Limited Network Bandwidth Research Articles

FedSparse: A Communication-Efficient Federated Learning Framework Based on Sparse Updates

Federated learning (FL) strikes a balance between privacy preservation and collaborative model training. However, the periodic transmission of model updates or parameters from each client to the federated server incurs substantial communication overhead, especially for participants with limited network bandwidth. This overhead significantly hampers the practical applicability of FL in real-world scenarios. To address this challenge, we propose FedSparse, an innovative sparse communication framework designed to enhance communication efficiency. The core idea behind FedSparse is to introduce a communication overhead regularization term into the client’s objective function, thereby reducing the number of parameters that need to be transmitted. FedSparse incorporates a Resource Optimization Proximal (ROP) term and an Importance-based Regularization Weighting (IRW) mechanism into the client update objective function. The local update process optimizes both the empirical risk and communication overhead by applying a sparse regularization weighted by update importance. By making minimal modifications to traditional FL approaches, FedSparse effectively reduces the number of parameters transmitted, thereby decreasing the communication overhead. We evaluate the effectiveness of FedSparse through experiments on various datasets under non-independent and identically distributed (non-IID) conditions, demonstrating its flexibility in resource-constrained environments. On the MNIST, Fashion-MNIST, and CIFAR datasets, FedSparse reduces the communication overhead by 24%, 17%, and 5%, respectively, compared to the baseline algorithm, while maintaining similar model performance. Additionally, on simulated non-IID datasets, FedSparse achieves a 6% to 8% reduction in communication resource consumption. By adjusting the sparsity intensity hyperparameter, we demonstrate that FedSparse can be tailored to different FL applications with varying communication resource constraints. Finally, ablation studies highlight the individual contributions of the ROP and IRW modules to the overall improvement in communication efficiency.

Protocol-Based Synchronization of Stochastic Jumping Inertial Neural Networks Under Image Encryption Application.

This work investigates the protocol-based synchronization of inertial neural networks (INNs) with stochastic semi-Markovian jumping parameters and image encryption application. The semi-Markovian jumping process is adopted to characterize INNs under sudden complex changes. To conserve the limited available network bandwidth, an adaptive event-driven protocol (AEDP) is developed in the corresponding semi-Markovian jumping INNs (S-MJINNs), which not only reduces the amount of data transmission but also avoids the Zeno phenomenon. The objective is to construct an adaptive event-driven controller so that the drive and response systems maintain synchronous relationships. Based on the appropriate Lyapunov functional, integral inequality, and free weighting matrix, novel criteria are derived to realize the synchronization. Moreover, the desired adaptive event-driven controller is designed under a semi-Markovian jumping process. The proposed method is demonstrated through a numerical example and an image encryption process.

Dynamic event-triggered adaptive positioning control of unmanned surface vehicles with system uncertainties

Abstract To solve the problems of unmanned surface vehicles (USVs), such as limited network bandwidth and computing resources, a finite-time command filtering backstepping (FTCFB) robust adaptive dynamic positioning control method with dynamic event-triggered mechanisms (DETM) for unmanned surface vehicles (USVs) is proposed. It addresses thruster dynamics, system uncertainties, and unknown external disturbances, ensuring finite-time convergence while mitigating issues like communication congestion and actuator wear. Adaptive neural networks are utilized to estimate the system uncertainties, and a DETM-based FTCFB control law, validated with Lyapunov stability theory, guarantees finite-time convergence of tracking and parameter estimation errors. Numerical simulations confirm the method’s effectiveness.

Enhancing Battlefield Awareness: Integration of IoMT Sensors and Networks in National Defense Systems

The integration of the Internet of Military Things (IoMT) into national defense systems presents both opportunities and challenges for enhancing battlefield awareness and decision-making. This study explores the technical and operational barriers to IoMT deployment, such as interoperability, network bandwidth limitations, energy consumption, and cybersecurity vulnerabilities. Using qualitative research methods based on secondary data from academic articles, defense reports, and technical documents, the study investigates how these challenges impede the effective use of IoMT in military operations. The research findings highlight key issues such as the lack of standardized communication protocols across devices, the need for low-latency networks, and the heightened cybersecurity risks posed by a large number of interconnected devices. Furthermore, the study underscores the importance of energy-efficient solutions to ensure continuous device operation in combat zones. Conclusions suggest that addressing these challenges through standardization, enhanced cybersecurity measures, and adopting emerging technologies such as AI and blockchain is essential for realizing IoMT's potential. Future research should focus on human and organizational factors, geopolitical concerns, and sustainability in IoMT deployment.

A Parallel Compression Pipeline for Improving GPU Virtualization Data Transfers.

