Attention-based Recurrent Neural Network Model for Radar High-resolution Range Profile Target Recognition
XU Bin①② CHEN Bo①② LIU Hongwei①② JIN Lin②
①(National Laboratory of Radar Signal Processing, Xidian University, Xi’an 710071, China) ②(Collaborative Innovation Center of Information Sensing and Understanding, Xidian University, Xi’an 710071, China)
To improve the performance of radar High-Resolution Range Profile (HRRP) target recognition, a new attention-based model is proposed based on time domain feature. This architecture encodes the time domain feature which can reveal the correlation inside the target with Recurrent Neural Network (RNN). Then, this model gives a weight to each part and sums the hidden feature with each weight for the final recognition. Experiments based on measured data show that the attention-based model is effective for radar HRRP recognition. Furthermore, the proposed method can still find the support areas even with the removed test data.
ZHANG Junying and BAI Xianghui. Study of the HRRP feature extraction in radar automatic target recognition[J]. Systems Engineering and Electronics, 2007, 29(12): 2047- 2053. doi: 10.3321/j.issn:1001-506x.2007.12.012.
LIANG Haitao, ZHANG Xueli, TONG Chuangming, et al. SVM target identification method based on wavelet decomposition and azimuth average HRRP[J]. Journal of Data Acquisition & Processing, 2010, 25(1): 29-35. doi: 10.3969/j.issn.1004-9037.2010.01.006.
[3]
DU Lan, LIU Hongwei, BAO Zheng, et al. Radar automatic target recognition using complex high-resolution range profiles[J]. IET Radar, Sonar & Navigation, 2007, 1(1): 18-26. doi: 10.1049/iet-rsn:20050119.
[4]
FENG B, DU L, LIU H W, et al. Radar HRRP target recognition based on K-SVD algorithm[C]. IEEE CIE International Conference on Radar, Chengdu, 2011: 642-645.
PAN Mian, WANG Penghui, DU Lan, et al. Radar HRRP target recognition based on truncated stick-breaking hidden Markov model[J]. Journal of Electronics & Information Technology, 2013, 35(7): 1547-1556. doi: 10.3724/SP.J.1146. 2012.01190.
[6]
PAN Mian, DU Lan, WANG Penghui, et al. Multi-task hidden Markov modeling of spectrogram feature from radar high-resolution range profiles[J]. EURASIP Journal on Advances in Signal Processing, 2012, 2012(1): 1-17. doi: 10.1109/CIE-Radar.2011.6159624.
[7]
JI S H, LIAO X J, and CARIN L. Adaptive multi-aspect target classification and detection with hidden Markov models[C]. International Conference on Acoustics, Speech and Signal Processing, Montreal, 2004: 125-129.
[8]
GREGOR K, DANIHELKA I, GRAVES A, et al. DRAW: A recurrent neural network for image generation[C]. Proceedings of the 32nd International Conference on Machine Learning, Lille, 2015: 1-8.
[9]
ZAREMBA W and SUTSKEVER I. Recurrent neural network regularization[C]. International Conference on Learning Representations, San Diego, 2015: 1-8.
[10]
SRIVASTAVE N, MANSIMOV E, and SALAKHUTDINOV R. Unsupervised learning of video representations using LSTMs[C]. Proceedings of the 32nd International Conference on Machine Learning, Lille, 2015: 1-9.
[11]
LI J W, LUONG M T, and JURAFSKY D. A hierarchical neural autoencoder for paragraphs and documents[C]. Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, Beijing, 2015: 1106-1115.
[12]
CHEN J X, YANG L, ZHANG Y Z, et al. Combining fully convolutional and recurrent neural networks for 3D biomedical image segmentation[C]. 29th Conference on Neural Information Processing Systems, Barcelona, 2016: 1-9.
[13]
ELMAN J L. Finding structure in time[J]. Cognitive Science, 1990, 14(2): 179-211.
[14]
CHOROWSKI J, BAHDANAU D, SERDYUK D, et al. Attention-based models for speech recognition[C]. 27th Conference on Natural Language Processing Systems, Montreal, 2015: 1-19.
[15]
SU B and LU S J. Accurate scene text recognition based on recurrent neural network[C]. 12th Asian Conference on Computer Vision, Singapore, 2015: 35-48.
