Linqing Guo
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Linqing Guo has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Jifei Pan
College of Electronic Engineering, National University of Defense Technology, Hefei 230037, China
3 shared publications - Shengli Zhang
College of Electronic Confrontation, National University of Defense Technology, No. 460 Huangshan Road, Shushan District, Hefei 230037, China
2 shared publications - Jingwei Xiong
College of Electronic Engineering, National University of Defense Technology, Hefei 230037, China
1 shared publication - Lingsi Xia
Hikvision Research Institute, Hangzhou Hikvision Digital Technology Co., Ltd., No. 555 Qianmo Road, Binjiang District, Hangzhou 310051, China
1 shared publication - Long Tan
College of Electronic Confrontation, National University of Defense Technology, No. 460 Huangshan Road, Shushan District, Hefei 230037, China
1 shared publication - Mingyang Du
College of Electronic Engineering, National University of Defense Technology, Hefei 230037, China
1 shared publication
Latest Publications
- Open Access
Self-Supervised Representation Learning for Quasi-Simultaneous Arrival Signal Identification Based on Reconnaissance Drones
Reconnaissance unmanned aerial vehicles are specifically designed to estimate parameters and process intercepted signals for the purpose of identifying and locating radars. However, distinguishing quasi-simultaneous arrival signals (QSAS) has become increasingly challenging in complex electromagnetic environments. In order to address the problem, a framework for self-supervised deep representation learning is proposed. The framework consists of two phases: (1) pre-train an autoencoder. For learning the unlabeled QSAS representation, the ConvNeXt V2 is trained to extract features from masked time–frequency images and reconstruct the corresponding signal in both time and frequency domains; (2) transfer the learned knowledge. For downstream tasks, encoder layers are frozen, the linear layer is fine-tuned to classify QSAS under few-shot conditions. Experimental results demonstrate that the proposed algorithm can achieve an average recognition accuracy of over 81% with the signal-to-noise ratio in the range of −16∼16 dB. Compared to existing CNN-based and Transformer-based neural networks, the proposed algorithm shortens the time of testing by about 11× and improves accuracy by up to 21.95%.
Open AccessSelf-Supervised Representation Learning for Quasi-Simultaneous Arrival Signal Identification Based on Reconnaissance Drones
Reconnaissance unmanned aerial vehicles are specifically designed to estimate parameters and process intercepted signals for the purpose of identifying and locating radars. However, distinguishing quasi-simultaneous arrival signals (QSAS) has become increasingly challenging in complex electromagnetic environments. In order to address the problem, a framework for self-supervised deep representation learning is proposed. The framework consists of two phases: (1) pre-train an autoencoder. For learning the unlabeled QSAS representation, the ConvNeXt V2 is trained to extract features from masked time–frequency images and reconstruct the corresponding signal in both time and frequency domains; (2) transfer the learned knowledge. For downstream tasks, encoder layers are frozen, the linear layer is fine-tuned to classify QSAS under few-shot conditions. Experimental results demonstrate that the proposed algorithm can achieve an average recognition accuracy of over 81% with the signal-to-noise ratio in the range of −16∼16 dB. Compared to existing CNN-based and Transformer-based neural networks, the proposed algorithm shortens the time of testing by about 11× and improves accuracy by up to 21.95%.
- Open Access
- 8 July 2022
- Published by MDPI AG in Electronics
- Vol. 11 (14) , 2142
Embedding Soft Thresholding Function into Deep Learning Models for Noisy Radar Emitter Signal Recognition
Radar emitter signal recognition under noisy background is one of the focus areas in research on radar signal processing. In this study, the soft thresholding function is embedded into deep learning network models as a novel nonlinear activation function, achieving advanced radar emitter signal recognition results. Specifically, an embedded sub-network is used to learn the threshold of soft thresholding function according to the input feature, which results in each input feature having its own independent nonlinear activation function. Compared with conventional activation functions, the soft thresholding function is characterized by flexible nonlinear conversion and the ability to obtain more discriminative features. By this way, the noise features can be flexibly filtered while retaining signal features, thus improving recognition accuracy. Under the condition of Gaussian and Laplacian noise with signal-to-noise ratio of −8 dB to −2 dB, experimental results show that the overall average accuracy of soft thresholding function reached 88.55%, which was 11.82%, 8.12%, 2.16%, and 1.46% higher than those of Sigmoid, PReLU, ReLU, ELU, and SELU, respectively.
- 8 July 2022
- Published by MDPI AG in Electronics
- Vol. 11 (14) , 2142
Open AccessEmbedding Soft Thresholding Function into Deep Learning Models for Noisy Radar Emitter Signal Recognition
Radar emitter signal recognition under noisy background is one of the focus areas in research on radar signal processing. In this study, the soft thresholding function is embedded into deep learning network models as a novel nonlinear activation function, achieving advanced radar emitter signal recognition results. Specifically, an embedded sub-network is used to learn the threshold of soft thresholding function according to the input feature, which results in each input feature having its own independent nonlinear activation function. Compared with conventional activation functions, the soft thresholding function is characterized by flexible nonlinear conversion and the ability to obtain more discriminative features. By this way, the noise features can be flexibly filtered while retaining signal features, thus improving recognition accuracy. Under the condition of Gaussian and Laplacian noise with signal-to-noise ratio of −8 dB to −2 dB, experimental results show that the overall average accuracy of soft thresholding function reached 88.55%, which was 11.82%, 8.12%, 2.16%, and 1.46% higher than those of Sigmoid, PReLU, ReLU, ELU, and SELU, respectively.
- Open Access
Recognition of Noisy Radar Emitter Signals Using a One-Dimensional Deep Residual Shrinkage Network
Signal features can be obscured in noisy environments, resulting in low accuracy of radar emitter signal recognition based on traditional methods. To improve the ability of learning features from noisy signals, a new radar emitter signal recognition method based on one-dimensional (1D) deep residual shrinkage network (DRSN) is proposed, which offers the following advantages: (i) Unimportant features are eliminated using the soft thresholding function, and the thresholds are automatically set based on the attention mechanism; (ii) without any professional knowledge of signal processing or dimension conversion of data, the 1D DRSN can automatically learn the features characterizing the signal directly from the 1D data and achieve a high recognition rate for noisy signals. The effectiveness of the 1D DRSN was experimentally verified under different types of noise. In addition, comparison with other deep learning methods revealed the superior performance of the DRSN. Last, the mechanism of eliminating redundant features using the soft thresholding function was analyzed.
Open AccessRecognition of Noisy Radar Emitter Signals Using a One-Dimensional Deep Residual Shrinkage Network
Signal features can be obscured in noisy environments, resulting in low accuracy of radar emitter signal recognition based on traditional methods. To improve the ability of learning features from noisy signals, a new radar emitter signal recognition method based on one-dimensional (1D) deep residual shrinkage network (DRSN) is proposed, which offers the following advantages: (i) Unimportant features are eliminated using the soft thresholding function, and the thresholds are automatically set based on the attention mechanism; (ii) without any professional knowledge of signal processing or dimension conversion of data, the 1D DRSN can automatically learn the features characterizing the signal directly from the 1D data and achieve a high recognition rate for noisy signals. The effectiveness of the 1D DRSN was experimentally verified under different types of noise. In addition, comparison with other deep learning methods revealed the superior performance of the DRSN. Last, the mechanism of eliminating redundant features using the soft thresholding function was analyzed.