GPUs are commonly used to accelerate the execution of applications in domains such as deep learning. Deep learning applications are applied to an increasing variety of scenarios, with edge computing being one of them. However, edge devices present severe computing power and energy limitations. In this context, the use of remote GPU virtualization solutions is an efficient way to address these concerns. Nevertheless, the limited network bandwidth might be an issue. This limitation can be alleviated by leveraging on-the-fly compression within the communication layer of remote GPU virtualization solutions. In this way, data exchanged with the remote GPU is transparently compressed before being transmitted, thus increasing network bandwidth in practice. In this paper, we present the implementation of a parallel compression pipeline designed to be used within remote GPU virtualization solutions. A thorough performance analysis shows that network bandwidth can be increased by a factor of up to 2×.

Swarm of Drones for Surveillance Monitoring of a Grounded Target: An event-triggered approach

In this paper, we present a novel formation control approach in the framework of Event-Triggered (ET) control to provide a solution to the surveillance problem. To do this, we identify two main challenges which are the switching topology of the drones and the limited bandwidth of the communication network, which are also valid in formation applications. To provide a solution to switching topologies, we propose a networked continuous controller that is robust in the presence of connection switching between drones and the target. Then, we propose a networked controller with ET communication in some aperiodic instants which reduces the required bandwidth and load within the communication network. We guarantee the stability of the developed ET controller and prove that the Zeno behavior cannot occur. To validate the method, we present realistic 3D simulation results conducted in the Simulink environment of Matlab® for different scenarios. The results of the study show the effectiveness of the proposed controller, especially for limited bandwidth channels as the ETC scheme has decreased the load within the communication network while resulting in a robust and efficient formation performance. We also considered moving target scenarios with missing possibilities to validate the robustness of the proposed method.

Fully Distributed Hierarchical ET Intrusion- and Fault-Tolerant Group Control for MASs With Application to Robotic Manipulators

This paper studies group synchronization tracking problem for a class of high-order multi-agent systems (MASs) with nonidentical and unknown direction faults (NUDFs) under multiple cyber attacks (i.e., denial-of-service (DoS) attacks and false data injection attacks (FDIAs)). Be motivated by this, a fully distributed hierarchical (cyber layer and physical layer) Zeno-free event-triggered (ET) intrusion-and fault-tolerant controller is presented based on Nussbaum-type gain technique, where a positive inter-event time exists in the state-dependent asynchronous ET mechanism (ETM). The constructed virtual cyber layer realizes the interaction among different agents so as to avoid the interaction in physical processes and thus reduce the error propagation. This two layer controller greatly increase the flexibility of the controller design compared with single-layer control strategies. In addition, a novel Nussbaum function stability lemma (Lemma lem4lem4) is developed for the first time. Based on this lemma, the fully distributed adaptive controller can adjust the direction of the input to match the fault direction with the help of Nussbaum function. Finally, a robotic manipulator example demonstrates the effectiveness and merit of the proposed control scheme. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —In industrial processes, NUDFs and cyber attacks often occur in many different systems, which usually lead to performance degradation or even serious accidents. Typically, NUDFs affect the performance of chemical and industrial processes, circuits, and sensors. Meanwhile, in driverless vehicle platoon, attackers change the control signal through the wireless network, and then issue emergency braking commands, etc. Therefore, this paper designs a fully distributed hierarchical ET intrusion-and fault-tolerant scheme for high-order MASs that are usually used to model ship dynamics, robotic manipulators, and quarter-car active suspension. The control scheme gives solutions to the problem of NUDFs, multiple cyber attacks, and network bandwidth limitations at different layers, and effectively integrates the physical and cyber layers to achieve group synchronization. Meanwhile, a new event-triggered mechanism under the framework of group synchronization is developed to save communication resources. Robotic manipulator simulation studies verify the validity of the proposed scheme.

A TabPFN-based intrusion detection system for the industrial internet of things

The industrial internet of things (IIoT) has undergone rapid growth in recent years, which has resulted in an increase in the number of threats targeting both IIoT devices and their connecting technologies. However, deploying tools to counter these threats involves tackling inherent limitations, such as limited processing power, memory, and network bandwidth. As a result, traditional solutions, such as the ones used for desktop computers or servers, cannot be applied directly in the IIoT, and the development of new technologies is essential to overcome this issue. One approach that has shown potential for this new paradigm is the implementation of intrusion detection system (IDS) that rely on machine learning (ML) techniques. These IDSs can be deployed in the industrial control system or even at the edge layer of the IIoT topology. However, one of their drawbacks is that, depending on the factory’s specifications, it can be quite challenging to locate sufficient traffic data to train these models. In order to address this problem, this study introduces a novel IDS based on the TabPFN model, which can operate on small datasets of IIoT traffic and protocols, as not in general much traffic is generated in this environment. To assess its efficacy, it is compared against other ML algorithms, such as random forest, XGBoost, and LightGBM, by evaluating each method with different training set sizes and varying numbers of classes to classify. Overall, TabPFN produced the most promising outcomes, with a 10–20% differentiation in each metric. The best performance was observed when working with 1000 training set samples, obtaining an F1 score of 81% for 6-class classification and 72% for 10-class classification.

