This paper introduces a novel automatic modulation classification (AMC) method using the graph-based fractional lower-order cyclic-spectrum analysis in the α-stable noise environment. The noise in the practical communication scenario usually exhibits impulse characteristics in the statistical sense, which could be modeled as the α-stable distribution. This would make the second- or higher-order statistics of the received signal vanish, and thus the performances of the conventional AMC algorithms designed for Gaussian noise significantly deteriorate when directly employed in the α-stable noise. In our proposed framework, the fractional lower-order cyclic spectrum (FLOCS) analysis is first invoked to acquire the polyspectra of the signal corrupted by the α-stable noise. Then, the graph-based AMC mechanism is systematically established upon the graph representation of the FLOCS to identify the modulation type according to the discrepancies between the graph features derived from the training and test data. The performance of our proposed new algorithm is theoretically analyzed, and the correct classification probability ${P_{cc}}$ over the modulation candidate set is formulated analytically. The remarkable scalability and efficiency of our proposed approach for the modulation candidate set variation are also theoretically investigated in detail. Monte Carlo simulation results demonstrate the effectiveness and superiority of the proposed AMC scheme.

A new feature-extraction paradigm for graph-based automatic modulation classification is proposed in this letter. In the proposed new framework, the modulation features are optimally constructed using the Kullback-Leibler divergence of the dominant entries in the adjacency matrices associated with the graph presentation of the cyclic spectra. Then, the Hamming distance is invoked to measure the discrepancies between the features derived from the training and test data to determine the modulation type. The proposed algorithm, which can select the most distinguishable features, leads to the promising solution to automatic modulation classification (AMC). Compared with the existing AMC approach based on cyclic spectra, Monte Carlo simulation results demonstrate that the proposed AMC method using the new feature-extraction scheme is much more effective.

Unmanned aerial vehicle (UAV) sensing based on the emitted radio-frequency (RF) signals is investigated for the UAV surveillance and control. A novel robust RF-based UAV sensing approach is introduced in this paper. In our proposed new approach, once an RF signal, such as UAV telemetry, track, and command (TT&C) signal and digital data-transmission (DDT) signal, emitted by a UAV is received, the corresponding minimum variance distortionless response (MVDR) spectrum is obtained to extract the feature vector. Then, a novel hybrid lightweight deep-learning network, namely UAV-CTNet, is designed by fusing transformer and convolution to facilitate the feature-map recognition based on the aforementioned feature vector so that the UAV can be detected and identified. Furthermore, Monte Carlo simulations are carried out to evaluate the effectiveness of our proposed new UAV sensing scheme in terms of the detection accuracy and correct recognition rate using the UAV’s TT&C and DDT signals in the presence of interference signals.

An innovative automatic modulation classification via graph-based constellation analysis for M-ary QAM signals is presented. In our framework, a unified mesh model for the constellation diagrams of the M-QAM signals within the modulation candidate set is first constructed, and exploited to transform the received M-QAM signal into graph domain. The concise graph representation of the received M-QAM signal is established from its constellation according to the positions of the recovered symbols in the mesh model. Then, the modulation feature vector is built from the eigenvector(s) corresponding to the maximum eigenvalue of its adjacency matrix. The modulation type can be identified by measuring the angle between the feature vectors resulting from the training data and the test data. Monte Carlo simulation results and theoretical analysis demonstrate that the proposed method with lower computational complexity can provide superior performance to the existing subtractive clustering technique, and is robust to the residual phase and timing offsets.

Based on a multiple layer perceptron neural networks, this paper presents a real-time channel prediction model, which could predict channel parameters such as path loss (PL) and packet drop (PD), for dedicated short-range communications (DSRC). The dataset used for training, validating, and testing was extracted from experiments under several different road scenarios including highways, local areas, residential areas, state parks, and rural areas. The study shows that the proposed PL prediction model outperforms conventional empirical models. Meanwhile, the proposed PD prediction model achieves higher prediction accuracy than the statistical one. Moreover, the prediction model can operate in real-time, through updating its training set, to predict channel parameters. Such a model can be easily extended to the applications of autonomous driving, the Internet of Things (IoT), 5th generation cellular network technology (5G) and many others.

Automatic composite-modulation classification (ACMC) is deemed to be an essential and important cognitive mechanism adopted for the next generation intelligent telemetry, tracking, and command (TT&C), cognitive aerospace communications, and space surveillance as it can automatically recognize the unknown composite-modulation (CM) scheme of the received signal. In this work, we introduce a novel ACMC method using the hybrid feature-vectors based on our proposed new cyclic-paw-print images extracted from the CM signals. In our proposed novel approach, according to the polyspectral analysis of the received CM signals, a received CM signal can be converted to a gray-scale image matrix, named cyclic-paw-print (CPP), which can be very robust against noise. Then, the discrete cosine transform (DCT) and discrete wavelet transform (DWT) are simultaneously applied to the obtained CPP matrix and the hybrid DCT/DWT feature-vector is further constructed thereby. Finally, the sequential minimal optimized support vector machine (SMO-SVM) is adopted as the classifier using such feature vectors. Our proposed new ACMC technique requires a lower computational complexity and leads to a higher recognition-accuracy than other existing ACMC methods according to Monte Carlo simulations and experiments using real CM signal data.