Internet Of Things Devices Research Articles

Federated Learning Enabled Fog Computing Framework for DDoS Mitigation in SDN Based IoT Networks

DDoS attacks require efficient detection due to challenges like latency, false positives, and resource inefficiency, especially in IoT and Fog-SDN setups. A framework combining ML and DL for real-time DDoS detection was evaluated against Logistic Regression, Random Forest, and CNN using benchmark datasets. Key metrics included accuracy, precision, recall, F1-score, false positive rate, latency, and resource use. The framework achieved 98.3% accuracy, surpassing CNN (95.6%), Random Forest (91.5%), and Logistic Regression (86.8%). Precision, recall, and F1-score were 98.7%, 97.8%, and 98.2%. False positive rates were 2.1%, compared to CNN (4.3%), Random Forest (6.4%), and Logistic Regression (8.2%). Latency was 30–110 ms for 100–500 requests in Fog-SDN versus 50–180 ms in cloud setups. Resource utilization was efficient: fog nodes 70%, cloud 60%, and IoT devices 40%. The proposed framework ensures high accuracy, low latency, and efficient resource use, perfect for real-time DDoS detection in Fog-SDN environments.

Designing of an efficient lightweight symmetric based block cipher algorithm for battery scarce Internet of Things (IoT) devices

ABSTRACT Lightweight cryptographic algorithms are critical for safeguarding resource-constrained contexts like IoT devices, RFID systems, and embedded software. This approach is intended to provide robust security while requiring minimal computing overhead, assuring efficiency in power-limited systems. The NYKS lightweight cryptographic technique is constructed and assessed using comprehensive security and performance measures to determine its applicability for such contexts. Security is evaluated using the Avalanche Effect, which takes into account non-linearity, confusion, diffusion rates, linear closeness, and Hamming distance to ensure resistance to cryptanalysis. Efficiency is determined by comparing encryption speed and throughput to computational efficiency. Experimental results show that NYKS has a greater Avalanche Effect (53%) than PRESENT (51%), Improved PRESENT (52%), and TWINE (50%), indicating superior diffusion features that improve cryptographic security. Additionally, NYKS exceeds its predecessors in efficiency, achieving an encryption speed of 100 ns, which is much quicker than PRESENT (310 ns), Improved PRESENT (310 ns), and TWINE (360 ns). Furthermore, NYKS has the maximum throughput at 640 Mbps, making it an excellent choice for high-speed, low-power applications, particularly in battery-constrained IoT devices and RFID systems.

Quantized convolutional neural networks: a hardware perspective

With the rapid development of machine learning, Deep Neural Network (DNN) exhibits superior performance in solving complex problems like computer vision and natural language processing compared with classic machine learning techniques. On the other hand, the rise of the Internet of Things (IoT) and edge computing set a demand on executing those complex tasks on corresponding devices. As the name suggested, deep neural networks are sophisticated models with complex structures and millions of parameters, which overwhelm the capacity of IoT and edge devices. To facilitate the deployment, quantization, as one of the most promising methods, is proposed to alleviate the challenge in terms of memory usage and computation complexity by quantizing both the parameters and data flow in the DNN model into formats with shorter bit-width. Consistently, dedicated hardware accelerators are developed to further boost the execution efficiency of DNN models. In this work, we focus on Convolutional Neural Network (CNN) as an example of DNNs and conduct a comprehensive survey on various quantization and quantized training methods. We also discuss various hardware accelerator designs for quantized CNN (QCNN). Based on the review of both algorithm and hardware design, we provide general software-hardware co-design considerations. Based on the analysis, we discuss open challenges and future research directions for both algorithms and corresponding hardware designs of quantized neural networks (QNNs).