Task offloading with enhanced Deep Q-Networks for efficient industrial intelligent video analysis in edge–cloud collaboration

With the rapid development of 5G and IoT technologies, industrial video surveillance systems face increasingly significant challenges in processing large-scale data. Traditional cloud-centric video analysis methods often lead to processing delays and inefficiencies in resource allocation due to centralized data processing. In this context, the emergence of edge computing, especially in edge–cloud collaborative intelligent video analysis, offers an innovative solution. Within this architecture, task offloading strategies focus on intelligent computation task distribution to minimize latency and optimize resource utilization. However, existing methods frequently overlook key application-specific metrics in video analysis, such as accuracy and timeliness. This study introduces a novel Edge–Cloud Collaborative Industrial Video Intelligent Analysis Architecture (ECIVA) based on the GRU-Enhanced Deep Q-Network (GEDQN). This architecture integrates the strengths of Gated Recurrent Units (GRU) for processing sequential data with the efficiency of Deep Q-Networks in optimizing task-offloading decisions. By incorporating a token bucket mechanism, the architecture optimally manages the task offloading rate within the constraints of limited network bandwidth, thus ensuring high accuracy and real-time responsiveness in video analysis. Empirical evaluation with offshore drilling platform video data shows that the architecture achieves 87.5% mean Average Precision (mAP) at a 0.5 IoU threshold and increases system throughput by 39.2% over traditional cloud-centric methods.

Efficient federated learning algorithm using sparse ternary compression based on layer variation classification

Federated learning is a distributed machine learning technique that ensures user privacy and enables multiple clients to jointly train a shared global model without transmitting local data. However, the frequent exchange of model parameters between numerous clients and the server results in heavy network delay and bandwidth limitation in federated learning. In view of that, we propose an efficient algorithm for federated learning using sparse ternary compression based on layer variation classification(LVC). First, use layer variation as a metric to assess the significance of each layer of the model parameters, and after client training, categorize the model parameters into different levels by the layer variation and sensitivity analysis. Then, during the upstream and downstream transmission of model parameters, we assign corresponding sparse and ternary quantization ratios for different levels to maximize compression efficiency while preserving crucial parameters. Finally, on the server side, a majority-layer aggregation strategy is adopted to further reduce the communication cost. Experimental results from image classification tasks conducted on MNIST and Fashion-MNIST datasets demonstrate that proposed LVC algorithm achieves high accuracy with minimal communication cost, thereby striking an optimal balance between communication efficiency and accuracy.

An improved hunger game search optimizer based IoT task scheduling in cloud–fog computing

Due to the rapid expansion of Internet of Things (IoT)-related applications, the utilization of cloud services is experiencing significant growth. Although cloud computing has proven its effectiveness in processing and storing data for various applications, it faces challenges in addressing certain requirements, such as the growing need for real-time or latency-sensitive applications and the limitations of network bandwidth. As a result, fog computing is often seen as a supplementary paradigm to cloud computing, providing additional capabilities and benefits to the traditional cloud paradigm, aiming to extend cloud services to edge devices and end-users. However, the limited capabilities of fog nodes require lighter tasks while other tasks that need more processing time are processed in the cloud. In the present research paper, we propose a novel algorithm that is customized for task scheduling within the context of cloud–fog computing on the Internet of Things (IoT) framework. Our approach builds upon the Hunger Game Search algorithm (HGS) as its foundation. To improve the exploitative capabilities of the HGS, our proposed method, called HGSMPA, incorporates the Marine Predator Algorithm (MPA). Through experimental evaluation using various workload traces, we have demonstrated the efficacy of HGSMPA. The findings reveal that HGSMPA surpasses alternative algorithms in terms of reducing energy consumption and minimizing the makespan time. Specifically, The empirical evaluation indicates that HGSMPA can reduce the makespan time by 19.31% for synthetic workloads and by 17.47% for real workloads as compared to similar scheduling algorithms. Moreover, HGSMPA can reduce energy consumption by 14.72% for synthetic workloads and by 17.68% for real workloads as compared to other methods.

