Higher-order cyclic cumulants (CCs) have been widely adopted as the features for automatic modulation recognition (AMR). However, the requirement of high-rate sampling has greatly restricted the application of CCs. In order to relieve this sampling burden, we attempt to bring AMR into the framework of compressive sensing. By exploiting the sparsity of CCs, the proposed AMR algorithm is implemented with a significantly small number of nonuniform samples of the observed signal. Simulation results demonstrate the availability and robustness of the proposed approach.

Automatic modulation recognition (AMR) for secondary modulated signals, remains as a bottleneck, since both analog and digital modulation are adopted in one signal. To address this issue, we focus on the original signal spectrum and the 2nd order spectrum of the pre-demodulated signal, which imply modulation features. Based on this idea, a novel statistic for discrimination is proposed and the feature vector is formed. Finally, the support vector machine (SVM) is adopted to implement the classification work.

Automatic modulation recognition (AMR) for some special phase shift keying (PSK) modulations, such as 8PSK versus π/4-differential quadrature reference PSK, or, quadrature PSK versus offset quadrature PSK, remains a challenge for practical applications, due to the same constellation shape. To address this issue, we shift our focus to the differences in coding schemes, which manifest themselves as fourth-order spectrum lines. Thus, the existence, position and amplitude of these lines are utilized as the robust features for AMR work. To alleviate the sampling burden, we also perform the proposed AMR algorithm with compressive measurements in the framework of compressive sensing.

Higher-order cyclic cumulants (CCs) have been widely adopted for automatic modulation recognition (AMR) in cognitive radio. However, the CC-based AMR suffers greatly from the requirement of high-rate sampling. To overcome this limit, we resort to the theory of compressive sensing (CS). By exploiting the sparsity of CCs, recognition features can be extracted from a small amount of compressive measurements via a rough CS reconstruction algorithm. Accordingly, a CS-based AMR scheme is formulated. Simulation results demonstrate the availability and robustness of the proposed approach.

In the field of millimeter wave communication, antenna subset modulation (ASM) is widely known as a low-complex physical layer secure technology. However, the randomly selected antenna subset, may produce severe pedestal secondary lobes. In this letter, an enhanced ASM scheme, simulated annealing based orthogonal matching pursuit (SA-OMP), is proposed. In this idea, the transmit array is initially formulated by SA to implement secure transmission, on the basis of which the receive array can be determined by solving an overdetermined problem with OMP. Hence an equivalent pattern with low sidelobes is derived. Simulation results demonstrate the availability of the proposed approach.

The theory of compressive sensing (CS) enables the recovery of sparse or approximately sparse signal from a significantly small set of measurements. This progress has inspired a new idea that many signal processing problems can be solved based on the framework of CS. In this work, we employ CS in the task of inferring the modulation of a communication signal without full signal recovery which is the usual focus in previous literatures. We show that the differences in the higher order power spectrum between MPSK signals can be used as the features for automatic modulation recognition (AMR). Subsequently, we present the sparsity of these features in the Fourier domain and provide an approach to extract features from nonuniform compressive samples of observed signals. The proposed method is validated by simulation experiments through the support vector machine (SVM).