Search

Source number estimation plays an important role in successful blind signal separation. At present, the application of machine learning allows the processing of signals without the time-consuming and complex work of manual feature extraction. However, the convolutional neural network (CNN) for processing complex signals has some problems, such as incomplete feature extraction and high resource consumption. In this paper, a lightweight source number estimation network (LSNEN), which can achieve a robust estimation of the number of mixed complex signals at low SNR (signal-to-noise ratio), is studied. Compared with other estimation methods, which require manual feature extraction, our network can realize the extraction of the depth feature of the original signal data. The convolutional neural network realizes complex mapping of modulated signals through the cascade of multiple three-dimensional convolutional modules. By using a three-dimensional convolution module, the mapping of complex signal convolution is realized. In order to deploy the network in the mobile terminal with limited resources, we further propose a compression method for the network. Firstly, the sparse structure network is obtained by the weight pruning method to accelerate the speed of network reasoning. Then, the weights and activation values of the network are quantified at a fixed point with the method of parameter quantization. Finally, a lightweight network for source number estimation was obtained, which was compressed from 12.92 MB to 3.78 MB with a compression rate of 70.74%, while achieving an accuracy of 94.4%. Compared with other estimation methods, the lightweight source number estimation network method proposed in this paper has higher accuracy, less model space occupation, and can realize the deployment of the mobile terminal.

Unmanned-aerial-vehicles (UAVs)-enabled wireless communication for Internet-of-Things (IoT) applications has attracted increasing attention. This article studies a UAV-assisted data dissemination system, where a rotary-wing UAV is dispatched to disseminate data to terrestrial IoT devices. We target to minimize the completion time via a joint optimization of the UAV trajectory and transmit power, while considering the indispensable constraints which cover the maximum energy budget, speed, transmit power of the UAV, and data requirement for each IoT device. First, we formulate the UAV data dissemination as a completion time minimization problem. To tackle the nonconvex optimization problem, the original problem is transformed into two subproblems: 1) the trajectory optimization and 2) the transmit power optimization, respectively, by introducing auxiliary variables and leveraging the concave–convex procedure. Then, we develop a joint trajectory and transmit power algorithm via tailoring the successive convex approximation and alternating descent method. We further improve the algorithm by maximizing the throughput instead of minimizing the completion time in the transmit power optimization process. The improved algorithm not only reduces the computational complexity but also enhances the achieved performance. In addition, simulation results demonstrate the superior performance of the proposed algorithms under various parameter configurations.

In the field of electronic countermeasures, it is extremely difficult to effectively identify communication behaviors of tactical radio station by analyzing communication protocols of the non-cooperative organization. Therefore, this work mainly explores the possibility of directly figuring out radio station communication behaviors from the perspective of the physical layer, conducts a theoretical analysis with the US military protocol standard as an example, and simulates the communication behaviors signals. Finally, combined with the autoencoder, experiment results show the feasibility of grasping communication behaviors from the perspective of the physical layer.

Signal recognition is one of the significant and challenging tasks in the signal processing and communications field. It is often a common situation that there's no training data accessible for some signal classes to perform a recognition task. Hence, as widely-used in image processing field, zero-shot learning (ZSL) is also very important for signal recognition. Unfortunately, ZSL regarding this field has hardly been studied due to inexplicable signal semantics. This paper proposes a ZSL framework, signal recognition and reconstruction convolutional neural networks (SR2CNN), to address relevant problems in this situation. The key idea behind SR2CNN is to learn the representation of signal semantic feature space by introducing a proper combination of cross entropy loss, center loss and reconstruction loss, as well as adopting a suitable distance metric space such that semantic features have greater minimal inter-class distance than maximal intra-class distance. The proposed SR2CNN can discriminate signals even if no training data is available for some signal class. Moreover, SR2CNN can gradually improve itself in the aid of signal detection, because of constantly refined class center vectors in semantic feature space. These merits are all verified by extensive experiments with ablation studies.

High frequency (HF) communication is widely spread due to some merits like easy deployment and wide communication coverage. Spectrum prediction is a promising technique to facilitate the working frequency selection and enhance the function of automatic link establishment. Most of the existing spectrum prediction algorithms focus on predicting spectrum values in a slot-by-slot manner and therefore are lack of timeliness. Deep learning based spectrum prediction is developed in this paper by simultaneously predicting multi-slot ahead states of multiple spectrum points within a period of time. Specifically, we first employ supervised learning and construct samples depending on long-term and short-term HF spectrum data. Then, advanced residual units are introduced to build multiple residual network modules to respectively capture characteristics in these data with diverse time scales. Further, convolution neural network fuses the outputs of residual network modules above for temporal-spectral prediction, which is combined with residual network modules to construct the deep temporal-spectral residual network. Experiments have demonstrated that the approach proposed in this paper has a significant advantage over the benchmark schemes.

