Satellite communication is a key aspect of future 6G networks, and the impact of artificial intelligence technology utilizing deep learning on satellite communications has garnered significant interest. This paper outlines the current research status of deep learning applications in satellite communication from the perspective of the physical layer, data link layer, and network layer. It also examines the limitations of deep learning in satellite communication applications and anticipates potential research directions for the future.
With the advancement of satellite imaging technology, the interpretation of remote sensing (RS) images has become an important subject. Especially in the object recognition field, accurate acquisition of vessel location and classification information is crucial to develop strategic plans. However, the lack of ship samples in RS images has been hindering the research of ship fine categorization. In this paper, a vessel generation method based on generative adversarial network is proposed to solve the insufficient samples in RS images. Dealing with sample insufficiency by prior global-local segmentation, category image generation and domain translation composition. Experiments on the HRSC2016 dataset show that the generated pseudo-images are highly similar to the real vessel, which verifies the effectiveness of the method. Besides, we constructed a ship dataset containing 10,000 images, which have great significance in vessel classification and localization.
LEO constellation network has become a promising approach for global communications, due to its global coverage, lower latency, and flexible construction. However, because of the limited coverage of a single satellite and the high-speed space-ground relative movement, how to leverage the precise satellite resources to satisfy the user terminals' requests is an important issue. Beam hopping technique is an efficient way to handle this issue by scheduling the onboard resource according to the distribution of terminals' requests. In this paper we propose a cooperative beam hopping for LEO constellation network, where multiple satellites provide service for the user terminals in the multi-coverage areas. The problem is innovatively modeled as a LASSO problem considering its sparsity, and machine learning methods are introduced to optimize it. Compared with the fixed beam system, our proposed scheme achieves better performance in terms of resource utilization and average user satisfaction rate.
LEO constellation networks, with their advantages of high throughput, wide coverage and strong robustness, have become a key solution to meet future demands for massive data communications and global coverage. However, due to the high-speed movement of low orbit satellites relative to the ground, mobility management is one of the most challenging research topics for enabling mobility service in LEO satellite networks. In this paper, we propose an IP tunneling based scheme for handling the relative mobility between space and ground. The LEO satellite network consists of the access network and the transmission network. The access network includes the user terminals, the space base station on satellite, and the core network in gateway. The transmission network includes the space router, inter-satellite link, feeder link, and the ground router. In the IP tunneling scheme, the access network and the transmission network belong to two networks respectively, where general routing algorithm is used to handle the relative mobility between the satellites and ground stations, and the IP tunneling is used to handle the relative mobility between the satellites and user terminals. Compared with the Mobile IPv6 (MIPv6), the IP tunneling scheme achieves better performance in terms of signaling overhead, memory overhead and switch delay. Analysis result shows that the IP tunneling scheme can greatly reduce the demand of on-board resources.
LEO satellite networks are becoming research hot and the utilization of software defined network (SDN) technology brings benefits in flexible network control. Controller placement is important for SDN-enabled networks. In this paper, the impact of delay and reliability on controller placement for SDN-enabled LEO satellite networks is analyzed firstly, following which a reliability-amended model is proposed based on programmable switches distributed on some satellites. The latency-reliability utility function and programmable switch-based enhanced controller placement algorithm are presented. Finally, the effectiveness of the proposed method is proved through simulation and the results show that about 10% reliability of controllers can be improved compared to that without amendment.
