With the rapid development of space-terrestrial integrated communication technologies, the satellite and terrestrial cellular communication industry ecosystems are gradually converging to meet the needs of ubiquitous broadband communication. In order to achieve efficient collaboration between satellite and terrestrial networks in terms of coverage, resources and scheduling, and improve the overall network performance, satellite-terrestrial collaborative networking will be an important direction of space-terrestrial integrated network in the future. Aiming at the goal and vision of space-terrestrial integrated network, the development status and technical challenges of space-terrestrial integrated network were explained, and the satellite-terrestrial integrated network architecture and technical system based on satellite-terrestrial collaborative deployment were proposed innovatively. The key technical schemes of satellite-terrestrial network collaboration at beam-level, service-level and planning-level were analyzed deeply, which provided guidance for the following research on satellite-terrestrial network collaboration technology.

Building a space-air-ground-sea integrated network is one of the visions of 6G, and it is also an important means to realize the “intelligent connection of all things”. The key element of the future space-air-ground-sea integrated network is satellite communication, which provides global coverage and wide-area connectivity. The basic architecture and development status of the space-air-ground-sea integrated network were discussed, and the satellite communication technology for the space-air-ground-sea integrated network was introduced. Then, several key technologies of satellite communication in the space-air-ground-sea integration scenario were analyzed, including adaptive air interface, multi-satellite cooperative robust beamforming, multi-dimensional resource on-demand scheduling, networking/network management, and network slicing technology. Finally, the challenges and application prospects of satellite communication were discussed under the development trend of space-air-ground-sea integration in the future.

Facing the emerging novel applications and service demands, space-air-ground-sea integrated network offers advantages such as global coverage, high efficiency, flexibility, elasticity, and reliability. It provides intelligent, collaborative, and efficient ubiquitous information services for diverse service scenarios in many fields, and is an important part of future mobile communication system. Firstly, the domestic and international development status, network architecture, demands, and challenges of space-air-ground-sea integration were analyzed. Then, driven by the demand for carrying massive diverse services, key network technologies, including network slicing, mobile edge computing, and communication-sensing-computing integration, were investigated. Finally, the future development direction of space-air-ground-sea integration was discussed from the perspectives of intelligence, integration, and security.

Satellites have the advantages of complementing terrestrial 5G network in terms of coverage, reliability and flexibility, especially the integration of low earth orbit (LEO) satellite and the terrestrial 5G network has been one of the hot topics in the industry. The enhanced 5G network architecture and related key technologies for satellite-terrestrial integration were primarily introduced. Firstly, the development trends in the field of satellite-terrestrial integration were analyzed and key application scenarios were listed. Secondly, two enhanced 5G network architectures were introduced, which involved transparent relay satellites and renewable satellites. Finally, the related key technologies were analyzed for the business requirements of satellite-terrestrial integrated communication and the brief summary was given, which provided useful reference for the application of 5G satellite-terrestrial integration.

Facing the future vision of ubiquitous network coverage, the deep integration and cooperative networking between satellite and terrestrial mobile communication network is a necessary technical path to build a space-air-ground integrated network, which can support access at anytime and anywhere. Firstly, current situation and development trend of space-air-ground integrated network were introduced. Then, the network architecture of satellite-terrestrial network and the design of space-based simplified core network were discussed, as well as the key technologies of satellite-terrestrial network. Finally, the proposed network architecture was verified by relevant experiments. and the follow-up research plan and outlook for future directions were discussed.

As the rapid deployment of low earth orbit (LEO) mega-constellations and the swift growth of satellite Internet unfolds, distributed satellite systems composed of a large number of LEO satellites become a significant feature of the space-air-ground integrated networks. In satellite networks, low-cost, high-precision time synchronization technology is one of the core technologies.The distributed timing synchronization problem of LEO communication constellations, influenced by the high dynamics and temporally varying topological structures of low orbit satellites was focused on. Considering the communication delay characteristics in the LEO satellite scenario, and addressing the deficiencies in synchronization accuracy and convergence speed of existing methods in dynamic network topologies, a timing synchronization method based on maximum likelihood estimation for LEO satellites was proposed. This method effectively solves the problem of unpredictable random communication delays and asymmetric inter-satellite distances in distributed dynamic networks. The simulation results show that the synchronization error of the time synchronization method based on maximum likelihood estimation is less than 5 ns under the dynamic scenario and topology of LEO satellites.

