A Hybrid CNN–LSTM Network for the Classification of Human Activities Based on Micro-Doppler Radar
Many deep learning (DL) models have shown exceptional promise in radar-based human activity recognition (HAR) area. For radar-based HAR, the raw data is generally converted into a 2-D spectrogram by using short-time Fourier transform (STFT). All the existing DL methods treat the spectrogram as an op...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2020-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8978926/ |
| id |
doaj-a523a04984a3476faeaea8509e9cef75 |
|---|---|
| record_format |
Article |
| spelling |
doaj-a523a04984a3476faeaea8509e9cef752021-03-30T02:36:32ZengIEEEIEEE Access2169-35362020-01-018247132472010.1109/ACCESS.2020.29710648978926A Hybrid CNN–LSTM Network for the Classification of Human Activities Based on Micro-Doppler RadarJianping Zhu0Haiquan Chen1Wenbin Ye2https://orcid.org/0000-0001-6978-813XCollege of Electronic and Information Engineering, Shenzhen University, Shenzhen, ChinaCollege of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen, ChinaCollege of Electronic and Information Engineering, Shenzhen University, Shenzhen, ChinaMany deep learning (DL) models have shown exceptional promise in radar-based human activity recognition (HAR) area. For radar-based HAR, the raw data is generally converted into a 2-D spectrogram by using short-time Fourier transform (STFT). All the existing DL methods treat the spectrogram as an optical image, and thus the corresponding architectures such as 2-D convolutional neural networks (2D-CNNs) are adopted in those methods. These 2-D methods that ignore temporal characteristics ordinarily lead to a complex network with a huge amount of parameters but limited recognition accuracy. In this paper, for the first time, the radar spectrogram is treated as a time sequence with multiple channels. Hence, we propose a DL model composed of 1-D convolutional neural networks (1D-CNNs) and long short-term memory (LSTM). The experiments results show that the proposed model can extract spatio-temporal characteristics of the radar data and thus achieves the best recognition accuracy and relatively low complexity compared to the existing 2D-CNN methods.https://ieeexplore.ieee.org/document/8978926/Radar signal processinghuman activity recognitionconvolutional neural networkrecurrent neural networkdeep learning |
| collection |
DOAJ |
| language |
English |
| format |
Article |
| sources |
DOAJ |
| author |
Jianping Zhu Haiquan Chen Wenbin Ye |
| spellingShingle |
Jianping Zhu Haiquan Chen Wenbin Ye A Hybrid CNN–LSTM Network for the Classification of Human Activities Based on Micro-Doppler Radar IEEE Access Radar signal processing human activity recognition convolutional neural network recurrent neural network deep learning |
| author_facet |
Jianping Zhu Haiquan Chen Wenbin Ye |
| author_sort |
Jianping Zhu |
| title |
A Hybrid CNN–LSTM Network for the Classification of Human Activities Based on Micro-Doppler Radar |
| title_short |
A Hybrid CNN–LSTM Network for the Classification of Human Activities Based on Micro-Doppler Radar |
| title_full |
A Hybrid CNN–LSTM Network for the Classification of Human Activities Based on Micro-Doppler Radar |
| title_fullStr |
A Hybrid CNN–LSTM Network for the Classification of Human Activities Based on Micro-Doppler Radar |
| title_full_unstemmed |
A Hybrid CNN–LSTM Network for the Classification of Human Activities Based on Micro-Doppler Radar |
| title_sort |
hybrid cnn–lstm network for the classification of human activities based on micro-doppler radar |
| publisher |
IEEE |
| series |
IEEE Access |
| issn |
2169-3536 |
| publishDate |
2020-01-01 |
| description |
Many deep learning (DL) models have shown exceptional promise in radar-based human activity recognition (HAR) area. For radar-based HAR, the raw data is generally converted into a 2-D spectrogram by using short-time Fourier transform (STFT). All the existing DL methods treat the spectrogram as an optical image, and thus the corresponding architectures such as 2-D convolutional neural networks (2D-CNNs) are adopted in those methods. These 2-D methods that ignore temporal characteristics ordinarily lead to a complex network with a huge amount of parameters but limited recognition accuracy. In this paper, for the first time, the radar spectrogram is treated as a time sequence with multiple channels. Hence, we propose a DL model composed of 1-D convolutional neural networks (1D-CNNs) and long short-term memory (LSTM). The experiments results show that the proposed model can extract spatio-temporal characteristics of the radar data and thus achieves the best recognition accuracy and relatively low complexity compared to the existing 2D-CNN methods. |
| topic |
Radar signal processing human activity recognition convolutional neural network recurrent neural network deep learning |
| url |
https://ieeexplore.ieee.org/document/8978926/ |
| work_keys_str_mv |
AT jianpingzhu ahybridcnnx2013lstmnetworkfortheclassificationofhumanactivitiesbasedonmicrodopplerradar AT haiquanchen ahybridcnnx2013lstmnetworkfortheclassificationofhumanactivitiesbasedonmicrodopplerradar AT wenbinye ahybridcnnx2013lstmnetworkfortheclassificationofhumanactivitiesbasedonmicrodopplerradar AT jianpingzhu hybridcnnx2013lstmnetworkfortheclassificationofhumanactivitiesbasedonmicrodopplerradar AT haiquanchen hybridcnnx2013lstmnetworkfortheclassificationofhumanactivitiesbasedonmicrodopplerradar AT wenbinye hybridcnnx2013lstmnetworkfortheclassificationofhumanactivitiesbasedonmicrodopplerradar |
| _version_ |
1724184874507567104 |