In the context of integrated aerial-space ocean monitoring and intelligent maritime management, ship target recognition based on multisource sensing data collected from Unmanned Aerial Vehicles (UAVs) and satellites plays a crucial role in navigation control, maritime law enforcement, and border surveillance. However, real-world ship recognition tasks face two major challenges. First, multimodal data fusion is difficult due to the heterogeneity and spatiotemporal misalignment between different modalities, such as optical images and electromagnetic radiation signals. Second, ship categories naturally exhibit a severe long-tailed distribution, where head classes dominate the sample population while tail classes remain scarce, significantly degrading overall recognition performance. To address these challenges, this paper proposes a long-tailed ship recognition method oriented toward aerial-space multimodal perception. The proposed method integrates a class-aware boundary optimization strategy and a category-based reweighting mechanism, effectively enhancing the discriminative capability of tail classes and improving the robustness of multimodal fusion. Experimental results demonstrate that the proposed method consistently outperforms existing approaches on representative long-tailed ship recognition tasks, showing strong practicality and generalization capability.

The decision-making and resource allocation of UAV swarms play a crucial role in dynamic, uncertain environments. In such complex scenarios, UAV swarms need to effectively collaborate and communicate in frequently changing interference conditions. However, existing resource allocation methods typically assume complete interference information or are suitable only for static environments, leading to significant performance degradation in the face of external uncertainties and incomplete information. To address these challenges, this paper employs fuzzy set theory to dynamically model the uncertainty of external interference and defuzzify its impact on the available frequency bands during iterative diagnostics. Additionally, a dynamic constrained multi-objective optimization model is developed, and a novel Dynamic Constrained Multi-Objective Evolutionary Algorithm based on Transfer Search (TrS-DCMOEA) is proposed. By integrating transfer learning and dynamic adjustment strategies, the algorithm quickly adapts to environmental changes, ensuring communication performance while maintaining the security of UAV swarm communications. Simulation results show that the proposed algorithm achieves superior decision-making and resource allocation efficiency in most time slots, with TrS-DCMOEA particularly excelling in tracking the Pareto front in dynamic environments.

The Space-Air-Ground-Sea Integrated Networks (SAGSIN) significantly enhance global communication by merging satellite, aviation, terrestrial, and marine networks. Crucial to SAGSIN’s functionality and security is spectrum monitoring using deep learning-based Automatic Modulation Classification (AMC), essential for processing and classifying complex modulation signals. However, these AMC models are susceptible to adversarial attacks. Thus, we introduce the Deep Time-Frequency Denoising Transformation (DTFDT) defense method to mitigate the impact of adversarial attacks. The DTFDT method is comprised of a deep denoising module and a transformation module. The denoising module maps signals into the time-frequency domain, amplifying the differences between benign and adversarial examples, aiding in the elimination of adversarial perturbations. Concurrently, the transformation module develops a learnable network, generating example-specific transformation matrices suited for signal data, which diminishes the effectiveness of attacks. Extensive evaluations on two datasets, RML2016.10a and DMRadio09.real, demonstrate the superior defense capabilities of DTFDT against various attacks.

In the past ten years, many high-quality datasets have been released to support the rapid development of deep learning in the fields of computer vision, voice, and natural language processing. Nowadays, deep learning has become a key research component of the Sixth-Generation wireless systems (6G) with numerous regulatory and defense applications. In order to facilitate the application of deep learning in radio signal recognition, in this work, a large-scale real-world radio signal dataset is created based on a special aeronautical monitoring system - Automatic Dependent Surveillance-Broadcast (ADS-B). This paper makes two main contributions. First, an automatic data collection and labeling system is designed to capture over-the-air ADS-B signals in the open and real-world scenario without human participation. Through data cleaning and sorting, a high-quality dataset of ADS-B signals is created for radio signal recognition. Second, we conduct an in-depth study on the performance of deep learning models using the new dataset, as well as comparison with a recognition benchmark using machine learning and deep learning methods. Finally, we conclude this paper with a discussion of open problems in this area.