Back to Search Start Over

Hybrid Quantum Convolutional Neural Networks for UWB Signal Classification

Authors :
Seon-Geun Jeong
Quang-Vinh Do
Hae-Ji Hwang
Mikio Hasegawa
Hiroo Sekiya
Won-Joo Hwang
Source :
IEEE Access, Vol 11, Pp 113726-113739 (2023)
Publication Year :
2023
Publisher :
IEEE, 2023.

Abstract

With the increasing requirements for location-based services for Internet of things (IoT) applications, ultrawideband (UWB) technology provides accurate indoor positioning capabilities. However, indoor environments contain various obstacles leading to significant signal propagation effects. This results in errors in the time-of-arrival-based UWB positioning system. Specifically, a non-line-of-sight (NLOS) signal induces additional distance and position errors owing to the path delay compared to a line-of-sight (LOS) signal. Therefore, UWB signal classification is essential for improving positioning accuracy. Recently, various approaches have successfully classified UWB signals, including machine-learning-based methods such as convolutional neural networks (CNNs) and long short-term memory (LSTM). This study proposes a hybrid quantum CNN (HQCNN) inspired by a CNN for UWB signal classification. HQCNN employs a classical layer before a quantum embedding circuit and variational quantum circuits for the convolutional filter. These structures enable efficient training and implementation. We used UWB channel impulse response data to demonstrate the performance of the proposed algorithm and compared the benchmarks with HQCNN using the evaluation metrics. The results showed that the HQCNN outperformed the others.

Details

Language :
English
ISSN :
21693536
Volume :
11
Database :
Directory of Open Access Journals
Journal :
IEEE Access
Publication Type :
Academic Journal
Accession number :
edsdoj.50345fb75fc04010b2536e74ca5e84c5
Document Type :
article
Full Text :
https://doi.org/10.1109/ACCESS.2023.3323019