Decentralized trust framework for smart cities: a blockchain-enabled cybersecurity and data integrity model

The rapid evolution of smart cities has led to transformative advancements through the integration of IoT devices, sensors, and data-driven systems, yet has simultaneously exposed critical vulnerabilities in cybersecurity, data integrity, and trust management. This research proposes a Decentralized Trust Framework that leverages blockchain technology, AI-driven threat detection, and a Lightweight Adaptive Proof-of-Stake (LA-PoS) consensus mechanism to address these challenges. The framework integrates three key layers: a Blockchain Layer for decentralized trust and immutability, a Cybersecurity Layer employing cryptographic standards and AI-based anomaly detection, and a Data Integrity Protocol Layer for real-time synchronization and tamper-proof data validation. Performance evaluations indicate the framework achieves a threefold increase in transaction throughput, a 30% reduction in latency, and enhanced energy efficiency compared to traditional blockchain systems. Security metrics highlight a 98.2% threat detection rate and a substantial reduction in false positives, while resource optimization nearly doubles IoT device battery life. The framework demonstrates applicability in critical smart city use cases, including smart traffic management, energy systems, and public safety, providing secure, scalable, and efficient solutions for urban infrastructures. Despite these advancements, challenges such as interoperability among heterogeneous systems, computational overhead for IoT devices, and policy adoption persist. Future research will focus on optimizing interoperability protocols, incorporating quantum-resistant cryptographic techniques, and extending the framework to emerging domains such as autonomous systems and smart healthcare. The proposed framework provides a robust foundation for building sustainable, resilient, and trustworthy urban ecosystems, bridging gaps in current smart city technologies.

Blockchain enhanced distributed denial of service detection in IoT using deep learning and evolutionary computation

The Internet of Things (IoT) is emerging as a new trend mainly employed in developing numerous vital applications. These applications endure on a federal storage framework primarily concerned with multiple issues. Blockchain technology (BC) is one of the supportive methods for developing IoT-based applications. It is employed to solve the problems encountered in IoT applications. The attack Distributed Denial of Service (DDoS) is one of the leading security attacks in IoT systems. Attackers can effortlessly develop the exposures of IoT gadgets and restrain them as fragments of botnets to commence DDoS threats. The IoT devices are said to be resource-constrained with computing resources and restricted memory. As a developing technology, BC holds the possibility of resolving security problems in IoT. This paper proposes the Metaheuristic-Optimized Blockchain Framework for Attack Detection using a Deep Learning Model (MOBCF-ADDLM) method. The main intention of the MOBCF-ADDLM method is to deliver an effective method for detecting DDoS threats in an IoT environment using advanced techniques. The BC technology is initially applied to mitigate DDoS attacks by presenting decentralized security solutions. Furthermore, data preprocessing utilizes the min-max scaling method to convert input data into a beneficial format. Additionally, feature selection (FS) is performed using the Aquila optimizer (AO) technique to recognize the most relevant features from input data. The attack classification process employs the deep belief network (DBN) technique. Finally, the red panda optimizer (RPO) model modifies the hyper-parameter values of the DBN model optimally and results in higher classification performance. A wide range of experiments with the MOBCF-ADDLM approach is performed under the BoT-IoT Binary and Multiclass datasets. The performance validation of the MOBCF-ADDLM approach portrayed a superior accuracy value of 99.22% over existing models.

Auto Slot Using AI

Parking slot management has become increasingly critical with the growing number of vehicles in urban areas. This paper presents a smart parking slot identification system that aims to minimize time spent searching for available. parking spaces and optimize space utilization. The proposed system employs image processing techniques and/or sensor-based technologies to detect vacant and occupied slots in real-time. Using tools such as cameras, ultrasonic sensors, or IoT devices, data is collected and processed through algorithms that analyze the status of each parking slot. The system can then display the availability to users via a mobile application or digital signage, guiding drivers directly to empty spots. Experimental results demonstrate improved accuracy in slot detection and a significant reduction in parking search time, contributing to reduced traffic congestion and enhanced user experience

SDT‐MCS: Topology‐Aware Microservice Orchestration With Adaptive Learning in Cloud‐Edge Environments

ABSTRACTThe exponential growth of IoT devices poses unprecedented challenges to cloud‐edge collaboration, particularly in microservice‐based industrial scenarios where real‐time sensor data flows through complex service chains. Current approaches suffer from critical issues: Inefficient service deployment, load imbalance, and service instability. Specifically, the performance degradation is particularly severe when critical path services experience sudden load spikes, often leading to cascading delays across entire service chains. This paper presents SDT‐MCS (Service Dependency Topology‐aware Microservice Collaborative Scheduling), a framework that combines topology‐oriented federated learning with lightweight Actor‐Critic reinforcement mechanisms. For resource collaboration, we propose a topology‐aware pre‐deployment algorithm that leverages federated learning to optimize global resource orchestration while considering both resource constraints and service dependencies. For service collaboration, we design a chain‐based scheduling mechanism that employs Actor‐Critic reinforcement learning for local dynamic adjustment, enabling rapid response to workload variations while maintaining service chain stability. We implement our framework on EdgeCloudSim and evaluate it using production workload traces from industrial robotics and smart city scenarios, with additional validation on a physical testbed. Experimental results demonstrate that our topology‐aware pre‐deployment reduces average latency by 31.6% in ETL scenarios compared to baseline approaches. Furthermore, our chain‐based scheduling achieves 35.4% latency reduction under high concurrency while maintaining service stability between 65%–70% through dynamic load balancing.

