volume | PIER Journals

Abstract

Aiming at the prevalent issue of limited dataset scale in radar target micro-Doppler effect-based human action recognition, this study constructs an improved Deep Convolutional Generative Adversarial Network (DCGAN) for radar sample data augmentation and integrates it with a Convolutional Neural Network (CNN) for human action classification. First, a millimeter-wave radar data acquisition system was established to collect human action echo signals. The raw data were preprocessed to extract micro-Doppler features, forming a 2D micro-Doppler time-frequency spectrogram dataset. Second, modifications were made to the original DCGAN by replacing the optimizer of the discriminator network and introducing an L2 regularization term to enhance the quality of micro-Doppler time-frequency spectrogram generation. Finally, a CNN architecture was implemented to classify the augmented human action samples. Experimental results demonstrate that the enhanced DCGAN-CNN framework achieves robust human action classification performance, achieving an accuracy rate of up to 97.5%. This validates the superior capability of generative adversarial networks in few-shot scenarios for radar-based human action recognition.

1. Wei, X. X., "Human posture recognition algorithm based on inertial sensors," Intelligent Computer and Applications, Vol. 12, No. 6, 97-101, 2022.

2. Xing, M. M., F. Yang, Z. H. Xin, and G. H. Wei, "Research on human action recognition using depth video," China Medical Devices, Vol. 38, No. 1, 36-41, 2023.

3. Rana, Soumya Prakash, Maitreyee Dey, Mohammad Ghavami, and Sandra Dudley, "Non-contact human gait identification through IR-UWB edge-based monitoring sensor," IEEE Sensors Journal, Vol. 19, No. 20, 9282-9293, 2019.

4. Ding, C. X., Y. H. Zhang, Z. T. Sun, et al., "Human complex motion recognition based on FMCW radar," Radar Science and Technology, Vol. 18, No. 6, 584-590, 2020.

5. Ding, Y. P., R. J. Liu, and X. M. Xu, "Micro-Doppler frequency estimation method for human targets based on continuous wave radar," Journal of Central South University (Natural Science Edition), Vol. 53, No. 4, 1273-1280, 2022.

6. Kim, Youngwook and Taesup Moon, "Human detection and activity classification based on micro-Doppler signatures using deep convolutional neural networks," IEEE Geoscience and Remote Sensing Letters, Vol. 13, No. 1, 8-12, 2016.

7. Yao, Z. P., Z. Y. Tang, and Y. J. Sun, "Human action recognition method based on convolutional neural networks," Journal of Air Force Early Warning Academy, Vol. 34, No. 5, 360-364, 2020.

8. Pan, R., "Human action recognition based on graph convolutional neural networks," Chongqing University of Posts and Telecommunications, Chongqing, China, 2022.

9. Gurbuz, Sevgi Zubeyde and Moeness G. Amin, "Radar-based human-motion recognition with deep learning: Promising applications for indoor monitoring," IEEE Signal Processing Magazine, Vol. 36, No. 4, 16-28, 2019.

10. Chen, Foji, Feng Zhu, Q. Xiao Wu, J. Zheng, and X. Zhang, "A review of generative adversarial networks and their applications in image generation," Chinese Journal of Computers, Vol. 44, No. 2, 347-369, 2021.

11. Alnujaim, Ibrahim, Daegun Oh, and Youngwook Kim, "Generative adversarial networks for classification of micro-Doppler signatures of human activity," IEEE Geoscience and Remote Sensing Letters, Vol. 17, No. 3, 396-400, 2019.

12. Qu, B. and H. Li, "Research on gesture recognition based on CNN and DCGAN," Journal of Heilongjiang Bayi Agricultural University, Vol. 33, No. 3, 79-84, 2021.

13. Chen, V. C., D. Tahmoush, and W. J. Miceli, Radar Micro-Doppler Signatures: Processing and Applications, IET Digital Library, London, 2014.
doi:10.1049/PBRA034E

14. Goodfellow, Ian, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio, "Generative adversarial networks," Communications of the ACM, Vol. 63, No. 11, 139-144, 2020.

15. Radford, Alec, Luke Metz, and Soumith Chintala, "Unsupervised representation learning with deep convolutional generative adversarial networks," arXiv preprint arXiv:1511.06434, 2015.

16. Yu, H. Y., L. Yin, S. F. Li, and S. Lyu, "Few-shot radar modulation signal recognition algorithm based on generative adversarial networks," Journal of Xidian University, Vol. 48, No. 6, 96-104, 2021.

17. Salimans, T., I. Goodfellow, W. Zaremba, V. Cheung, A. Radford, and X. Chen, "Improved techniques for training GANs," 30th Conference on Neural Information Processing Systems (NIPS 2016), Vol. 29, 1-9, Barcelona, Spain, 2016.

Dr. Cheng Luo

Professor Qiusheng Li

Dr. Yingjie Zhong