A signal design method was proposed for communication and navigation integration,utilizing optimized cyclic code shift keying (CCSK) as communication spread spectrum modulation. The positive value of the cross-correlation function of the communication signal aligns with the peak value of the autocorrelation function of the navigation signal to overcome the cross-correlation interference and enhence the communication rate. The frequency bands of communication signal and navigation signal were divided by frequency division multiplexing. Based on orthogonal frequency division multiplexing (OFDM), three sub-carrier distribution schemes of continuous allocation, alternating allocation and random allocation were proposed. Simulation and analysis of the autocorrelation peak performance of integrated signal with optimized CCSK and CCSK modulation were conducted. The results demonstrate that under three subcarrier allocation schemes, the autocorrelation peak performance of the integrated signal employing optimized CCSK modulation surpasses that using only CCSK modulation. And the cross-correlation characteristic of communication signal modulated by optimized CCSK will enhance the autocorrelation peak value of the corresponding navigation signal, and improve the code phase synchronization performance of navigation signal under the integrated signal system.

The security challenges confronting the space-ground integrated network in 6G era were mainly disscussed, and an innovative scheme based on quantum security was proposed. Firstly, the developmental backdrop and evolving trends of the integrated network’s architecture were outlined, highlighting the distinct features and application contexts of both the transparent forwarding and on-board regeneration frameworks. Secondly, the security challenges faced by the space-ground integrated network were analyzed, especially the risks of data transmission security and access authentication. In response to these challenges, the innovative solutions were proposed to ensure the security of the space-ground integrated network based on quantum technology. By leveraging the unconditional security of quantum communication and the advantages of post-quantum cryptographic algorithms, end-to-end secure communication protection for the space-ground integrated network was provided. Finally, the feasibility of using quantum technology was verified to ensure the security of the space-ground integrated network.

The current methods for identifying network traffic are generally designed and tested for specific network environments or datasets, making it difficult to generalize and apply to complex and ever-changing actual network environments. A robust traffic recognition algorithm based on graph neural networks was proposed for achieving accurate traffic recognition in practical network scenarios. Firstly, in response to the current algorithm’s neglect of network environment fluctuations and the decrease in accuracy caused by pattern changes, network flows were clustered and filtered by selecting high-level protocol features to reduce the impact of network bandwidth fluctuations on website access traffic behavior. Secondly, due to the fact that most current algorithms only perform single stream recognition and ignore the interrelationships between flows, the various types of feature information and their correlations of network flows were considered, and spatiotemporal correlation features between network flows were extracted through graph neural networks to fully learn network traffic characteristics. By complementing multiple flows and features, the robustness of the algorithm was improved. Finally, a Transformer model that could capture global data information was used as a classifier to analyze the multi type features of network data flow, achieving robust network traffic recognition. Approximately 1 500 and 1 400 visits to 21 target websites in different network environments were collected as datasets for training and testing, achieving an accuracy of 90.7%. Compared with the latest ProGraph algorithm, the accuracy is improved by 7.3%, and the experimental results verify the effectiveness of the proposed method.

Cooperative spectrum sensing is regarded as the foundation and a key stage of cognitive radio networks. The node allocation strategy during the spectrum detection process was directly determined by the results of joint spectrum sensing. Various methods for allocating cognitive terminals to enhance the efficiency and fairness of spectrum sensing were introduced. Aiming at the perceptual efficiency of different sub-bands, an indicator called the inefficient transmission parameter (ITP) was proposed to evaluate communication performance, and a closed-form expression solution to the perceptual efficiency optimization problem was provided. The designed scenarios included terminal pairs with the same frequencies having different perceptual properties and the same perceptual properties. For the perceived fairness among different sub-bands, two allocation algorithms were proposed: the arcuate allocation algorithm and the class division allocation algorithm. Fairness between sub-bands was measured by evaluating the worst perceived performance in the sub-band. In order to be applicable to actual scenarios, the frequency band property parameter was added to enhance fairness. This parameter was taken into account the priority and anti-interference ability of the main user using different frequency bands. Simulation results show that the proposed strategy significantly improves ITP in cognitive radio networks, especially when sub-band utilization is different, and the proposed arcuate allocation algorithm significantly improves the perceived fairness of the system.

