In this paper, we investigate the anti-jamming channel selection problem for interference mitigation (IM)-based dense wireless networks in dynamic environment, in which the active user set is variable due to their specific traffic demands. We jointly consider the mutual interference among users and external jamming in IM-based dense wireless networks, and propose a generalized maximum protocol interference and jamming (GMPIJ) model to accurately capture the mutual interference and external jamming. Then, the anti-jamming channel selection problem is formulated as an anti-jamming dynamic game, and subsequently it is proved to be an exact potential game, which has at least one pure strategy Nash equilibrium (NE). Based on the stochastic learning theory, a distributed anti-jamming channel selection algorithm (DACSA) is proposed to find the NE solution. Moreover, the simulation results are presented to demonstrate the effectiveness of the proposed DACSA algorithm.

With the advancement of communication technology, there is a higher demand for high-precision and high-generalization channel path loss models as it is fundamental to communication systems. For traditional stochastic and deterministic models, it is difficult to strike a balance between prediction accuracy and generalizability. This paper proposes a novel deep learning-based path loss prediction model using satellite images. In order to efficiently extract environment features from satellite images, residual structure, attention mechanism, and spatial pyramid pooling layer are developed in the network based on expert knowledge. Using a convolutional network activation visualization method, the interpretability of the proposed model is improved. Finally, the proposed model achieves a prediction accuracy with a root mean square error of 5.05 dB, demonstrating an improvement of 3.07 dB over a reference empirical propagation model.

Precise path loss models play a crucial role in design and optimization of wireless communication systems, requiring a careful balance between accuracy and efficiency. Deep learning presents a promising approach for improving both aspects. This paper proposes a novel approach that uses satellite images to construct a comprehensive dataset with rich environmental information. By incorporating environmental features into a convolutional neural network based model, accuracy of path loss prediction is significantly enhanced. Comparisons and validation demonstrate the approach in improving accuracy of path loss prediction, particularly with the assistance of road information.

In this paper, we investigate the anti-jamming problem in heterogeneous wireless networks. Although there are many studies on the anti-jamming defense problem in both power domain and spectrum domain, these two important aspects were addressed separately. In this paper, to cope with the jamming attacks flexibly, we study the anti-jamming defense problem from a multi-domain perspective, which includes both power domain and spectrum domain, and a multi-domain anti-jamming scheme (MDAS) is proposed. To be more specific, a Stackelberg power game is formulated in the power domain to fight against the jamming attacks, and a multi-armed bandit-based channel selection with a channel switching cost and unknown channel availability state information is formulated in the spectrum domain. Besides, we analyze the performance of the formulated Stackelberg power game and derive the optimal power strategy and utility of a legitimate user. In addition, it is proved that the proposed anti-jamming scheme has a logarithmic regret. Finally, extensive simulations are conducted to validate the performance of the proposed MDAS.

In spite of unique advantages like higher recognition accuracy and better generalization capability, automatic modulation classification (AMC)-oriented deep neural networks (ADNNs) are still vulnerable to adversarial examples (AEs). Recent results revealed that an attacker can easily fool ADNNs through adding a small and imperceptible perturbation to the original signal. Among different AE generation methods, universal adversarial perturbation (UAP) has unique characteristics, including input agnostic and shift invariance. However, applying UAP directly to RF signals faces three main challenges, i.e., perturbation neutralization, high perceptibility, and dependency of original signals. In this backdrop, a novel UAP under frequency and data constraints (UAP-FD) attack is put forward for solving these problems in this article. First, an individual perturbation is filtered based on the representation visualization algorithm to counter the neutralization problem in perturbation integration. Second, the high-frequency components in the integrated UAP is eliminated through signal decomposition and reconstruction for promoting the imperceptibility. Third, a proxy signal generation method is proposed to help UAP-FD adapt to data-free black-box settings. A series of experiments is conducted to evaluate the aggressiveness and imperceptibility of UAP-FD attack in different settings on a public data set. Results show that, compared with the existing proposal, UAP-FD has a 40% higher fooling rate, and it can reduce the accuracy of the ADNN model from 83% to 9% while maintaining a good imperceptibility and shift-invariance property. In addition, UAP-FD is applied to real-world captured signals over the transmission channel; and it can reduce the model accuracy from 98.3% to 12.5%.