Emitter Identification of Digital Modulation Transmitter Based on Nonlinearity and Modulation Distortion of Power Amplifier

Abstract

Specific transmitter identification (SEI) is a technology that uses a received signal to identify to which individual radiation source the transmitted signal belongs. It can complete the identification of the signal transmitter in a non-cooperative scenario. Therefore, there are broad application prospects in the field of wireless-communication-network security, spectral resource management, and military battlefield-target communication countermeasures. This article demodulates and reconstructs a digital modulation signal to obtain a signal without modulator distortion and power-amplifier nonlinearity. Comparing the reconstructed signal with the actual received signal, the coefficient representation of the nonlinearity of the power amplifier and the distortion of the modulator can be obtained, and these coefficients can be used as the fingerprint characteristics of different transmitters through a convolutional neural network (CNN) to complete the identification of specific transmitters. The existing SEI strategy for changing the modulation parameters of a test signal is to mix part of the test signal with the training signal so that the classifier can learn the signal of which the modulation parameter was changed. This method is still data-oriented and cannot process signals for which the classifier has not been trained. It has certain limitations in practical applications. We compared the fingerprint features extracted by the method in this study with the fingerprint features extracted by the bispectral method. When SNR < 20 dB, the recognition accuracy of the bispectral method dropped rapidly. The method in this paper still achieved 86% recognition accuracy when SNR = 0 dB. When the carrier frequency of the test signal was changed, the bispectral feature failed, and the proposed method could still achieve a recognition accuracy of about 70%. When changing the test-signal baud rate, the proposed method could still achieve a classification accuracy rate of more than 70% for four different individual radiation sources when SNR = 0 dB.

Keywords: convolutional neural network; modulator distortion; power-amplifier nonlinearity; specific emitter identification.

Figure 1

Feature-extraction flowchart.

Figure 2

Hammerstein model.

Figure 3

Phase-modulation flowchart.

Figure 4

Phase deviation caused by modulator distortion. (a) Offset of phase center point caused by I/Q gain imbalance; (b) offset of origin of phase coordinate caused by DC offset.

Figure 5

(a) Result of bispectral analysis of signal by Emitter 1 under signal-to-noise ratio of 15 dB. (b) Four paths for integrating two-dimensional bispectral features.

Figure 6

Influence of different model structures on accuracy.

Figure 7

Signal simulation flowchart: entire signal process from original message bits to demodulation after passing through the channel; partial signal enlargement after low-pass filtering. Jitter reflects nonlinear characteristics of the power amplifier in the signal-amplification process.

Figure 8

Accuracy of ASK signal features under different signal-to-noise ratio conditions through CNN and SVM classifiers.

Figure 9

With a SNR ratio of 15 dB: (a) three-dimensional visualization map that composed 1-, 7-, and 9-dimensional data of feature vector e; (b) confusion matrix corresponding to CNN classification result.

Figure 10

Recognition results of integral bispectrum algorithms, the algorithm in [27], and the proposed algorithm under different signal-to-noise ratio conditions.

Figure 11

Experimental results when using different carrier-frequency data and (a) bispectral method or (b) proposed method as the test set.

Figure 12

Relationship between recognition accuracy and SNR of the proposed method and bispectral method on test-set data of different carrier frequencies.

Figure 13

After changing the data information rate of the test set, the classification accuracy results using (a) bispectral method and (b) proposed method under different signal-to-noise ratio conditions.

Figure 14

Classification-accuracy results of the proposed method and bispectral method under different SNR conditions after changing test-set information rate.

Figure 15

Baseband signal after passing through the power amplifier.

Figure 16

Classification-accuracy rate of three different models through CNN classifier under different signal-to-noise ratio conditions.

Figure 17

Classification-accuracy results of the proposed method under different signal-to-noise ratio conditions after changing (a) the data information rate and (b) the carrier frequency of the test set.