Co-design of adaptive event-triggered mechanism and controller for networked control systems with uncertainty and time-varying delay

ABSTRACT This paper investigates the co-design problem of event-triggered mechanisms and state feedback controllers for networked control systems with parameter uncertainty and network-induced delay. Firstly, to overcome the network bandwidth limitation, an adaptive event-triggered mechanism is used to filter the required sampling signals to reduce the network resource occupancy. Second, establish a unified model subject to event triggering mechanisms, parameter uncertainty, and network delay constraints, and derive the stability conditions of the system for the model by combining Lyapunov stability theory and the linear matrix inequality method. Then, the state feedback controller is designed to satisfy the event-triggering condition and certain robust performance requirements, and an optimisation method is proposed to compromise network resource consumption and control performance. Finally, the effectiveness of the proposed method is verified by a numerical example.

Enabling Fast and Privacy-Preserving Broadcast Authentication With Efficient Revocation for Inter-Vehicle Connections

Many vehicular applications, especially safety-related ones, rely on spatial-temporal messages periodically broadcast by vehicles. In the absence of a secure authentication scheme, invalid spatial-temporal messages may be sent out by malicious vehicles. Meanwhile, malicious applications may also collect a lot of personal information from spatial-temporal messages. Since inter-vehicle connections are often deployed in high-moving traffic, any authentication must be implemented in real-time. To meet all these properties, we propose a Fast and Anonymous Spatial-Temporal Trust (FastTrust) scheme for inter-vehicle connections. In contrast to most authentication protocols which rely on fixed infrastructures, FastTrust is mostly designed on hash chains and an entropy-based commitment, and is able to secure periodic spatial-temporal messages. FastTrust also protects vehicles' privacy by deploying a pseudonym-varying scheduling mechanism to satisfy the anonymity and unlinkability requirements. Finally, in order to efficiently isolate malicious vehicles, a lightweight certificate management scheme is proposed for the limited bandwidth of vehicular networks. We provide analytical evaluations to show that our FastTrust achieves the security and privacy properties. Extensive validations are done to show that FastTrust can authenticate dozens of times faster than the existing signature algorithms, and isolate malicious vehicles at a low cost in terms of communication and computational resources.

Adaptive event-triggering mechanism based Takagi-Sugeno fuzzy automatic generation controller design for offshore wind power system

Interconnected offshore wind power system (OWPS) attracts widespread attention due to its special advantage. With the development of wind power technology, higher requirements are put forward for the reliability of networked OWPS. Considering the nonlinear characteristics of OWPS and limited network bandwidth resource, an adaptive multi-event triggering mechanism based distributed Takagi-Sugeno fuzzy robust H∞ control scheme is investigated in this paper. A modeling method for OWPS integrating communication constraints is proposed. By Lyapunov stability theory, the sufficient condition for the asymptotical stability of OWPS has been derived. Compared with traditional control scheme, the proposed Takagi-Sugeno fuzzy controller provides smoother output power, smaller torque oscillation and higher wind energy utilization. Furthermore, the co-design strategy increases the utilization rate of network bandwidth. The feasibility of adaptive multi-event triggering mechanism based Takagi-Sugeno fuzzy controller is demonstrated through a simulation example.

Observer-based event triggering security load frequency control for power systems involving air conditioning loads

&lt;p&gt;This paper presents a power system frequency control strategy that integrates an observer-based event-triggered mechanism (ETM) to defend against denial-of-service (DoS) attacks and accommodates the integration of renewable energy sources. The proposed strategy incorporates demand response by enabling air conditioning loads (ACs) to participate in frequency regulation, thereby enhancing system flexibility and stability. To address the challenges posed by limited network bandwidth and potential message blocking, the ETM minimizes communication while defending against DoS attacks. The stability of the closed-loop system is guaranteed by deriving an $ H_{\infty} $ stability criterion using the Lyapunov–Krasovskii function method, with controller parameters determined through linear matrix inequalities (LMIs). A two-area power system simulation is conducted to validate the feasibility and effectiveness of the proposed approach, demonstrating its ability to maintain stable frequency control under cyber-attack scenarios and varying renewable energy contributions.&lt;/p&gt;

Synchronize Only the Immature Parameters: Communication-Efficient Federated Learning By Freezing Parameters Adaptively

