Summary
ASPIRE Workshop
2024
Session Number:P
Session:
Number:P-4
Blind Array Algorithm Adaptation by CNN-assisted SIR Estimation using IQ Constellation
Kazuki Maruta, Shun Kojima,
pp.-
Publication Date:2024/03/05
Online ISSN:2188-5079
Summary:
This paper presents a blind interference suppression performance enabled by deep learning assisted interference estimation using visualized wireless signal information, that is, IQ constellation. Co-channel interference becomes more extensive due to frequency resource exhaustion and small cell deployment which had been triggered by mobile traffic explosion. Multi-antenna signal processing known as blind adaptive array (BAA) is an effective means to suppress co-channel interference without any a priori information such as channel state information. These algorithms, e.g. constant modulus algorithm (CMA) and power inversion (PI) should be optimally selected according to interference level. We previously verified the possibility of signal-to-interference (SIR) classification by multi-layered deep convolutional neural network (CNN), utilizing IQ constellation images where includes the desired and interference signals for model training. This paper clarifies the interference suppression performance of BAAs adaptively switched according to estimated SIR values.