An Asynchronous Quasi‐Cloud/Edge/Client Collaborative Federated Learning Mechanism for Fault Diagnosis

Ma, Xue; Wen, Chenglin

doi:10.1049/cje.2021.07.008

Cited by 12 publications

(9 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We can extend the method in this paper to distributed filtering. Similarly, for the distributed model of Federated learning, we will also applicable [ 39 , 40 ].…”

Section: Discussionmentioning

confidence: 99%

A High-Order Kalman Filter Method for Fusion Estimation of Motion Trajectories of Multi-Robot Formation

Wang

Liu

Wen

2022

Sensors

View full text Add to dashboard Cite

Multi-robot motion and observation generally have nonlinear characteristics; in response to the problem that the existing extended Kalman filter (EKF) algorithm used in robot position estimation only considers first-order expansion and ignores the higher-order information, this paper proposes a multi-robot formation trajectory based on the high-order Kalman filter method. The joint estimation method uses Taylor expansion of the state equation and observation equation and introduces remainder variables on this basis, which effectively improves the estimation accuracy. In addition, the truncation error and rounding error of the filtering algorithm before and after the introduction of remainder variables, respectively, are compared. Our analysis shows that the rounding error is much smaller than the truncation error, and the nonlinear estimation performance is greatly improved.

show abstract

“…We can extend the method in this paper to distributed filtering. Similarly, for the distributed model of Federated learning, we will also applicable [ 39 , 40 ].…”

Section: Discussionmentioning

confidence: 99%

A High-Order Kalman Filter Method for Fusion Estimation of Motion Trajectories of Multi-Robot Formation

Wang

Liu

Wen

2022

Sensors

View full text Add to dashboard Cite

show abstract

“…There are two main reasons for EMD frequency aliasing: (1) there is discontinuity in the signal; (2) there is interaction between signals. In order to continuously improve the EMD in this method, Wu and Huang [19] proposed the improved method EEMD.…”

Section: Improved Eemd Methodsmentioning

confidence: 99%

“…By conducting collaborative reasoning and training on the edge and cloud, the problem of industrial intelligence diagnosis being applied in enterprise applications is solved. MA Xue, WEN Chenglin [2] and others proposed an asynchronous quasi cloud/edge/client collaborative joint learning mechanism for fault diagnosis, established a new asynchronous quasi cloud/edge/client collaborative Federated learning mechanism, and verified the effectiveness of the algorithm through the data of rotating machinery. Sun Ming [3] proposed a new task unloading strategy based on Deep reinforcement learning CTOSDRL to solve the problem of task unloading for multi-user collaboration in the cloud.…”

Section: Literature Review 21 Development Of Cloud Edge Collaboration...mentioning

confidence: 99%

Fault Diagnosis of Transfer Learning Equipment Based on Cloud Edge Collaboration + Confrontation Network

2023

Teh. vjesn.

View full text Add to dashboard Cite

With the continuous improvement of product quality, production efficiency, and complexity, higher requirements are put forward for the reliability and stability of equipment, and the difficulty of real-time diagnosis of faults and functional failures is also increasing. The traditional fault diagnosis methods based on signal processing and Convolutional neural network cannot meet the requirements of on-site online real-time fault diagnosis of equipment. One is that the vibration signals on the industrial site are superimposed on each other, nonlinear and unstable and traditional feature extraction methods take a long time, resulting in unstable extraction results. Second, massive data and fault diagnosis algorithms need rich computing and storage resources. The traditional Convolutional neural network method conflicts with the real-time response requirements of fault diagnosis. At the same time, different models of fault diagnosis models have poor generalization ability, and the diagnostic accuracy is not high or even impossible to diagnose. To solve the above problems, this paper proposes a fault diagnosis method based on industrial Internet platform, which is equipment cloud edge collaboration + adaptive countermeasure network Transfer learning. On the edge side, the vibration signals collected from key components of the model are processed using empirical mode decomposition (EEMD) to solve the problem of signal nonlinearity and stationarity. In the cloud, EEMD signals of different models are decomposed into source domain and target domain for confrontation training, which is used as the input of the improved domain adversarial network model DANN (Domain Adversarial Neural Networks), so as to improve the accuracy of fault diagnosis of different models by using cloud computing power and the improved adversarial network Transfer learning algorithm. Through the analysis of experimental data, this paper verifies that the model after the confrontation network Transfer learning is more accurate than the traditional fault diagnosis method. Through the coordination of computing resources and real-time requirements, real-time diagnosis of cloud side collaborative bearing fault is realized.

show abstract

“…A comparison of the performance of several algorithms is shown in Table 6. [36] 29.50% No Federated + SKF [36] 77.72% Yes Ave-FL [37] 27.67% Yes α-FedAvg(α = 25%) [38] 91.32% No CEC + Average [39] 92.89% No i-MFN…”

Section: Experimental Analysismentioning

confidence: 99%

Multi-Level Federated Network Based on Interpretable Indicators for Ship Rolling Bearing Fault Diagnosis

Wang

Zhang

2022

JMSE

View full text Add to dashboard Cite

The federated learning network requires all the connection weights to be shared among the server and clients during training which increases the risk of data leakage. Meanwhile, the traditional federated learning method has a poor diagnostic effect for non-independently identically distributed data. In order to address these issues, a multi-level federated network based on interpretable indicators was proposed in this manuscript. Firstly, an interpretable adaptive sparse deep network is constructed based on the interpretability principle. Secondly, the relevance map of the network is constructed based on interpretable indicators. Based on this map, the contribution of the connection weights in the network is used to build a multi-level federated network. Finally, the effectiveness of the proposed algorithm has been proved through experimental validation in the paper.

show abstract

An Asynchronous Quasi‐Cloud/Edge/Client Collaborative Federated Learning Mechanism for Fault Diagnosis

Cited by 12 publications

References 20 publications

A High-Order Kalman Filter Method for Fusion Estimation of Motion Trajectories of Multi-Robot Formation

A High-Order Kalman Filter Method for Fusion Estimation of Motion Trajectories of Multi-Robot Formation

Fault Diagnosis of Transfer Learning Equipment Based on Cloud Edge Collaboration + Confrontation Network

Multi-Level Federated Network Based on Interpretable Indicators for Ship Rolling Bearing Fault Diagnosis

Contact Info

Product

Resources

About