Enhancing anomaly detection and prevention in Internet of Things (IoT) using deep neural networks and blockchain based cyber security

The rapid adoption of Internet of Things (IoT) devices has significantly increased cybersecurity risks, making them vulnerable to anomalies, attacks, and unauthorized access. Traditional security mechanisms struggle to handle the massive data flow, real-time processing requirements, and evolving cyber threats in IoT networks. This paper presents an integrated approach using Deep Neural Networks and Blockchain technology (DNNs-BCT) to enhance anomaly detection and prevention in IoT environments. Our proposed framework employs DNNs for intelligent anomaly detection, leveraging multi-layer feature extraction and adaptive learning mechanisms. The DNN model is trained on IoT traffic datasets to classify network behavior as normal or anomalous, effectively detecting threats such as Distributed Denial of Service (DDoS) attacks, malware injections, and insider threats. Unlike traditional rule-based intrusion detection systems (IDS), the DNN continuously learns and adapts to new attack patterns, improving detection accuracy and false-positive reduction. This study integrates Blockchain technology into the IoT ecosystem to ensure data integrity, transparency, and decentralized security. Each IoT device logs its activity onto a private blockchain network, preventing data tampering, unauthorized access, and single points of failure. The blockchain employs smart contracts for automated threat response, instantly mitigating malicious activity without human intervention. This distributed ledger approach enhances trust, authentication, and secure communication across IoT devices. The synergy between DNN-based anomaly detection and Blockchain-based security provides a robust, scalable, and adaptive solution for real-time cybersecurity threats in IoT networks. With a low false-positive rate of 15.42% and a strong detection accuracy of 99.18%, the proposed model successfully identifies malicious activity, including malware injections and Distributed Denial of Service (DDoS) assaults. Blockchain technology replaces single points of failure and forbids illegal changes by providing data integrity, openness, and decentralizing powers. Furthermore, smart contracts allow autonomous, real-time attack responses, enhancing reaction time efficiency (95.25%) and general system scalability (94.96%).

Research on the Strategy of Promoting Rural Tourism Development Through IoT Technology in Rural Revitalization

Rural areas increasingly rely on Internet of Things (IoT) technologies to enhance tourist experiences, optimize resource organization, and ensure the sustainability of tourism operations. However, despite the possibility of these methods, difficulties like insufficient infrastructure, poor data integration, and inadequate tourism strategies prevent rural tourism from reaching its full potential. The research objective is to examine the function of IoT in promoting rural tourism growth, specifically through the optimization of tourism images and the use of deep learning (DL) models for personalized experiences. The data consist of multiple components, including rural tourism images, IoT data (such as sensor data), and tourist-related activities or behavior data. The data are preprocessed using normalization, and principal component analysis (PCA) is employed for dimension reduction. Waterwheel plant optimized intelligent shuffleNet (WPO-IntSNet) helps to recommend particular images based on tourist behavior data. Waterwheel plant optimization (WPO) helps to optimize various criteria, including resource allocation, energy use, and even individualized tourism experiences. The IntSNet was modified to enhance its performance for tasks, such as recognizing images, enhancement, and content recognition in rural tourism images. It would be great for real-time image processing on IoT devices in rural areas. The findings show that IoT-enabled solutions considerably improve resource management, whereas DL algorithms create tourist experiences by assessing preferences and behaviors. The results outperformed the traditional method in terms of accuracy (98.82%), precision (97.9%), recall (96.4%), and F1-score (96.9%). Tourism images developed with DL algorithms are shown to attract more visitors, resulting in increased engagement and satisfaction. Finally, IoT and DL have the potential to alter rural tourism, helping to revitalize rural communities.

