Low Probability of Intercept Radar Signal Recognition Based on the Improved AlexNet Model
In: Proceedings of the 2nd International Conference on Digital Signal Processing, 2018-02-25
Online
unknown
Zugriff:
In order to solve the problem that low recognition rate and less signal type of low probability of intercept (LPI) radar signal at -6dB of low signal-to-noise ratio (SNR), the paper presents a method based on Smooth Pseudo Wigner-Ville distribution (SPWVD) for signal time-frequency analyze and an improved-AlexNet deep convolutional neural network (DCNN) model for low probability of intercept radar signal to classification. First of all, the time-frequency images of radar signals are accessed by time-frequency analysis of SPWVD. Next, to fit for input of the model size selected later and weaken the influence of noise, time-frequency images must be denoised and clipped processing by wavelet threshold filtering and bi-cubic interpolation. After that, employing TensorFlow frame and GPU to built improved-AlexNet and that accelerate the training of model. Last but not least, The model will extract feature and classify 10 type of radar signals include that CW, LFM, NLFM, BPSK, Costas, Frank, T1, T2, T3 and T4. The simulation results show that the overall correct recognition rate(CRR) of radar signals is 92.5% at -6dB that higher than existing methods.
Titel: |
Low Probability of Intercept Radar Signal Recognition Based on the Improved AlexNet Model
|
---|---|
Autor/in / Beteiligte Person: | Li-min, Guo ; Xin, Chen |
Link: | |
Zeitschrift: | Proceedings of the 2nd International Conference on Digital Signal Processing, 2018-02-25 |
Veröffentlichung: | ACM, 2018 |
Medientyp: | unknown |
DOI: | 10.1145/3193025.3193037 |
Schlagwort: |
|
Sonstiges: |
|