Federated learning allows edge devices to collaboratively train a global model without sharing their local private data. Yet, with limited network bandwidth at the edge, communication often becomes a severe bottleneck. In this paper, we find that it is unnecessary to always synchronize the full model in the entire training process, because many parameters already become mature (i.e., stable) prior to model convergence, and can thus be excluded from later synchronizations. This allows us to reduce the communication overhead without compromising the model accuracy. However, challenges are that the local parameters excluded from global synchronization may diverge on different clients, and meanwhile some parameters may stabilize only temporally. To address these challenges, we propose a novel scheme called Adaptive Parameter Freezing (APF), which fixes (freezes) the non-synchronized stable parameters in intermittent periods. Specifically, the freezing periods are tentatively adjusted in an additively-increase and multiplicatively-decrease manner—depending on whether the previously-frozen parameters remain stable in subsequent iterations. We also extend APF into APF# and APF++, which freeze parameters in a more aggressive manner to achieve larger performance benefit for large complex models. We implemented APF and its variants as Python modules with PyTorch, and extensive experiments show that APF can reduce data transfer amount by over 60%.

Formula omitted] containment control for multi-unmanned aerial vehicle systems: A self-triggered control scheme

This paper is concerned with the containment control problem for multi-unmanned aerial vehicle (multi-UAV) systems with data transmission among agents during a limited-bandwidth network. A good deal of data transmitted during a limited-bandwidth network may result in data collision, which brings negative effects on controller design and containment tracking implementation. To reduce the network load during data transmission, a consensus condition and a triggering subsidiary condition are first proposed to realize the containment tracking with H∞ performance by using the triggered states. Then a self-triggered scheme is established under a multi-agent framework, in which the next triggering instant is calculated beforehand by using the past states. Combined with the self-triggered scheme, a containment controller is obtained via solving the derived consensus condition, which ensures that follower UAVs are driven into a convex set surrounded by leader UAVs. Finally, the validity of the self-triggered containment control method is proved through a simulation experience.

Dynamic output feedback control for UAVs with limited network bandwidth: A GA-assisted design approach

This paper studies the dynamic output feedback (OF) control problem for unmanned aerial vehicles (UAVs) with limited network bandwidth by a genetic-algorithm-assisted design approach. In order to make full use of the limited network bandwidth, only partial measurement information is transmitted to the controller via the sensor-to-controller network. Round-Robin protocol is introduced to administrate the transmission rule. Sufficient conditions are presented to make sure that the closed-loop UAV system is exponentially stable with the required L2-gain performance. However, due to the nonlinear coupling terms of the dynamic OF controller parameters and some redundant variables in the system modeling, the established conditions are non-convex and difficult to be solved. Fortunately, genetic algorithms (GA) show good performance in solving optimization problem with nonlinear and non-convex constraints. Therefore, this paper skillfully combines GA with linear matrix inequality techniques to solve the established conditions. Finally, all the dynamic OF controller parameters and the minimal L2-gain can be obtained simultaneously. At the end of this paper, an UAV networked control example and a numerical example are performed to verify the effectiveness and superiorities of our method.

An adaptive event‐triggered H∞control framework for nonlinear networked control systems with dual‐channel quantization

AbstractNetworked control system (NCS) is widely used in various fields, but the nonlinear of the controlled plants and the limited network bandwidth bring stability issues, which will affect the performance or even instability of the NCS. To address these aforementioned challenges, this paper proposes a novel nonlinear NCS framework, by considering network‐induced delay, adaptive event‐triggered mechanism, state quantization, control input quantization, and external disturbance based on T‐S fuzzy model. Also, based on this framework, we derive an stability criterion for NCS using linear matrix inequality and propose a collaborative design method for event triggers and controllers. Finally, we verify the superiority and effectiveness of the method by giving two numerical examples.

Secure Control of Markovian Jumping Systems Under Deception Attacks: An Attack-Probability-Dependent Adaptive Event-Triggered Mechanism

This article considers the secure control of Markovian jumping systems (MJSs) under stochastic deception attacks that occurred in the communication network. Therein, the stochastic deception attack is described by a Bernoulli random variable. Considering the limited network bandwidth and the impact of deception attacks simultaneously, we propose an attack-probability-dependent adaptive event-trigger mechanism (AETM). It can not only reduce the number of the controller updates, but also adapt to the variation of the system dynamics subject to deception attacks. A mathematical model is established for the closed-loop system under stochastic deception attacks. Then, a time-dependent looped-functional is constructed to reduce the conservatism of stability results. The norm of the system state is estimated, and based on the discrete-time Lyapunov theory, a less conservative stability criterion is derived. Then, an easy-to-implement design algorithm for the controller gain is given so that the exponential stabilization in mean square sense can be realized for Markovian jumping system subject to stochastic deception attacks. Finally, an electrical circuit example is provided to validate the feasibility and superiority of the presented method.