Spectrum prediction is a promising technique of improving the spectrum exploitation in cognitive radio networks (CRNs). Accurate spectrum prediction can assist in reducing the energy consumption of spectrum sensing and improving the network throughput, etc. Recently, a significant amount of research efforts has been devoted to this area, especially deep learning (DL) methods, and greatly advanced spectrum prediction abilities. This paper reviews the state-of-the-art developments in DL for spectrum prediction. Specifically, we first summarize the existing spectrum prediction methods and give a taxonomy from different aspects. Second, we formulate long-term, multi-dimensional, and non-ideal spectrum prediction problems to provide theoretical support for spectrum prediction models. Third, we divide the DL models into four categories: basic models, models for the long-term domain, models for the multi-dimensional domain, and models for the non-ideal domain base on application scenarios and offer an in-depth tutorial on these models. Fourth, we comprehensively collect and organize widely used common spectrum data sources in the existing literature and also overview the spectrum data simulation methods to facilitate other researchers. Furthermore, we summarize and analyze the selection of evaluation metrics for spectrum prediction models based on DL. Finally, we summarize the research trends and highlight the critical research challenges.

Cognitive electronic warfare (CEW) has received increasing attention, and it is widely recognized that it will play a significant role. Cognitive jamming decision-making, as one of the critical technologies of CEW, dramatically impacts the global battlefield situation. In this paper, we introduce the A2C algorithm into cognitive jamming decision-making and propose a heuristic improved A2C algorithm. As a fusion algorithm of DQN and Policy Gradient, on the one hand, the Actor-Critic algorithm has the advantages of iterative updating of policies and high efficiency in complex spaces, compared to the DQN algorithm. On the other hand, compared with the Policy Gradient algorithm, it features fast convergence. But, it suffers from high variance, making convergence difficult. First, we establish a cognitive jamming decision-making model to address the above issues. Then, we develop an improved A2C algorithm by introducing a baseline and dueling networks. The baseline reduces the variance of the critic network, while dueling networks further decrease variance, enhancing the convergence of the A2C algorithm. Additionally, the improved A2C algorithm does not rely on prior information, and enhances the adaptive capability of the jammer when interacting with the target radar. We conducted numerical simulations based on the designed cognitive jamming decisionmaking model. The results demonstrated that compared with the four algorithms (DQN, Policy Gradient, Actor-Critic, A2C), the convergence speed of the improved A2C algorithm is improved by 50%, 57.8%, 34.5%, and 13.64%, respectively, and verified the excellent performance of the improved A2C. Finally, we introduce the heuristic reward function and propose the heuristic improved A2C algorithm. Compared with the improved A2C algorithm, the convergence speed of this algorithm is improved by 31.58%. The simulation results prove that the algorithm can greatly improve our advantage in electronic countermeasures

Spectrum prediction is a promising technique to release spectrum resources and plays an essential role in cognitive radio networks and spectrum situation generating. Traditional algorithms normally focus on one-dimensional or predict spectrum values in a slot-by-slot manner and thus cannot fully perceive the spectrum states in complex environments and lack timeliness. In this paper, a deep learning-based prediction method with a simple structure is developed for temporal–spectral and multi-slot spectrum prediction simultaneously. Specifically, we first analyze and construct spectrum data suitable for the model to simultaneously achieve long-term and multi-dimensional spectrum prediction. Then, a hierarchical spectrum prediction system is developed that takes advantage of the advanced Bi-ConvLSTM and the seq2seq framework. The Bi-ConvLSTM captures time–frequency characteristics of spectrum data, and the seq2seq framework is used for long-term spectrum prediction. Furthermore, the attention mechanism is used to address the limitations of the seq2seq framework that compresses all inputs into fixed-length vectors, resulting in information loss. Finally, the experimental results have shown that the proposed model has a significant advantage over the benchmark schemes. Especially, the proposed spectrum prediction model achieves 6.15%, 0.7749, 1.0978, and 0.9628 in MAPE, MAE, RMSE, and $R^2$, respectively, which is better than all the baseline deep learning models.

Radio frequency (RF) fingerprint (RFF) refers to a unique hardware-related feature of RF devices, characterized by non-reproducible physical characteristics. This feature makes RFF suitable for equipment identity authentication in wireless communication, particularly in the Internet of Things (IoT). Therefore, we propose a more generalized and intuitive definition of the RFF, namely the differences between the received and ideal RF signals. The ideal RF signal is defined as having the same input information source, encoding, and modulation schemes as the received signal, but without any influences from hardware devices or wireless transmission channels. Our approach begins by analyzing the effects of transmission channels, such as attenuation, dispersion, multi-path propagation, and noise, treating them as a series of processes. We represent both the received and ideal signals in a designed multi-domain space.We then develop a Siamese-based neural network to extract features from the multi-domain representations of paired signals simultaneously. A difference block is used to eliminate common features and retain the expected generalized RFF (GRFF). Identification is achieved by treating each device as a specific class, utilizing a softmax classifier. Besides, we introduce a two-stage strategy involving nearest neighbor and particle swarm optimization to discriminate the out-of-distributionsignals. Experiments are carried out using a public ADS-B signal dataset, and the trained network achieves a classification accuracy of 99.13% on a test dataset with 7442 samples from 58 classes in total and out-of-distributiondetection accuracy of 95.45% under a class confidence 0.96 on 5560 samples from additional 206 classes.

Energy conservation, cost, and emission reduction are the research topics of most concern today. The aim of this paper is to reduce the cost and carbon emissions and improve the sustainable development of sheep transportation. Under the typical case of the “farmers–middlemen–slaughterhouses” (FMS) supply model, this paper comprehensively analyzed the factors, sources, and types of cost and carbon emissions in the process of sheep transportation, and a quantitative evaluation model was established. The genetic algorithm (GA) was proposed to search for the optimal path of sheep transportation, and then the model solving algorithm was designed based on the basic GA. The results of path optimization indicated that the optimal solution can be obtained effectively when the range of basic parameters of GA was set reasonably. The optimal solution is the optimal path and the shortest distance under the supply mode of FMS, and the route distance of the optimal path is 245.6 km less than that of random path. From the cost distribution, the fuel power cost of the vehicle, labor cost in transportation, and consumables cost account for a large proportion, while the operation and management cost of the vehicle and depreciation cost of the tires account for a small proportion. The total cost of the optimal path is 26.5% lower than that of the random path, and the total carbon emissions are 36.3% lower than that of random path. Path optimization can thus significantly reduce the cost of different types and significantly reduce the proportion of vehicle fuel power cost and consumables cost, but the degree of cost reduction of different types is different. The result of the optimal path is the key to be explored in this study, and it can be used as the best reference for sheep transportation. The quantitative evaluation model established in this paper can systematically measure the cost and carbon emissions generated in the sheep transportation, which can provide theoretical support for practical application.