Objective Vascular leiomyomas are rare begin tumor comprising mature vascular smooth muscles that originate in the tunica media of the blood vessels. Most of the tumors arise from the veins. Only a dozen cases of artery-arising vascular leiomyoma have been reported, most of which are presented as small nodules in the hand. Methods Here we report an interesting case of a large artery-arising vascular leiomyoma and perform a systematic review. Results A 55-year-old man complained a 6 × 3 cm firm, mobile, tender and pulseless mass in the medial region of his left thigh. The computed tomography showed a well-demarcated fusiform tumor encircling the superficial femoral artery and was enhanced significantly with contrast. The patency of the superficial femoral artery was intact. The magnetic resonance image exhibited slightly hypointense relative to skeletal muscle on T1-weighted images and a heterogeneous appearance on T2-weighted images. We performed the en bloc resection of the tumor and used a vascular graft to revascularize the artery through end-to-end anastomosis. After histopathological assessments, the tumor was diagnosed as an artery-arising vascular leiomyoma. We also performed a systematic review on artery-arising leiomyomas, discovering 21 cases. Most of the artery-arising vascular leiomyomas were small nodules (mean length: 2.4cm) and most of them were superficial solitary mass located in the hand (13 cases, 62%). Excision of the tumor was an effective treatment. The histological subtype of the artery-arising vascular leiomyoma in all cases was solid type. During the follow-up of each patient, there was no recurrence. Conclusions Artery-arising vascular leiomyomas are extremely rare. Most of them are painless and locate in the hand. Their pathological subtype is solid type in all patients. Due to their begin nature, excision is a cure with little chance of recurrence.

This paper expounds the connotation of cognitive electronic warfare system and development status at home and abroad, analyzes the framework of cognitive electronic warfare system, and introduces three closed-loop processes: cognitive reconnaissance, cognitive countermeasure, and cognitive effectiveness evaluation; finally, to explore the realization approach of cognitive electronic warfare system, proposes a cognitive electronic warfare platform architecture, and focuses on the analysis of the functional requirements of the platform and key technologies.

Spectrum data, which is usually characterized by many dimensions, such as location, frequency, time and signal strength, presents formidable challenges in terms of acquisition, processing and visualization. In practice, a portion of spectrum data entries may be unavailable due to the interference during the acquisition process or the compression during the sensing process. However, the completion work in multi-dimensional spectrum data has drawn few attention. In this paper, we first put forward the concept of spectrum tensor to depict the multi-dimensional data, and then we invoke the recent theocratical advances of tensor completion algorithms to retrieve the incomplete measurements. Moreover, we also build an experimental platform using Universal Software Radio Peripheral (USRP) to collect real-world spectrum tensor data. Experimental results demonstrate that the effectiveness of the proposed spectrum tensor processing scheme is over the state-of-the-art spectrum matrix processing scheme.

Target localization and tracking have the inseparable connection. On the one hand, localization is the basis of tracking. On the other hand, tracking can provide localization with prior information to correct the error of localization to some extent. At the same time, making the monitoring area be the grid area can let the localization and tracking system easy to be applied to different monitoring area. This paper mainly studies on the indoor passive localization and tracking in the grid area where ‘passive’ means there is no communication between sensors and the target. The indoor localization and tracking issue is modeled as a vector Kalman filtering problem. Two schemes are developed for the issue: one is based on the optimized weighted centroid algorithm and the other is based on the fingerprint algorithm. Both simulation results and real-world experimental results are used to demonstrate the effectiveness of the developed schemes.

Due to possessing the multi-direction anisotropic basis functions, directionlet transform can capture the inherent geometrical feature of the image. A directionlet transform -based edge detection approach is proposed in this paper. In the approach, we filter along any two directions with two slopes to capture abundant edge information. In addition, we use the multiscale characteristics of directionlet transform. Combine the advantage of small and large scales to get the desirable edge. The experiments indicate that the proposed approach outperforms wavelet and any other traditional approaches.