Radio broadcasting plays an important role in our daily life. Meanwhile, with the development of wireless communications, the application of software-defined radio platforms gives rise to cheap and easy design of illegal broadcasting stations. These unauthorized broadcasting stations sometimes illegally occupy licensed frequency band, especially associated with amateur radios and unlicensed personal communication devices and services. These unauthorized broadcasting stations may severely interfere the authorized broadcasting and further disrupt the management of spectrum resource in civil applications, such as emergency services and air traffic control. However, it still remains a challenging task to automatically and effectively identify the unauthorized broadcasting in complicated electromagnetic environments. Aiming at developing an intelligent and efficient unauthorized broadcasting identification system, in this paper, a novel identification approach is proposed based on long short-term memory (LSTM) recurrent neural network (RNN) and LabVIEW software. In our approach, firstly, a series of LabVIEW applications are developed to drive USRP 2930s for the acquisition of broadcasting signals. Then, the LSTM identification network is proposed to recognize unauthorized broadcasting. Through the special gate structure inside, the proposed LSTM framework can effectively extract the distinguishing features, such as channel state information and RF device fingerprinting. Simulation results show that the proposed LSTM based approach perform better than other contrastive methods, especially in identification accuracy. Implementation results also demonstrate that the proposed method has an outstanding unauthorized broadcasting identification performance with a high accuracy, i.e., identify the unauthorized broadcasting signals with 99.83% accuracy at the licensed frequency of 107.8 MHz, in realistic electromagnetic environments.

Automatic modulation classification (AMC) plays an important role in many military and civilian communication applications. However, it remains a challenging task to support such AMC mechanisms under impulsive noise environments. Aiming at improving the classification performance in impulsive noise, in this paper, a novel modulation classification method is proposed by using the cyclic correntropy spectrum (CCES). In the proposed method, CCES is introduced into AMC for effectively suppressing impulsive noise. Specifically, it is verified that modulation types can be distinguished through CCES. Then, Multi-slices are extracted at different cycle-frequencies from CCES as the original features for AMC. Following the extraction, the principal component analysis (PCA) is applied to these slices to further optimize the original features. Finally, the radial basis function (RBF) neural network is used as a classifier to perform modulation classification. Monte Carlo simulations demonstrate that the proposed algorithm outperforms other existing schemes in impulsive noise cases, especially with a low generalized signal to noise ratio (GSNR).

Integrating cognitive radio into Internet of Things (IoT) is conducive to reducing spectrum scarcity for large-scale IoT deployment, where a core technology is the design of spectrum access algorithms for effective assignment of spectrum holes. However, due to the partially observable channels and increased number of users in the cognitive radio Internet of Things (CRIoT) network, the secondary users have difficulty avoiding interferences and accessing the spectrum quickly. This study presents a distributed dynamic spectrum access (DSA) algorithm that employs a priority experience replay deep echo state Q-network (PER-DESQN) for CRIoT networks with multiple users and channels. To accelerate the Q-network convergence, we use an echo state network based on the underlying temporal correlation to estimate Q-values. Then, to resolve the Q-value overestimation and improve prediction accuracy, the estimated Q-value and decision action process are trained using a double deep Q-network (DDQN). Moreover, a priority experience replay mechanism that uses the Sum-Tree combined with importance sampling weights is proposed to optimize the DDQN to address the instability of the Q-value resulting from random sampling. As the simulation results demonstrate, the proposed algorithm can make fast and accurate DSA decisions and boost the network channel capacity significantly.

In this paper, we propose a novel approach for performing high-quality edge-preserving image filtering. Based on a local linear model and using the principle of Stein's unbiased risk estimate as an estimator for the mean squared error from the noisy image only, we derive a simple explicit image filter which can filter out noise while preserving edges and fine-scale details. Moreover, this filter has a fast and exact linear-time algorithm whose computational complexity is independent of the filtering kernel size; thus, it can be applied to real time image processing tasks. The experimental results demonstrate the effectiveness of the new filter for various computer vision applications, including noise reduction, detail smoothing and enhancement, high dynamic range compression, and flash/no-flash denoising.

Automatic modulation classification plays an essential role in cognitive communication systems. Traditional automatic modulation classification approaches are primarily developed under Gaussian noise assumptions. Nevertheless, recent empirical studies show that the impulsive noise has emerged in numerous wireless communication systems. The bursty nature of impulsive noise fundamentally challenges the applicability of traditional automatic modulation classification approaches. In order to accurately identify the modulation schemes in impulsive noise environment, in this paper, we propose a novel modulation classification approach through using hyperbolic-tangent cyclic spectrum and multi-branch attention shuffle neural networks. First, based on the designed hyperbolic-tangent autocorrelation function, hyperbolic-tangent cyclic spectrum is proposed to effectively suppress impulsive noise and extract the discriminating features. Then, based on the hyperbolic-tangent cyclic spectrum, a novel deep shuffle neural network is proposed as a classifier to perform the modulation classification through multi-branch attention mechanism to reweight all features. Both numerical and real-data experimental results demonstrate that the proposed algorithm can correctly classify modulation schemes with high accuracy and robustness in impulsive noise.