A modified federated learning based novel multi-stage intrusion detection system for real-time identification of attacks in a smart home ecosystem that ensures privacy-preserving model aggregation across distributed IoT devices while minimizing communication overhead and adapting to dynamic network conditions is introduced.
: Federated Learning (FL) has emerged as a promising approach to build collaborative Intrusion Detection Systems (IDSs) in the IoT, e.g., in smart homes. FL allows models to be shared without exposing sensitive training data, thus protecting the privacy of IoT users. However, existing FL-based IDSs rely on assumptions that rarely hold in practice, namely homogeneous devices, synchronous participation, and benign contributors. We argue that, in real-world smart homes, IoT devices are highly heterogeneous, resource-constrained, and attractive targets for adversaries, which makes conventional FL less effective or vulnerable to poisoning attacks. We present H OME G UARD , a collaborative IDS specifically designed for the constraints and threat model of practical smart home IoT infrastructures. In our approach, we rethink FL deployment by (1) of-floading model training to gateways to manage computational heterogeneity of IoT devices and (2) organizing anomaly detection models into device-specific communities based on privacy-preserving traffic fingerprints which do not expose sensitive data. Within communities and across smart homes, H OME G UARD implements an asynchronous, hierarchical FL architecture that tolerates device churn, uneven data availability, and Byzantine participants. Further, H OME G UARD applies Byzantine-robust aggregation at two levels: within local communities, and globally in the cloud to limit the impact of compromised devices. Experimental evaluation shows that H OME G UARD achieves an average true positive rate of 97.86% locally and 97.53% globally with a 0% false positive rate, while maintaining robustness against both targeted and untargeted poisoning attacks.
Philipp Eichhammer, Christian Berger, Hans P. Reiser· Proceedings of the 23rd Inte...· 0 citations
The growing usage of Internet of Things gadgets has facilitated smart environments including automated homes, smart lighting systems, and smart industrial applications. In spite of these advantages, the IoT devices are frequently not well secured because of their poor computing power and compact architecture. This makes them very susceptible to cyberattacks that may cripple the functionality of their devices, interfere with network processes or reveal confidential information. The identification of malicious activity on the IoT networks has thus become a key challenge to cybersecurity. This paper has discussed a machine learning-driven intrusion detection system that aims to detect attacks on IoT devices. An actual IoT system based on ESP32 microcontrollers and a smart bulb was used to create network traffic both in the regular operation environment and in the simulated attack environment. To enhance the efficiency of the model, the data obtained on the traffic was subjected to various preprocessing steps such as data cleaning, feature encoding, normalization and feature selection. Local Outlier Factor (LOF) based anomaly detection method was employed to determine abnormal network behavior with a random forest classifier used to determine the category of attack. The system is constantly watching the IoT traffic and sends automatic email notifications in case of suspicious activity. The outcomes of the experiment suggest that machine learning methods can be successfully used to differentiate between legitimate and malicious network behavior, which can be used as a viable solution to enhance the security and surveillance of IoT-based systems. The proposed system achieved an accuracy of 88.22%, precision of 90.41%, recall of 85.51%, and an F1-score of 87.89%, demonstrating effective detection of malicious IoT network activity.
P. Praveen, K. Sridhar, B. Rao et al.· International journal of com...· 0 citations
Smart homes depend on interconnected sensors, cameras, routers, mobile applications, and cloud services. This connectivity improves automation and convenience, but it also expands the attack surface for Distributed Denial of Service (DDoS), Denial of Service (DoS), Mirai botnet, brute-force, spoofing, reconnaissance, and man-in-the-middle attacks. Traditional signature-based security is often insufficient because IoT devices are resource-constrained, heterogeneous, and frequently deployed with weak authentication or delayed firmware updates. This study evaluates supervised machine-learning classifiers for detecting cyberattacks in smart-home IoT network traffic using the CICIoT2023 dataset. Four algorithms, namely Random Forest, Decision Tree, k-Nearest Neighbour, and Support Vector Machine, were compared under 50:50, 70:30, and 80:20 train-test split settings. The models were evaluated using accuracy, precision, recall, and F1-score, with emphasis on DDoS, Mirai, and brute-force attack classes that are particularly relevant to smart-home environments. The findings show that tree-based classifiers are highly effective for IoT attack detection. Random Forest achieved the strongest overall accuracy and precision, while Decision Tree showed the most stable recall and F1-score for brute-force detection. The results indicate that Random Forest is suitable as a general-purpose smart-home IDS classifier, whereas Decision Tree or a hybrid ensemble strategy should be considered when missed brute-force attacks carry high operational risk. The paper contributes a clearer empirical comparison of lightweight supervised learning models and provides implementation guidance for smart-home intrusion detection systems.
