The rapid development of Industrial Internet of Things (IIoT) technology has highlighted the critical role of wireless sensor networks in enabling intelligent production and equipment monitoring. Effective sensor deployment is essential for ensuring communication quality and transmission speed in IIoT environments. This paper presents a novel sensor deployment strategy that integrates four key metrics: deployment cost, energy consumption, network connectivity, and sensing probability. To address the challenges of multi-dimensional optimization, the proposed method normalizes these metrics and assigns appropriate weights based on their relative importance. A major innovation of this approach is the inclusion of larger-scale environmental obstacles, which enhances its adaptability to diverse industrial settings and specific deployment scenarios. Through a comprehensive set of simulation experiments across different scenarios, the proposed particle swarm/genetic hybrid algorithm demonstrates superior performance compared to existing methods, even surpassing 10% in performance. Specifically, it excels in optimizing the newly introduced network performance metric and significantly improves search convergence time, making it a highly efficient and effective solution for sensor network optimization in IIoT applications.

The rapid development of the Industrial Internet of Things (IIoT) has brought enormous opportunities to the industrial environment, but has also posed significant challenges to network security, especially physical layer security, which is crucial for the reliable transmission of messages among interconnected devices and sensors. Focusing on enhancing the physical layer security in IIoT communication, this paper proposes a mechanism for the Spatial Secrecy Outage Probability (SSOP) to ensure the secure transmission of data. The SSOP algorithm assumes that the location of legitimate devices is known, while the eavesdropper devices have a random distribution of locations, forming the SSOP model related to the device locations from the perspective of insecure regions (ISRs) and secure regions (SRs), and the closed-form expression for its upper bound is derived. Subsequently, under the constraint of the SSOP conditions, we establish an optimization model with the objective of system throughput maximization based on the Spatial Secrecy Outage Probability (SRM-SSOP). By optimizing the transmission rate of legitimate channels or controlling the transmission power, the maximization of the system throughput can be achieved, thus reducing the consumption of communication resources. Through sophisticated modeling and simulation in MATLAB, we verified the accuracy of the definition and derived the upper bound of the SSOP. The experimental results also show that the SSOP algorithm gradually converges to a specific value as the number of antenna elements increases. In addition, we observed that, when the receiver device is aligned with the normal direction of the transmit antenna array, the SSOP reaches a minimum value. This performance is crucial for understanding and optimizing the security performance of IIoT communication systems.

With the development of the Industrial Internet of Things (IIoT) technology, fault diagnosis has emerged as a critical component of its operational reliability, and machine learning algorithms play a crucial role in fault diagnosis. To achieve better fault diagnosis results, it is necessary to have a sufficient number of fault samples participating in the training of the model. In actual industrial scenarios, it is often difficult to obtain fault samples, and there may even be situations where no fault samples exist. For scenarios without fault samples, accurately identifying the unknown faults of equipment is an issue that requires focused attention. This paper presents a method for the normal-sample-based mechanical equipment unknown fault detection. By leveraging the characteristics of the autoencoder network (AE) in deep learning for feature extraction and sample reconstruction, normal samples are used to train the AE network. Whether the input sample is abnormal is determined via the reconstruction error and a threshold value, achieving the goal of anomaly detection without relying on fault samples. In terms of input data, the frequency domain features of normal samples are used to train the AE network, which improves the training stability of the AE network model, reduces the network parameters, and saves the occupied memory space at the same time. Moreover, this paper further improves the network based on the traditional AE network by incorporating a convolutional neural network (CNN) and a long short-term memory network (LSTM). This enhances the ability of the AE network to extract the spatial and temporal features of the input data, further improving the network’s ability to extract and recognize abnormal features. In the simulation part, through public datasets collected in factories, the advantages and practicality of this method compared with other algorithms in the detection of unknown faults are fully verified.

With the development of the Industrial Internet of Things (IIoT) technology, fault diagnosis has emerged as a critical component of its operational reliability, and machine learning algorithms play a crucial role in fault diagnosis. To achieve better fault diagnosis results, it is necessary to have a sufficient number of fault samples participating in the training of the model. In actual industrial scenarios, it is often difficult to obtain fault samples, and there may even be situations where no fault samples exist. For scenarios without fault samples, accurately identifying the unknown faults of equipment is an issue that requires focused attention. This paper presents a method for the normal-sample-based mechanical equipment unknown fault detection. By leveraging the characteristics of the auto-encoder network (AE) in deep learning for feature extraction and sample reconstruction, normal samples are used to train the AE network. Whether the input sample is abnormal is determined via the reconstruction error and a threshold value, achieving the goal of anomaly detection without relying on fault samples. In terms of input data, the frequency domain features of normal samples are used to train the AE network, which improves the training stability of the AE network model, reduces the network parameters, and saves the occupied memory space at the same time. Moreover, this paper further improves the network based on the traditional AE network by incorporating a convolutional neural network (CNN) and a long short-term memory network (LSTM). This enhances the ability of the AE network to extract the spatial and temporal features of the input data, further improving the network’s ability to extract and recognize abnormal features. In the simulation part, through public datasets collected in factories, the advantages and practicality of this method compared with other algorithms in the detection of unknown faults are fully verified.

To ensure end-to-end Quality of Service (QoS) management in 5G-TSN converged networks, this paper proposes a dynamic weighted QoS mapping method based on Weighted Fuzzy C-Means and Three-Way Decisions (WFCM-TDwQM). The WFCM algorithm is employed to cluster Time-Sensitive Networking (TSN) flows based on their QoS attributes, reducing computational complexity. A three-way decision-based method is used to assign a reasonable and approximate set of 5G QoS Identifier (5QI) values to each cluster. Finally, dynamic weights are adjusted by considering QoS similarity and the residual load rate, enabling the system to adapt to network load changes. The experimental results show that, compared with three other mapping algorithm combinations, WFCM-TDwQM not only ensures end-to-end QoS consistency but also achieves better load balancing under varying network loads. Moreover, its mapping performance is evaluated under different network scenarios.

To ensure end-to-end Quality of Service (QoS) management in 5G-TSN Converged Networks, this paper proposes a dynamic weighted QoS mapping method based on Weighted Fuzzy C-Means and Three-Way Decisions (WFCM-TDwQM). The WFCM algorithm is employed to cluster Time-Sensitive Networking (TSN) flows based on their QoS attributes, reducing computational complexity. A three-way decision-based method is used to assign a reasonable and approximate set of 5G QoS Identifier (5QI) values to each cluster. Finally, dynamic weights are adjusted by considering QoS similarity and residual load rate, enabling the system to adapt to network load changes. Experimental results show that, compared with three other mapping algorithm combinations, WFCM-TDwQM not only ensures end-to-end QoS consistency but also achieves better load balancing under varying network loads. Moreover, its mapping performance is evaluated under different network scenarios.