To improve the coding efficiency of rate-distortion optimization quantization (RDOQ) for the third-generation audio video coding standard (AVS3), a parallelized RDOQ algorithm based on single instruction multiple data (SIMD) was proposed. Firstly, in the optimal coefficient decision (OCD) stage, the dependencies within the scan line were retained through optimization. Secondly, in the last non-zero position decision (LNPD) stage, the Zig-Zag scan line was split into various independent lines based on the partitioning strategy. The optimal coefficients on each scan line were allowed to be calculated in parallel by the proposed method. Finally, SIMD instructions were utilized for parallelized acceleration to boost calculation efficiency in the RDOQ process. Experimental results show that the proposed algorithm can achieve a 29.46% coding time reduction with only 0.25% BD-Rate loss for AI configuration.

Person re-identification (re-ID) involves the cross-camera retrieval and matching of target pedestrian images, facilitating pedestrian association in scenarios where biometric features such as faces and fingerprints may prove ineffective. It has become a pivotal technology in intelligent video surveillance systems, playing a crucial role in domains like smart security and smart cities. Traditional re-ID algorithms typically employ either representation learning or metric learning methods. A novel approach was proposed which combined representation learning and metric learning methods based on the multi-task learning machine learning paradigm. By capitalizing on the advantages of both feature representation and distance metric, and concurrently training the model using classification loss and triplet loss, comprehensive training for both feature extraction and similarity measurement was ensured. Experimental results validate the effectiveness of the proposed approach, demonstrating superior performance in re-ID tasks and underscoring the robustness and superior generalization capability.

In recent years, the world has entered a period of vigorous development in intelligent computing. As deep learning models with huge parameters and complex structures, large model training requires fast synchronization of training parameters between multiple cards and servers, which imposes higher requirements on the bandwidth, latency, reliability, scalability and security of datacenter networks. The requirements and related key technologies of intelligent computing networks for large model training were studied, and the standard specifications, academic research, and case practices of intelligent computing networks were analyzed, in order to promote the development of intelligent computing networks.

AI large model is leading the hot ICT(information and communications technology) industry in the next decade. Intelligent computing center network is a communication base to support the distributed training of AI large model, and it is one of the key factors to determine the efficiency of AI clusters. The data volume and the number of parameters of AI large model are expanding continuously, which brings the network of intelligent computing centers serious challenges, and also brings an opportunity for intergenerational innovation of key network technologies. In the process of AI large model training and inferencing, providing high performance and high security transmission of data are the two core requirements of AI business for intelligent computing network. Efficient load balancing, congestion control technologies and network security protocols are the key network technologies. To address the challenge brought by large-scale AI business, global scheduling ethernet (GSE) was proposed as a corresponding solution, and realistic test environment was built to compare the performance of GSE and RoCE. The test results show that GSE significantly improves JCT compared with RoCE network.

With the development of artificial intelligence, significant achievements are made in various fields such as natural language processing and computer vision through the utilization of large-scale pre-trained models,which promotes the construction of intelligent computing centers. Key technologies related to large-scale pre-trained models in intelligent computing networks were studied. The latest standardization progress of intelligent computing network at home and abroad was systematically reviewed. A target architecture for intelligent computing network was proposed, and the principles of key technologies, including remote direct memory access (RDMA), IB, RoCE, and collective communication, were explored. Moreover, the current issues and future development trends of intelligent computing networks were analyzed. This research holds crucial importance in advancing the development of intelligent computing network technology and providing guidance for the establishment of intelligent computing centers.

With the rapid development of artificial intelligence technology, large language models have been widely applied in numerous fields. However, the potential of large language models to generate inaccurate, misleading, or even harmful contents has raised concerns about their reliability. Adopting alignment techniques to ensure the behavior of large language models is consistent with human values has become an urgent issue to address. Recent research progress on alignment techniques for large language models were surveyed. Common methods for collecting instruction data and human preference datasets were introduced, research on supervised tuning and alignment adjustments was summarized, commonly used datasets and methods for model evaluation were discussed, and future research directions were concluded.