An essential aspect of Industrial Internet of Things (IIoT) systems lies in their reliability and resilience against failures. Fault detection serves as a crucial method for mitigating errors, leading to reduced downtime. Previous studies have predominantly focused on fault detection using centralized Artificial Intelligence (AI) approaches, wherein participant information is centralized and forwarded to a central server. However, Federated Learning (FL) offers a solution to these issues, enhancing the system’s reliability. In this study, we propose a Decentralized FL (DFL) approach for collaborative learning in bearing fault detection. DFL is preferred over centralized FL due to its elimination of a single point of failure. By leveraging the decentralized FL concept, the vulnerability of the collaborative framework to attacks can be minimized. Our proposed DFL integrates continual learning techniques to reduce communication overhead. The results demonstrate that decentralized collaborative learning achieves satisfactory performance, with an accuracy rate of 96.08% and a learning time reduction of up to 37.52%.

The distributed and heterogeneous connections in the digital twin (DT)-based industrial Internet of Things (IIoT) are vulnerable to cyber-attacks and malicious activities. This study proposes a permissioned blockchain-assisted collaborative and decentralized cyber threat detection for securing DT-based IIoT networks. A context-aware network intrusion detection system (C-NIDS) model was developed using factorized and grouped convolution structures to detect adversarial attacks in virtual and physical environments. A verifiable off-chain aggregation technique with a digital signature is implemented to provide a trustworthy and anti-tampering aggregated model with minimum transaction time. The results exhibit the robustness of the proposed model by achieving an attack detection accuracy of 99.50% using a lightweight model structure with trainable parameters of 4,634 and MFLOPs calculation of 0.0088. Moreover, the verifiable off-chain aggregation performs a total transaction time of 0.0244 seconds.

The heterogeneous connections in metaverse environments pose vulnerabilities to cyber-attacks. To prevent and mitigate malicious network activities in a distributed metaverse, conventional intrusion detection systems (IDS) have communication overhead and privacy concerns. Federated learning (FL) techniques are widely employed to develop IDS frameworks and enable privacy-preserving collaborative learning schemes in decentralized ecosystems. However, the vanilla FL system utilizes a centralized FL aggregation technique, which introduces a single point of failure (SPoF) and potential unauthorized aggregators, allowing malicious clients to inject false data parameters, known as poisoning attacks. Furthermore, low-quality clients in the FL system can result in degraded model performance and hinder convergence. This study proposes a secure and reliable blockchain-aided federated learning (BFL)-based IDS framework using a lightweight model for securing metaverse. An authorized federated IDS is proposed to establish a trustworthy decentralized aggregation mechanism, utilizing proof-of-authority (PoA) consensus. The proposed federated IDS implemented a hybrid client selection (HCS) technique, considering the accuracy and reputation of client histories, to select high-quality metaverse edge devices. Additionally, a fairness ERC-20 token-based incentive mechanism was developed to reward selected FL clients as a token of appreciation for their contribution to the FL training processes. According to the IDS framework measurements, the proposed model performs better than the existing approaches for detecting cyber-attacks in metaverse environments, achieving an accuracy of 99.28% with trainable parameters of 1.8K and mega floating-point operations (MFLOPs) of 0.0016.

Software-defined networking (SDN)-based Industrial Internet of Things (IIoT) networks have a centralized controller that is a single attractive target for unauthorized users to attack. Cybersecurity in IIoT networks is becoming the most significant challenge, especially from increasingly sophisticated Distributed Denial-of-Service (DDoS) attacks. This situation necessitates efficient approaches to mitigate recent attacks following the incompetence of existing techniques that focus more on DDoS detection. Most existing DDoS detection capabilities are computationally complex and are no longer efficient enough to protect against DDoS attacks. Thus, the need for a low-cost approach for DDoS attack classification. This study presents a competent feature selection method extreme gradient boosting (XGBoost) for determining the most relevant data features with a hybrid convolutional neural network and long short-term memory (CNN-LSTM) for DDoS attack classification. The proposed model evaluated the CICDDoS2019 data set with improved accuracy and low-complexity capability for low latency IIoT requirements. Performance results show that the proposed model achieves a high accuracy of 99.50% with a time cost of 0.179 ms.

Unauthorized users may attack centralized controllers as an attractive target in software-defined networking (SDN)-based industrial cyber-physical systems (CPS). Managing high-complexity deep learning (DL)-based intrusion classification to recognize and prevent attacks in the industrial Internet of Things (IIoT) networks with low-latency requirements is challenging. Moreover, a centralized DL-based intrusion detection system (IDS) leads to privacy concerns and communication overhead issues during data uploading to a cloud server for training processes in IIoT environments. This study proposes federated learning (FL)-based low-complexity intrusion detection and classification in SDN-enabled industrial CPS. This framework utilizes Chi-square and Pearson correlation coefficient (PCC) feature selection methods to select potential features, which help reduce the model’s complexity and boost performance. The proposed model evaluated the SDN and IIoT-related InSDN and Edge-IIoTset datasets. The model measurement shows that the proposed model achieves high accuracy, low computational cost, and a low-complexity model architecture compared with state-of-the-art approaches.

Software-defined networks (SDN)-based industrial cyber-physical systems (CPS) enable customizing development opportunities with integrated network interconnection to perform monitoring, measurement, control system, and security tasks. The extensive connectivity and the vast amount of data exchange in the SDN-based industrial CPS environment make it vulnerable to cyberattacks. Furthermore, an SDN controller is a single attractive target for an attack. It is challenging when the SDN controller manages DL-based high-complexity intrusion detection in an IIoT network with low latency requirements to identify and prevent attacks. This study proposes a lightweight intrusion detection model in an SDN-based industrial CPS environment. The proposed model was evaluated using a recent publicly SDN-related cyber-security InSDN dataset. The experimental results show that the proposed model outperforms the state-of-the-art by achieving 98.95% accuracy, 99.00% precision, 98.91% recall, and a 0.164 ms time cost when using the LightGBM feature selection technique.