Application of Internet of Things Data Mining in Sports Teaching and Training System Optimization

In the digital era, the integration of advanced technologies such as the Internet of Things (IoT), data mining, and Deep Learning (DL) has significantly transformed various sectors, including education and sports. Traditional sports teaching methods, such as manual observation, video analysis, verbal feedback, and repetitive drills, largely rely on manual observation and subjective assessment, which often result in inconsistencies and limited feedback. This research proposes a novel sports teaching and training system that combines IoT-based data mining with machine learning techniques to enhance the effectiveness, precision, and intelligence of Physical Education (PE). With the growing demand for personalized and data-driven training approaches, IoT devices such as wearable sensors, smart garments, and motion trackers are employed to continuously collect real-time physiological and motion-related data from students and athletes during training sessions. The collected data undergo preprocessing through data mining techniques aimed at eliminating noise and normalization to identify behavioral patterns critical to training optimization. Spatial features are extracted and refined using Power Spectral Density (PSD) analysis. To enhance prediction accuracy and model scalability, an Artificial Gorilla Troops Optimizer-driven Scalable Extreme Learning Machine (AGTO-SELM) is introduced for efficient classification and forecasting of training outcomes. The proposed system delivers adaptive feedback and personalized training recommendations, enabling coaches and educators to make informed decisions in real time and closely monitor progress. To assess its performance and practical utility, the model is evaluated through experimental simulations in actual sports teaching environments. The findings reveal that the proposed model achieves a substantial accuracy improvement of 97.4%, along with higher specificity, sensitivity, and F1-score. This research underscores the transformative potential of intelligent technologies in evolving traditional physical education into a smart, adaptive, and data-informed learning experience.

Hybrid Proximal Policy Optimization based Adaptive Link Load Balancing with INT in Secure SDN-IoT Networks

Background/Objectives: The Internet of things with software defined networking has evolved and is facing vigorous traffic patterns and security threats in the realm of data communication. Balancing the load becomes a challenge with IoT devices. The objective is to develop Hybrid Proximal Policy Optimization based adaptive link load balancing method to address the issues of dynamic traffic patterns and security threats in SDN-IoT, thereby enhancing network performance. Method: The research follows a two level architecture which includes a centralized SDN controller which is responsible for training policy with global telemetry and switches that perform real time inference using local state observation. The system constructs the state representations using live INT metrics. Link utilization, queue length, delay, packet-loss and trust score help to find the robust path cost estimation. The PPO selects the optimal forwarding paths aided by a dynamic reward function. The policy is synchronized with edge switches for low-latency, decentralized decision-making. Findings: The effectiveness of the proposed algorithm is validated under normal and adversarial conditions. While comparing ECMP, DQN, A3C and Trust AODV, the proposed method demonstrates superior scalability with 37% reduction in 99th percentile latency. Efficient load balancing is achieved with 60-62% stable link utilization and 75 ms threat response latency at 90% attack intensity. The F1 score of 95% for DDoS, 92% for INT spoofing indicates its higher detection accuracy. These results show that the significant improvements are achieved in load distribution, reduced latency, and in security resiliency. Novelty: This research article uniquely integrates real time telemetry, trust based security model and deep reinforcement learning model to achieve load balancing in SDN IoT network. The real time INT data and trust metrics enables decentralized and lowlatency path selection. The balance between QoS and security is achieved by the dynamic reward function providing a robust and scalable solution. Keywords: Load balancing; SDN-IoT; Link load balancing; INT; Trust aware routing; QoS

Fuzzy-Based Multi-Modal Query-Forwarding in Mini-Datacenters

The rapid growth of Internet of Things (IoT) enabled devices in industrial environments and the associated increase in data generation are paving the way for the development of localized, distributed datacenters. In this paper, we have proposed a novel mini-datacenter in the form of wireless sensor networks to efficiently handle query-based data collection from Industrial IoT (IIoT) devices. The mini-datacenter comprises a command center, gateways, and IoT sensors, designed to manage stochastic query-response traffic flow. We have developed a duplication/aggregation query flow model, tailored to emphasize reliable transmission. We have developed a dataflow management framework that employs a multi-modal query forwarding approach to forward queries from the command center to gateways under varying environments. The query forwarding includes coarse-grain and fine-grain strategies, where the coarse-grain strategy uses a direct data flow using a single gateway at the expense of reliability, while the fine-grain approach uses redundant gateways to enhance reliability. A fuzzy-logic-based intelligence system is integrated into the framework to dynamically select the appropriate granularity of the forwarding strategy based on the resource availability and network conditions, aided by a buffer watching algorithm that tracks real-time buffer status. We carried out several experiments with gateway nodes varying from 10 to 100 to evaluate the framework’s scalability and robustness in handling the query flow under complex environments. The experimental results demonstrate that the framework provides a flexible and adaptive solution that balances buffer usage while maintaining over 95% reliability in most queries.

