文献类型：期刊文献

英文题名：LSTM-Based Intelligent Fault Detection for Fuzzy Markov Jump Systems and Its Application to Tunnel Diode Circuits

作者：Xue, Min[1];Yan, Huaicheng[1,2];Wang, Meng[1];Shen, Hao[3];Shi, Kaibo[2]

机构：[1]East China Univ Sci & Technol, Key Lab Smart Mfg Energy Chem Proc, Minist Educ, Shanghai 200237, Peoples R China;[2]Chengdu Univ, Sch Informat Sci & Engn, Chengdu 610106, Peoples R China;[3]Anhui Univ Technol, Sch Elect & Informat Engn, Maanshan 243002, Peoples R China

年份：2022

卷号：69

期号：3

起止页码：1099

外文期刊名：IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS

收录：;EI(收录号：20213310762926);WOS:【SCI-EXPANDED(收录号:WOS:000770045800090)】；

基金：This work was supported in part by the National Natural Science Foundation of China under Grant 62073143, Grant 61922063, and Grant 62003139; in part by the Program of Shanghai Academic Research Leader under Grant 19XD1421000; in part by the Shanghai and HongKong-Macao-Taiwan Science and Technology Cooperation Project under Grant 19510760200; in part by Shanghai Natural Science Foundation under Grant 20ZR1415200; in part by the Shanghai Shuguang Project under Grant 18SG18; and in part by the Innovation Program of Shanghai Municipal Education Commission under Grant 2021-01-07-00-02-E00107.

语种：英文

外文关键词：Computer architecture; Fault detection; Logic gates; Microprocessors; Feature extraction; Circuits and systems; Recurrent neural networks; Fault detection; fuzzy Markov jump systems; long short-term memory (LSTM); attention mechanism

摘要：This brief investigates the intelligent fault detection problem for a class of Takagi-Sugeno fuzzy Markov jump systems. The considered systems are modeled by a set of fuzzy rules and membership functions. With the help of deep learning technique, an intelligent fault detection method based on long short-term memory (LSTM) neural network with an attention mechanism is proposed. The improved LSTM model can accurately capture the fault features by making full use of previous information. Finally, the proposed approach is applied to the tunnel diode circuits and corresponding experiments are conducted to demonstrate its effectiveness, which achieves better performance in comparison with other conventional method for detection.