This paper investigates robust transmit (TX) beamforming from the satellite to user terminals (UTs), based on statistical channel state information (CSI). The proposed design specifically targets the mitigation of satellite-to-terrestrial interference in spectrum-sharing integrated terrestrial and satellite networks. By leveraging the distribution information of terrestrial UTs, we first establish an interference model from the satellite to terrestrial systems without shared CSI. Based on this, robust TX beamforming schemes are developed under both the interference threshold and the power budget. Two optimization criteria are considered: satellite weighted sum rate maximization and mean square error minimization. The former achieves a superior achievable rate performance through an iterative optimization framework, whereas the latter enables a low-complexity closed-form solution at the expense of reduced rate, with interference constraints satisfied via a bisection method. To avoid complex integral calculations and the dependence on user distribution information in inter-system interference evaluations, we propose a terrestrial base station position-aided approximation method, and the approximation errors are subsequently analyzed. Numerical simulations validate the effectiveness of our proposed schemes.

Integrating narrowband Internet of things (NB-IoT) into low earth orbit (LEO) satellite communications plays a promising role in serving massive devices and extensive coverage. In the receiver, the active user equipments (UEs) detection and timing synchronization in the random access procedure is essential but greatly challenged by the large frequency offset and long delay characterized by the LEO satellite. This paper presents the preamble allocation and receiver design of beam domain random access for NB-IoT integrated LEO satellite communications equipped with large-scale array antenna. Leveraging the inherent sparsity of the LEO satellite beam domain channel, we first design a UE grouping-based preamble allocation scheme to alleviate the collision by channel orthogonality for UEs simultaneously accessing. Building on the designed scheme, we propose a joint Doppler-beam domain active UE detection method that resorts to large frequency offset, which significantly improves the detection rate and obtains the coarse estimation of frequency offset. To achieve accurate timing synchronization, we formulate the timing offset estimation problem based on the maximum likelihood criterion. Following this, a modified and refined phase difference timing offset estimation algorithm is proposed, which employs the complete frequency hopping pattern of the preamble. To reduce the computational complexity, we propose an estimation algorithm possessing high performance with the prior information of estimated frequency offset. In comparison with the conventional method, the simulation demonstrates the superior performance of the proposed active UE detection and timing synchronization algorithms.

This paper aims at designing a joint on-ground and on-board beamforming scheme for multi-gateway multibeam satellite systems. The proposed scheme exploits the degrees of freedom of on-board processing to mitigate feeder link interference and enhance satellite resource flexibility without requiring additional feeder link bandwidth, thereby improving the overall throughput. Based on the practical consideration of not escalating the complexity and real-time computational burden of the satellite payload, the two-timescale design is employed. Specifically, the long-term on-board beamforming network (BFN) is designed based on the statistical channel state information (S-CSI) of all links, while the short-term on-ground precoding is designed based on the instantaneous CSI (I-CSI) of the effective channel with optimized BFN. We formulate an average sum rate maximization problem that incorporates the lossless constraint of the BFN, as well as the average power constraints at the satellite. We propose an algorithm called two-timescale joint on-ground and on-board beamforming (TJGBB) to tackle this mixed-timescale non-convex stochastic optimization problem. Simulation results demonstrate the effectiveness of the proposed algorithm and its advantages over the benchmark schemes.

Multibeam geostationary (GEO) satellites play a pivotal role in contemporary satellite communication by ad-dressing the escalating demands. In this paper, we assume a flexible beam-user mapping scheme, wherein users have the liberty to select a specific beam for access, while power and bandwidth allocation is performed on a per-beam basis. To effectively manage radio resources and cater to user demands, we employ a Time Division Multiple (TDM) technique within each beam. To address the resource management challenge, we propose an iterative algorithm for optimizing time fraction and power. Each subproblem is transformed and solved using the Successive Convex Approximation (SCA) method and branch and bound algorithm. The results demonstrate that the proposed algorithm outperforms fixed user mapping and fixed time fraction algorithms by reducing the average user unmet degree by 4.0% and 9.9%, respectively. The flexibility inherent in user mapping and resource allocation enhances resource utilization and alleviates the pressure on satellite from higher user demands.

An orbit design method for repeat ground track (RGT) satellites aiming at maximizing average access ratio (AAR) of terrestrial-satellite terminals (TSTs) is proposed in this paper. Several TSTs are located in different places, their AAR is deduced and employed as the objective to design the orbit, constraints of the regression error in orbit are also deduced and the optimization problem which the optimal orbit satisfies is obtained. The objective is transformed by using the penalty function and the Particle Swarm Optimization (PSO) is adopted to solve the optimization problem. Compared with previous work, ours has two contributions: one is that the expression of the objective to design the orbit is deduced, the other is that the adopted algorithm to solve the optimization problem has a better performance for outputting the global optimal solution.

The low earth orbit (LEO) satellite constellation network can realize the global seamless coverage, and provide data offload services to remote areas, such as mountain areas. This paper considers a LEO satellite-ground network, in the network each user can choose a basestation equipped with a special antenna through C-band, and then the basestation will transmit the collected data to the LEO satellite through backhaul link over Ka-band, so as to realize the data offloading services for mobile users. We aim to maximize the system energy efficiency within the constraints of meeting the backhaul capacity determined by the LEO satellite backhaul link and the users minimum data rate requirements. The optimization problem is a nonconvex problem, which is solved by an iterative algorithm based on successive convex approximation(SCA). The final simulation results show that the proposed allocation scheme for data offloading transmission has higher system energy efficiency than the traditional access allocation schemes.

A math word problem (MWP) is a coherent narrative which reflects the underlying logic of math equations. Successful MWP generation can automate the writing of mathematics questions. Previous methods mainly generate MWP text based on inflexible pre-defined templates. In this paper, we propose a neural model for generating MWP text from math equations. Firstly, we incorporate a matching model conditioned on the domain knowledge to retrieve a MWP instance which is most consistent with the ground-truth, where the domain is a latent variable extracted with a domain summarizer. Secondly, by constructing a Quantity Cell Graph (QCG) from the retrieved MWP instance and reasoning over it, we improve the model’s comprehension of real-world scenarios and derive a domain-constrained instance sketch to guide the generation. Besides, the QCG also interacts with the equation encoder to enhance the alignment between math tokens (e.g., quantities and variables) and MWP text. Experiments and empirical analysis on educational MWP set show that our model achieves impressive performance in both automatic evaluation metrics and human evaluation metrics.

As the mobile communication developing from 4G to 5G, it is an important technical trend to develop cellular backhaul by using Non-Terrestrial Network (NTN). This paper introduces the requirement of transmission rate and time delay for eNodeB (eNB) and gNodeB (gNB), as well as the ability of transmission rate and time delay which Non-Terrestrial Network can provide. Then, the improvement scheme of UDP, TCP, SCTP and voice service that are affected by Non-Terrestrial Network in cellular backhaul is studied.