The advent of modern communication systems has led to the widespread application of deep learning-based automatic modulation recognition (DL-AMR) in wireless communications. However, existing networks still cannot effectively capture the complex relationships between signals under low signal-to-noise (SNR) ratio conditions. This paper proposes an automatic modulation recognition (AMR) method using multi-modal hybrid neural networks, named IQFormer. It is based on I/Q signals and time-frequency (T-F) transform distribution matrix inputs. To capture the inherent connection between spatio-temporal and T-F features in cross-modal features, we design a Dynamic Fusion Embedding (DFE) module. Within this module, feature information from multiple modalities is dynamically aggregated during the embedding stage, resulting in semantically enriched token sequences. Moreover, we develop a staged Transformer block scheme that allows IQFormer to efficiently extract local and global features from embedded tokens at different scales using convolution and attention mechanisms. Experimental results on RadioML2016.10a, RadioML2016.10b and HisarMod2019.1 datasets demonstrate the superior performance of IQFormer compared to the state-of-the-art (SOTA) DL-AMR methods. Code is available at https://github.com/WestdoorSad/IQFormer.

Ocean sound speed is important for underwater acoustic applications, such as communications, navigation and localization, where the assumption of uniformly distributed sound speed profiles (SSPs) is generally used and greatly degrades the performance of underwater acoustic systems. The acquisition of SSPs is necessary for the corrections of the sound ray propagation paths. However, the inversion of SSPs is challenging due to the intricate relations of interrelated physical ocean elements and suffers from the high costs of calculations and hardware deployments. This paper proposes a novel sound speed inversion method based on multi-source ocean remote sensing observations and machine learning, which adapts to large-scale sea regions. Firstly, the datasets of SSPs are generated utilizing the Argo thermohaline profiles and the empirical formulas of the sound speed. Then, the SSPs are analyzed utilizing the empirical orthogonal functions (EOFs) to reduce the dimensions of the feature space as well as the computational load. Considering the nonlinear regression relations of SSPs and the observed datasets, a general framework for sound speed inversion is formulated, which combines the designed machine learning models with the reduced-dimensional feature representations, multi-source ocean remote sensing observations and water temperature data. After being well trained, the proposed machine learning models realize the accurate inversion of the targeted ocean region by inputting the real-time ocean environmental data. The experiments verify the advantages of the proposed method in terms of the accuracy and effectiveness compared with conventional methods.

Specific Emitter Identification (SEI) is vital for maritime traffic safety in the Automatic Identification System (AIS). Current deep learning SEI methods rely heavily on large amounts of annotated data to learn radio frequency fingerprint (RFF) features, which is challenging to obtain under non-cooperative communication conditions and time-consuming to annotate manually. In this paper, we propose a novel momentum-based asymmetric algorithm called the Contrastive and Non-Contrastive Self-Supervised Learning (CoNSSL) method for few-shot SEI. Specifically, we first perform data augmentation on the emitter signals to construct positive and negative samples. Then, we design an asymmetric dual-network architecture, consisting of an online network and a target network, to map the positive and negative sample pairs into the RFF representation spaces of both networks. A contrastive loss function is employed to maximize the similarity between positive pairs and minimize the similarity between negative pairs. Finally, the RFF representations of positive samples obtained by the online network are introduced into another space and compared with the RFF representations of positive samples from the target network for consistency, further enhancing the learning of robust and generalizable RFF features. Experimental results show that CoNSSL effectively learns universal RFF features on a 50-class unlabeled AIS signal dataset and a 5-class universal software radio peripheral (USRP) dataset. In a 10-shot scenario, CoNSSL achieves recognition accuracies of 93.29% and 78.40%, respectively, with a simple linear classifier, outperforming state-of-the-art Self-Supervised Learning (SSL) SEI methods.

Specific emitter identification (SEI) is a non-password authentication method that adds an extra layer of security to wireless devices. However, existing SEI methods are unable to continuously learn new classes from a limited number of training examples due to data scarcity, which is more challenging than the catastrophic forgetting and overfitting problems associated with the widely studied class-incremental learning (CIL). In this paper, we propose a novel few-shot class-incremental specific emitter identification (FSCIL-SEI) framework to address the challenge of catastrophic forgetting and overfitting in CIL. Specifically, to ensure inter-class discriminability during the incremental process, we first employ prototype learning training methods in the base task and introduce a self-supervised contrastive learning (SSCL) that increases inter-class distances and reduces intra-class distances in the feature space. Secondly, we propose a separation of class weights (SCW) to isolate old and new class weights in the classification layer, which effectively mitigates the issue of catastrophic forgetting. Finally, to alleviate the problem of overfitting due to insufficient samples during incremental training, we introduce a three-stage course learning (CL) approach that advances from simple to complex tasks, which not only mitigates overfitting but also improves the generalization ability of the model. Experimental results demonstrate that our method outperforms other FSCIL methods in terms of both plastic degree (PD) and incremental accuracy when evaluated on automatic identification system (AIS) and automatic dependent surveillance-broadcast (ADS-B) datasets.

Automatic Modulation Recognition (AMR) is an essential part of Intelligent Transportation System (ITS) dynamic spectrum allocation. However, current deep learning-based AMR (DL-AMR) methods are challenged to extract discriminative and robust features at low signal-to-noise ratios (SNRs), where the representation of modulation symbols is highly interfered by noise. Furthermore, current research on GNN methods for AMR tasks generally suffers from issues related to graph structure construction and computational complexity. In this paper, we propose a Spatial-Temporal-Frequency Graph Convolution Network (STF-GCN) framework, with the temporal domain as the anchor point, to fuse spatial and frequency domain features embedded in the graph structure nodes. On this basis, an adaptive correlation-based adjacency matrix construction method is proposed, which significantly enhances the graph structure’s capacity to aggregate local information into individual nodes. In addition, a PoolGAT layer is proposed to coarsen and compress the global key features of the graph, significantly reducing the computational complexity. The results of the experiments confirm that STF-GCN is able to achieve recognition performance far beyond the state-of-the-art DL-AMR algorithms, with overall accuracies of 64.35%, 66.04% and 70.95% on the RML2016.10a, RML2016.10b and RML22 datasets, respectively. Furthermore, the average recognition accuracies under low SNR conditions from -14dB to 0dB outperform the state-of-the-art (SOTA) models by 1.20%, 1.95% and 1.83%, respectively. Code is available at https://github.com/WestdoorSad/STF-GCN.

Integrated detection and communication (IDC) is vital for development of underwater integration technology. Existing approaches heavily rely on either communication or detection signals, requiring complex integration structures and demodulation techniques. In IDC systems, the superimposed waveforms offer enhanced flexibility, simplified waveform design and improved spectral efficiency. However, a significant challenge of superimposed waveforms is the interference between the communication and detection signal. To address this problem, a novel superimposed waveform adopting detection interference cancellation scheme based on compressed sensing is investigated in this paper. This scheme effectively eliminates detection signal and obtains residual communication signal. By combining the sparse estimation of channel impulse response with prior knowledge of detection signals, it reconstructs the detection interference at the receiver. Furthermore, we demonstrate the superiority of proposed scheme based on compressing sensing by comparing it with correlation matching estimation. Considering the influence of multi-path effects and ambient noise on residual communication signal, this paper introduces a virtual time reversal mirror (VTRM) for channel equalization and Gaussian kernel soft decision algorithm. The multi-path effects and ambient noise are modeled by the Bellhop and Middleton Class A model, respectively. Simulation results and sea trial show that the effectiveness of the proposed scheme.