In this paper, beetle antennae search algorithm-based mixed kernel relevance vector regression (BASA-MkRVR) model is presented and applied to predict the dissolved gases content in power transformer, and beetle antennae search algorithm (BASA) is used to select the appropriate kernel parameters and controlled parameter. The RVR model with RBF kernel (RBFRVR) and the RVR model with Sigmoid kernel (SigmoidRVR) are, respectively, used to compare with the proposed BASA-MkRVR model in order to testify the superiority of BASA-MkRVR compared with RBFRVR and SigmoidRVR. The experimental results indicate that BASA-MkRVR has more excellent prediction ability for the dissolved gases content in power transformer oil than RBFRVR and SigmoidRVR.

With the advent of 5G, the amount of data generated by the Internet of Things (IoT) will explode, and the amount of information processed by prognostics health management (PHM) will also increase dramatically. The traditional cloud computing framework will generate requirements of higher network bandwidth and real-time data processing. This paper studies the system requirement analysis, system structure and function in detail, and an intelligent fault diagnosis system based on the edge computing of vibration signals is proposed for the equipment faults of rotating components such as gears and bearings, and this paper studies the system requirement analysis, system structure and function in detail. The edge of the system implements online diagnostics based on the real-time equipment data and the downloaded cloud model, and preprocesses the vibration signal at the edge. The cloud uses the data reported by the edge to train the model. In this paper, we use wavelet packet transform, Fisher discrimination criterion, combined with the edge of the Support Vector Machine (SVM) and Relevance Vector Machine (RVM) proved in this paper, computing framework can be real-time processing of sensor data, reduces network bandwidth resource consumption and latency, and can continuously update the computing model with the latest data sets. The intelligent fault diagnosis system based on vibration signal edge computing has application potential in highly time-sensitive fault diagnosis of rotating components as well as in vibration signal monitoring systems with large data volumes.

Matching area selection is a core technology for underwater gravity aided navigation. To improve the matching accuracy and track continuity, a curvature directional feature based matching area clustering method (CD-MAC) is proposed. Based on the directivity of principal curvature and the gravity surface spatial variation of gaussian curvature, curvature directional feature is designed. After classification of matching areas by k-means, a two-stage matching area re-closing mechanism that combines morphological operations and status updates derived based on cluster information is presented. This approach improves path continuity while maintaining the navigational feasibility of the added areas. The experimental results demonstrated that the optimal areas selected by the proposed feature is effectively matched in test navigation directions compared with the other gravity features. For the 4-connected and 8-connected, the connectivity performance of the matching area re-closing mechanism improves by 72.73% and 60.00%, respectively. Additionally, 83.89% directions are successfully matched. It indicates that the selected areas not only have strong connectivity, but also have good directional adaptation, which can improve the positioning accuracy of underwater vehicles.

This article proposes an observer-based predefined-time robust formation controller for uncertain multi-UAV systems with external disturbances by integrating the sliding-mode technique with neural networks. The predefined-time strategy is developed to enhance formation tracking performance, including faster convergence speed, higher accuracy, and better robustness, while the sliding-mode scheme, integrated with the neural network, is effectively utilized to handle uncertain dynamics and external disturbances, ensuring adaptivity, availability, and robustness. Furthermore, the stability of the closed-loop control system is analyzed using Lyapunov’s method applied to the formulation of the quadrotor Newton–Euler model. This analysis fully guarantees that the desired formation position tracking and attitude stabilization goals for multi-UAV (quadrotor) systems can be achieved. Finally, the effectiveness of the theoretical results is validated through comprehensive simulations.