ISBN: 978-981-18-1791-5 DOI: 10.18178/wcse.2021.06.048
EMI Signal Classification Based on Improved HHT and CNN
Abstract— This paper proposes a new single channel EMI signal classification method. In this method, we propose to use an improved HHT algorithm, BEMD algorithm and HT algorithm, to convert nonlinear and non-steady-state single-channel electromagnetic interference signals into time-frequency domain matrices, which solves the problem of traditional EMD algorithms. After that, we use the time-frequency domain matrix as the input of the CNN to classify the EMI signals. By the simulation signal experiment and EEG signal experiment verification, the proposed algorithm has a good classification effect
Index Terms— Signal classification, HHT, EMD, EMI
Hongyi Li, Di Zhao, Kejia Hu
School of Mathematical Sciences, Beihang University, CHINA
Cite: Hongyi Li, Di Zhao, Kejia Hu, "EMI Signal Classification Based on Improved HHT and CNN," 2021 The 11th International Workshop on Computer Science and Engineering (WCSE 2021), pp. 333-338, Shanghai, China, June 19-21, 2021.