Classifying IoT firmware security threats using image analysis and deep learning

<p>As the internet of things (IoT) grows, its embedded devices face increasing vulnerability to firmware-based attacks. The lack of robust security mechanisms in IoT devices makes them susceptible to malicious firmware updates, potentially compromising entire networks. This study addresses the classification of IoT firmware security threats using deep learning and image-based analysis techniques. A publicly available dataset of 32×32 grayscale images, derived from IoT firmware samples and categorized as benignware, hackware, and malware, was utilized. The grayscale images were converted into three-channel RGB format to ensure compatibility with convolutional neural networks (CNNs). We tested multiple pre-trained CNN architectures, including SqueezeNet, ShuffleNet, MobileNet, Xception, and ResNet50, employing transfer learning to adapt the models for this classification task. Both ResNet50 and ShuffleNet achieved exceptional performance, with 100% accuracy, precision, recall, and F1-score. These results validate the effectiveness of our methodology in leveraging transfer learning for IoT firmware classification while maintaining computational efficiency, making it suitable for deployment in resource-constrained IoT environments. T</p>

Enhanced security for IoT cloud environments using EfficientNet and enhanced football team training algorithm

The growing implementation of Internet of Things (IoT) technology has resulted in a significant increase in the number of connected devices, thereby exposing IoT-cloud environments to a range of cyber threats. As the number of IoT devices continues to grow, the potential attack surface also enlarges, complicating the task of securing these systems. This paper introduces an innovative approach to intrusion detection that integrates EfficientNet with a newly refined metaheuristic known as the Enhanced Football Team Training Algorithm (EFTTA). The proposed EfficientNet/EFTTA model aims to identify anomalies and intrusions in IoT-cloud environments with enhanced accuracy and efficiency. The effectiveness of this model is measured using a standard dataset and is compared against some other methods during performance metrics. The results indicate that the proposed method surpasses existing techniques, demonstrating improved accuracy over 98.56% for NSL-KDD and 99.1% for BoT-IoT in controlled experiments for the protection of IoT-cloud infrastructures.

Enhancing IoT Smart Door Lock Security Using Chameleon Swarm Algorithm, SHA-256, and ECC

The rapid proliferation of Internet of Things (IoT) devices, spanning from smart home appliances to wearable technology, has significantly heightened concerns regarding security and privacy across various sectors. As cyber threats become increasingly sophisticated and frequent, the urgency for robust, adaptable security frameworks within IoT infrastructures is more critical than ever. This study introduces a cutting-edge security framework tailored for IoT-based smart door locks, which employs a novel integration of the Chameleon Swarm Algorithm (CSA), Secure Hash Algorithm SHA-256, and Elliptic Curve Cryptography (ECC). We conducted comprehensive performance evaluations in a Microsoft Visual Studio 2012 environment, where our proposed framework was benchmarked against conventional hybrid methods based on Genetic Algorithms (GA) and Firefly Algorithm such as - SHA-256-ECC-GA and SHA-256-ECC-FA. These evaluations demonstrated that our framework significantly enhances security performance, achieving up to 15.17% faster encoding times at 100 iterations and markedly quicker decoding times at 150 iterations compared to the benchmark techniques. The improvements confirm the framework’s effectiveness in not only bolstering IoT device security but also in its potential for scalability and adaptability across diverse IoT applications. Furthermore, the integration of advanced algorithms like CSA, SHA-256, and ECC effectively addresses critical vulnerabilities, particularly in smart door locks, thereby paving the way for safer, more secure IoT environments. This study exemplifies how innovative cryptographic techniques can be strategically applied to meet the evolving security demands of the IoT landscape.

