Integrated sensing and communication (ISAC) is an up-and-coming technique for future 6G networks. However, the communication message carried by the detection waveform will face the risk of being eavesdropped, which leads to the challenge of wireless security for ISAC networks. In this paper, we leverage non-orthogonal multiple access (NOMA) to support more users for the ISAC network, with the precoding well designed to guarantee the security. Specifically, we formulate a joint precoding optimization problem to maximize the sum secrecy rate for multiple users via artificial jamming, where the superimposed signal for NOMA users can be concurrently employed for the target detection. Since the optimization problem is non-convex, it is transformed into a convex one based on successive convex approximation (SCA), where the Taylor’s approximation and second-order cone (SOC) constraint are further applied. Then, we propose an iterative algorithm, through which the original optimization problem can be solved effectively. Simulation results show that the proposed secure NOMA-ISAC scheme can guarantee the secure transmission while ensuring the sensing performance.

Automatic modulation classification (AMC) is a core technique in noncooperative communication systems. In particular, feature-based (FB) AMC algorithms have been widely studied. Current FB AMC methods are commonly designed for a limited set of modulation and lack of generalization ability; to tackle this challenge, a robust AMC method using convolutional neural networks (CNN) is proposed in this paper. In total, 15 different modulation types are considered. The proposed method can classify the received signal directly without feature extracion, and it can automatically learn features from the received signals. The features learned by the CNN are presented and analyzed. The robust features of the received signals in a specific SNR range are studied. The accuracy of classification using CNN is shown to be remarkable, particularly for low SNRs. The generalization ability of robust features is also proven to be excellent using the support vector machine (SVM). Finally, to help us better understand the process of feature learning, some outputs of intermediate layers of the CNN are visualized.

In this letter, we consider employing a non-orthogonal multiple access (NOMA) unmanned aerial vehicle (UAV) as relay to ferry data from a remote base station to multiple ground users. Our objective is to minimize the power consumption of UAV by jointly optimizing the location and transmission power of UAV. The formulated problem is non-convex with highly coupled binary discrete constraints, which is intractable to solve. Firstly, it is transformed into a more tractable form, where the logarithmic constraints are transformed into polynomial form, and the discrete constraints are transformed into continuous form. Then the non-convex equality constraints are handled by penalty method, while the non-convex inequality constraints are handled by successive convex approximation technique. Based on these measures, we propose a double-loop iterative algorithm to obtain the solution approximately. Numerical results show the effectiveness of our proposed algorithm.

Convolutional neural networks have an immense impact on computer vision tasks. However, the accuracy of convolutional neural networks on a dataset is tremendously affected when images within the dataset highly vary. Test images of plant leaves are usually taken in situ. These images, apart from the region of interest, contain unwanted parts of plants, soil, rocks, and/or human body parts. Segmentation helps isolate the target region and a deep convolutional neural network classifies images precisely. Therefore, we combined edge and morphological based segmentation, background subtraction, and the convolutional neural network to help improve accuracy on image sets with images containing clean and cluttered backgrounds. In the proposed system, segmentation was applied to first extract leaf images in the foreground. Several images contained a leaf of interest interposed between unfavorable foregrounds and backgrounds. Background subtraction was implemented to remove the foreground image followed by segmentation to obtain the region of interest. Finally, the images were classified by a pre-trained classification network. The experimental results on two, four, and eight classes of datasets show that the proposed method achieves 98.7%, 96.7%, and 93.57% accuracy by fine-tuned DenseNet121, InceptionV3, and DenseNet121 models, respectively, on a clean dataset. For two class datasets, the accuracy obtained was about 12% higher for a dataset with images taken in the homogeneous background compared to that of a dataset with testing images with a cluttered background. Results also suggest that image sets with clean backgrounds tend to start training with higher accuracy and converge faster.