Specific emitter identification (SEI) refers to the technology that uses the inherent defects of the physical layer of a hardware device to identify and uniquely associate a single emitter. Given its unique capability to automatically extract high-dimensional features, deep learning has recently shown great potential in SEI. However, the training process of supervised neural networks largely depends on the generalization of sample distribution and integrity of data labels. In specific non-cooperative areas such as electronic reconnaissance and spectrum monitoring, the unknown radio frequency devices and the complexity of the electromagnetic environment make it difficult to utilize sufficient labeled samples and capture the potential distribution of the data. Therefore, we introduce semi-supervised learning into SEI and propose a self-classification generative adversarial network (GAN) using bispectrum-based feature extraction. The bispectrum estimation of the signal is used as the feature representation of emitters, and the label information is embedded into the input latent layer to guide the training of the GAN. The shared weight update of the classification network is realized through semi-supervised training, and the optimization function of the generator is redefined to solve the mode collapse caused by the game principle. We innovatively extend SEI to communication scenarios with multiple relays and evaluate the effectiveness of our algorithm in different communication scenarios with contaminated radio-frequency fingerprints. The algorithm is also verified using Universal Software Radio Peripheral based on software-defined radio platforms. The numerical experimental results on six modulation signals demonstrate the excellent semi-supervised classification performance of the proposed SEI scheme in multiple scenarios.

Better neural networks, more powerful computer hardware and signal Big Data make deep learning increasingly important in Specific Emitter Identification (SEI). However, its implementation uses large amounts of resources and releases CO2. With the requirement of green and low-carbon resource sustainability, deep learning-based SEI faces the great challenge of compressing the model while ensuring its performance. In this article, we propose a novel green and low resource specific emitter identification method using complex networks and Fisher pruning (GLR-SEI). Specifically, we divide the learning process into two stages: knowledge distillation and fisher pruning, which involves transferring knowledge from complex number networks to real number networks. The proposed approach is evaluated using aircraft Automatic Dependent Surveillance Broadcast (ADS-B) data collected in the real world and many pruning experiments are performed. In addition, we add descriptive experiments on Raspberry Pi to provide a more effective data presentation in order to better demonstrate our research results. The experimental results show that the proposed lightweight network in this article can reduce the recognition rate by only 0.7% and the inference time by 10.1% compared to the complex neural network.

This letter presents our initial results for real-time orthogonal frequency division multiplexing (OFDM) signal modulation classification based on deep learning and software-defined radio. We generate a modulation classification dataset under a dynamic fading channel, including 6 different OFDM modulation signals, and propose a novel neural network with triple-skip residual stack (TRS) as the basic unit. Each TRS has multiple residual units with gradually increasing convolutional layers. Finally, a near real-time classification system is designed based on the proposed network and GNU Radio. The processing delay incurred by the detection and modulation classification is about 4 ms. It is worth mentioning that the classification accuracy can reach 64% at −10 dB, which is about 7% higher than ResNet and VGG.

Communication signal modulation recognition has important research value in the fields of cognitive electronic warfare, communication countermeasures and non-collaborative communication. However, traditional signal recognition methods usually suffer some drawbacks, such as low accuracy, poor scalability, dependence on expert characteristics, and poor applicability to real-world environments. Therefore, in this paper, a real-time modulation recognition system based on deep learning and software-defined radio (SDR) technology is designed. In the first step, an improved residual neural network is designed. A multi-skip residual stack (MRS) is designed to preserve more initial residuals information on the multi-scale feature map, which can simultaneously learn the deep and shallow characteristics of the signal. Then, a multi-skip residual network is designed with the MRS as the basic unit, and the network is trained using an adaptive moment estimation optimization algorithm. Finally, the network is tested on public datasets. In the second step, the network is embedded in a SDR platform composed of a GNU Radio and an universal software radio peripheral to realize real-time recognition of the input signal. Experiments show that this system has strong real-time capabilities, high recognition accuracy and considerable robustness.

Specific emitter identification (SEI) uses the unintentional modulation information carried by the emitter waveform, i.e., radio frequency fingerprints, to realize the matching identification of the received signal and its corresponding emitter. We propose an emitter identification scheme based on deep learning (DL). The received signal is subjected to time-varying filtered empirical mode decomposition (tvf-EMD). The obtained intrinsic mode functions (IMFs) are subjected to Hilbert transformation to obtain a 3D-Hilbert spectrum, using amplitude frequency aggregation characteristics obtained by projection to express the nonstationary information of the signal and calculate its bispectral diagonal slice as the second feature. A squeeze-and-excitation block (SEB) is introduced to a convolutional neural network (CNN) to focus on the effective area in the feature map, and support vector machine (SVM) is used to fuse the decision information of the two types of features. We collect four types of steady-state signals on a software-defined radio (SDR) sensor platform based on GNU Radio and universal software radio peripherals (USRP), and study the algorithm performance. The results show that our scheme has strong generalization, high recognition accuracy, and broad application prospects compared with other SEI methods.