Attack detection in internet of things networks with deep learning using deep transfer learning method

Cybersecurity becomes a crucial part within the information management framework of internet of things (IoT) device networks. The large-scale distribution of IoT networks and the complexity of communication protocols used are contributing factors to the widespread vulnerabilities of IoT devices. The implementation of transfer learning models in deep learning can achieve optimal performance faster than traditional machine learning models, as they leverage knowledge from previous models that already understand these features. Base model was built using the 1-dimension convolutional neural network (1D-CNN) method, using training and test data from the source domain dataset. Model 1 was constructed using the same method as base model. The test and training data used for model 1 were from the target domain dataset. This model successfully detected known attacks at a rate of 99.352%, but did not perform well in detecting unknown attacks, with an accuracy of 84.645%. Model 2 is an enhancement of model 1, incorporating transfer learning from the base model. Its results significantly improved compared to model 1 testing. Model 2 has an accuracy and precision rate of 98.86% and 99.17 %, respectively, allowing it to detect previously unknown attacks. Even with a slight decrease in normal detection, most attacks can still be detected.

KronNet a lightweight Kronecker enhanced feed forward neural network for efficient IoT intrusion detection

The rapid expansion of Internet of Things (IoT) networks necessitates efficient intrusion detection systems (IDS) capable of operating within the stringent resource constraints of IoT devices. This study introduces KronNet, a lightweight feed-forward neural network enhanced with Kronecker product operations, designed for real-time IoT intrusion detection. KronNet leverages Gaussian Mixture Model (GMM)-based oversampling and a hybrid loss function combining Focal Loss and Cross-Entropy with adaptive class weighting to address class imbalance, ensuring robust detection across diverse attack types. Evaluated on the CICIoT2023 and BoT-IoT datasets, KronNet achieves exceptional performance, with accuracies of 99.01% and 99.91%, weighted F1-scores of 99.01% and 99.91%, and low false positive rates of 0.03% and 0.01%, respectively. The model operates with minimal computational overhead, utilizing 5,074 parameters (19.82 KB) for CICIoT2023 and 4,703 parameters (18.37 KB) for BoT-IoT, with inference times of 0.209 ms and 0.208 ms. Post-quantization, memory usage reduces to 4.96 KB and 4.59 KB, with negligible accuracy degradation (0.06% and 0.01% loss). Compared to state-of-the-art models, KronNet demonstrates up to 15,829× lower FLOPS and 12,010× faster inference, making it a highly efficient solution for edge deployment in resource-constrained IoT environments. This work advances IoT cybersecurity by delivering a scalable, accurate, and lightweight IDS capable of real-time threat detection.

A compact dual-band reconfigurable antenna with metamaterial for IoT applications

This study presents a compact antenna designed for Internet of Things (IoT) applications, utilizing advanced wireless communication technologies. The antenna is designed to operate at dual frequencies (2.4 GHz and 6.04 GHz) with 12 multipath T-structured metamaterials. In Mode 1 (D1 OFF), the antenna operates at 2.4 GHz with a bandwidth of 80 MHz (2.36–2.44 GHz). In Mode 2 (D1 ON), it functions at 6.04 GHz with a bandwidth of 300 MHz (5.9–6.2 GHz). The design employs a commercially available FR-4 substrate with a relative permittivity of 4.3 and a loss tangent of 0.025, all within a compact size of (0.16λ₀ × 0.12λ₀ × 0.0112λ₀). The antenna radiator integrates a single PIN diode (SMP1340-079LF) along with a complete biasing circuit to achieve reconfigurability. The proposed design overcomes the conventional limitations by integrating T-structured metamaterials to achieve dual-band operation in a compact size. This antenna is ideal for wireless communication applications due to its manufacturability, enhanced gain, and low return loss. It is well suited for widely used frequency ranges, including Wi-Fi and Bluetooth. The results demonstrate that a miniaturized antenna with excellent efficiency has achieved, making it a promising solution for next-generation IoT devices.

Towards IoT device privacy & data integrity through decentralized storage with blockchain and predicting malicious entities by stacked machine learning

Towards IoT device privacy & data integrity through decentralized storage with blockchain and predicting malicious entities by stacked machine learning

A hybrid and self-adaptive QoS and QoE-driven RAT selection strategy for IoT devices

A hybrid and self-adaptive QoS and QoE-driven RAT selection strategy for IoT devices