Basic study of electromagnetic noise waveform extraction using independent component analysis
Nao Takahashi, Shinobu Ishigami, Ken Kawamata · 2021 IEEE Asia-Pacific Microwave Conference (APMC) · 2021
In device such as a drone that has a build-in 5G (5thgeneration mobile communications) to control itself, electromagnetic noises generated by the device itself may cause electromagnetic interference to radio waves for the communications and control. This is the so-called a problem of intra system electromagnetic compatibility (EMC). On the other hand, when a measuring electromagnetic noise with an antenna, not only electromagnetic noise that interferes with the communications, but also electromagnetic fields that include all waveform components are measured at the same time. In this paper, we focused on a technology of independent component analysis (ICA), which has been used to separate and extract specific waveforms from mixed signals, for example, brain waves or acoustic waves. Applying ICA to the extraction of electromagnetic waves, it was examined whether a specific electromagnetic noise component could be extracted from the waveform observed by antennas. In this study, Assuming electromagnetic waves generated from a drone and mobile terminals, the result of extracting electromagnetic waves in a specific frequency band using ICA for multiple observation signals are shown. From the result, we also confirmed the applicability of ICA to a problem of the intra system EMC.