Nurin Abyana Balqis Jailani, Haniza Nahar, N. M. Salleh et al.· International journal of res...· 0 citations
The rapid growth of the Internet of Things (IoT) in critical domains such as healthcare, smart cities, cybersecurity, and finance has led to the generation of large volumes of distributed data and increased susceptibility to cyberattacks. Although federated learning (FL)-based intrusion detection systems (IDS) have been introduced to support distributed learning and improve privacy, they still face several challenges, including performance limitations, high computation and communication overhead, and potential privacy attacks. To address these challenges, we propose a Random Projection-Based Personalized Federated Learning (RPPFL) framework for IoT intrusion detection. In the proposed framework, random projection, a lightweight one-way dimensionality reduction transformation, is applied at the IoT device level, which enhances privacy while lowering computational and communication costs. Furthermore, personalized federated learning at the fog layer reduces the computational burden on resource-constrained IoT devices and improves model robustness in environments where IoT data are non-independent and identically distributed (non-IID). We also introduce a conditional generative adversarial network (cGAN)-based privacy attack to evaluate the resilience of the proposed framework and demonstrate the effectiveness of our approach in preserving the privacy of IoT data. Experimental results on the RT-IoT 2022 and CIC-IoT 2023 datasets demonstrate that RPPFL provides high detection accuracy (above 95.0%) while preserving data privacy and reducing computation and communication overhead in dynamic IoT environments. The proposed framework is generalizable and applicable to a wide range of IoT intrusion detection scenarios.
Md. Morshedul Islam, Hossain Shahriar, Alfredo Cuzzocrea et al.· Annual International Compute...· 0 citations
With the rapid adoption of smart home solutions and related technologies, edge computing has emerged as a key enabler by offering low-latency data processing, increased efficiency and improved scalability. However, this integration in IoT systems introduces complex security challenges in smart home edge environments, increasingly susceptible to cyber threats such as denial-of-service (DoS), malware injection, passive surveillance, and unauthorized access. This paper investigates intelligent intrusion detection and attack classification strategies specifically designed for smart home edge systems. Using the comprehensive ML-EdgeIIoT dataset, this study designs and evaluates a machine learning-based intrusion detection framework for multiclass classification of eight categories of IoT network attacks, namely Backdoor, MITM, DDoS, Ransomware, Password Attack, SQL Injection, Prob-attacks, and Normal traffic while minimizing false positives and false negatives. The framework incorporates data cleaning, correlation- and feature importance-based feature selection, hyperparameter optimization using gridsearchCV, model training, and ensemble learning. A set of machine learning models comprising Artificial Neural Network, Balanced Random Forest, K-Nearest Neighbours, Random Forest, and Logistic Regression was implemented and comparatively evaluated. Two ensemble techniques were subsequently developed using the three best-performing classifiers: (1) a stacking ensemble with Logistic Regression as the meta-learner and (2) a Top-3 majority voting ensemble. Model performance was evaluated using accuracy, precision, recall, F1-score, confusion matrix, and ROC-AUC. Robustness and generalization of the individual machine learning models were assessed through stratified 10-fold cross-validation for the three best-performing classifiers. The Top-3 voting ensemble subsequently achieved the highest performance on the independent test set, with accuracy of 99.24%, average precision of 98.75%, recall of 99.00%, and an F1-score of 99.00% for all attack classes, while reducing misclassification compared with individual classifiers. The findings of this study significantly enhance the understanding of smart home edge computing security, which will pave the way for more robust and intelligent threat detection frameworks.
Abhay Kumar Ray, Rupak Sharma, Sunil Kumar Pandey· International Journal of Wir...· 0 citations
This paper introduces an innovative ML-based security paradigm that improves the attack detection accuracy by combining adaptive feature extraction techniques with a context-attentive hybrid mechanism and maximizes detection accuracy and computational efficiency.
P. P. Bairagi, Ashish Bagwari, Sailen Dutta Kalita et al.· international journal of eng...· 0 citations