL) enables collaboration among multiple devices within the Metaverse to train a fall detection model, all while preserving individual data privacy. The approach also leverages parallelism in Federated Learning (FL) and Split Learning (SL) by decomposing training tasks between clients and servers. Moreover, we devise an efficient client selection mechanism to ensure timely training and model convergence performance. We implemented our architecture and assessed its performance using a sensory dataset. The evaluation results with the baseline demonstrate our architecture's superiority in terms of convergence time. Our approach mitigates data heterogeneity and privacy concerns, creating secure and efficient fall detection systems for the Metaverse.

Network delays cause a reduction in the Quality-of-Service (QoS) for Internet of Things (IoT) applications, and even render time-critical applications inoperative. The paper tackles the problem of forming fog federations that aim to improve the QoS. However, instabilities within fog federations might cause some providers to withdraw from the federation, and thus decrease the profit of the federations and the expected QoS. Moreover, federation formation techniques could potentially create privacy risks for end-users whose data is utilized in the process. This paper introduces a decentralized evolutionary game theoretic algorithm that tackles the problem of fog federation formation, as well as, providing a decentralized privacy-aware federated learning algorithm that predicts the QoS between fog servers for optimizing the formation procedure. The devised method provides better stability and increased QoS when compared to other benchmarks.

Federated fog computing is an answer for horizontally upscaling fog resources to improve the Quality of Service (QoS) of Internet of Things (IoT) applications. However, the dynamic nature of some IoT’s crucial components, such as the ones of Internet of Vehicles (IoV), may hinder the QoS improvement and result in its deterioration instead. Specifically, delays can occur due to the unoptimized distribution of services and unbalanced network traffic loads on the fog nodes. The current federated fog architectures ignore the mobility of users during the formation of fog federations. In this work, we present an adaptive and efficient fog federation formation scheme using game theory according to the service requirements. The problem formulation in terms of forming the federations and offloading requests among fog members is formulated as an integer program, then modeled as a Hedonic game. We adopt the Merge & Split as a formation technique, where the federations that are not satisfied in terms of QoS merge with other federations that would enhance the service performance. Our adaptive fog federation formation mechanism is designed to cope with the environmental changes in the IoV paradigm. Experimental evaluation shows that our framework can acquire better QoS and lower time to form the federations compared to the literature.

Fog computing empowers the internet of vehicles (IoV) paradigm by offering computational resources near the end users. In this dynamic paradigm, users tend to move in and out of the range of fog nodes which has implications for the quality of service of the vehicular applications. To cope with these limitations, scholars addressed forming federations of fog providers for task offloading purposes. Nonetheless, a few challenges remain a burden for the formation of the federations. The formation mechanisms used to structure the federations of providers are still not fully stable. This causes a problem because a structureless federation can lead to an underperforming infrastructure. Furthermore, most of the literature ignored the honesty metrics of the providers and how trustworthy they are in allocating the agreed-upon resources for processing the tasks. Moreover, adopting a central reputation mechanism is questionable in terms of reliability due to many complications including the lack of consensus. In this work, we develop a Blockchain-based reputation mechanism for assisting the formation of fog federations for IoV applications. Our mechanism comprises on-chain smart contracts for storing and manipulating the providers’ reputations, and an off-chain Hedonic-based formation process that considers the parameters extracted from the chain to build the federations. We develop smart contracts using Solidity and deploy them on the Ethereum Blockchain. We test our mechanism using the EUA dataset as a proof of concept and compare it to other works in the literature. The results obtained show that our approach is able to enhance the overall payoff and quality of service in the IoV paradigm.

Federated learning extends the centralized machine learning architecture by enabling data privacy for its providers. The distributed structure of the emerged federated architecture imposes a problem of the data being not independent and identically distributed (non-IID), which drastically affects the performance of the learning process. While the majority of the recent works in the federated learning domain have accepted this limitation, only a few scholars addressed the non-IID problem straightforwardly. Nevertheless, these works lack the fundamental analysis of the data’ IIDness, and/or contradict the privacy feature of the federated learning paradigm. In this paper, we focus on evaluating the harmony of the participants by studying their data distribution and calculating their level of compatibility. The devised tool, in this work, is an assessment technique integrated within the federated learning framework to analyze the data distribution among the trainers. Our proposed method is proven by experimenting with several scenarios, and results show that our utility can fairly assess the selected participants before initiating the learning process.

Federated learning using fog computing can suffer from the dynamic behavior of some of the participants in its training process, especially in Internet-of-Vehicles where vehicles are the targeted participants. For instance, the fog might not be able to cope with the vehicles’ demands in some areas due to resource shortages when the vehicles gather for events, or due to traffic congestion. Moreover, the vehicles are exposed to unintentionally leaving the fog coverage area which can result in the task being dropped as the communications between the server and the vehicles weaken. The aforementioned limitations can affect the federated learning model accuracy for critical applications, such as autonomous driving, where the model inference could influence road safety. Recent works in the literature have addressed some of these problems through active sampling techniques, however, they suffer from many complications in terms of stability, scalability, and efficiency of managing the available resources. To address these limitations, we propose a horizontal-based federated learning architecture, empowered by fog federations, devised for the mobile environment. In our architecture, fog computing providers form stable fog federations using a Hedonic game-theoretical model to expand their geographical footprints. Hence, providers belonging to the same federations can migrate services upon demand in order to cope with the federated learning requirements in an adaptive fashion. We conduct the experiments using a road traffic signs dataset modeled with intermodal traffic systems. The simulation results show that the proposed model can achieve better accuracy and quality of service than other models presented in the literature.