LEO satellite communication systems experience significant space-to-ground propagation delays, which lead to outdated channel state information (CSI). As a consequence, the CSI available at a given time instance reflects the channel conditions from several time instances earlier, a phenomenon known as channel aging. Therefore, accurate channel prediction is crucial. Existing channel prediction methods often rely on sequential models that predict the CSI at each future time instance based on the CSI of only the previous time instance. However, this approach tends to suffer from error accumulation over time, resulting in degraded prediction accuracy as the prediction horizon extends. To address this challenge, we propose a Transformer-Long Short-Term Memory (T-LSTM) channel predictor that combines the Transformer and LSTM architectures for future channel prediction. The model predicts the CSI at the current time instance by utilizing a sequence of outdated CSI from several instances earlier. Simulation experiments conducted on a practical doubly selective satellite channel model demonstrate that the proposed T-LSTM model effectively mitigates the impact of channel aging. Moreover, it improves the prediction accuracy of satellite communication channels affected by long propagation delays, specifically compared to LSTM and Transformer methods operating individually.

In this paper,..........

Intelligent reflecting surface (IRS) is a promising technology that provides high throughput in future communication systems and is compatible with various communication techniques, such as non-orthogonal multiple-access (NOMA). This paper studies the downlink transmission of IRS-assisted NOMA communication, considering the practical case of imperfect channel state information (CSI). Aiming to maximize the system sum rate, a robust IRS-aided NOMA design is proposed to jointly find the optimal beamforming vector for the access point and the passive reflection matrix for the IRS. This robust design is realised using the penalty dual decomposition (PDD) scheme, and it is shown that the results have a close performance to their upper bound obtained from the corresponding perfect CSI scenario. The presented method is compatible with both continuous and discrete phase shift elements of the IRS. Our findings show that the proposed algorithms, for both continuous and discrete IRS, have low computational complexity compared to other schemes in the literature. Furthermore, we conduct a performance comparison between the IRS-aided NOMA and the IRS-aided orthogonal multiple access (OMA). This comparison shows that robust beamforming techniques are crucial for the system to reap the advantages of IRS-aided NOMA communication in the presence of CSI uncertainty.

This paper examines the uplink transmission of a single-antenna handsheld user to a cluster of satellites, with a focus on utilizing the inter-satellite links to enable cooperative signal detection. Two cases are studied: one with full CSI and the other with partial CSI between satellites. The two cases are compared in terms of capacity, overhead, and bit error rate. Additionally, the impact of channel estimation error is analyzed in both designs, and robust detection techniques are proposed to handle channel uncertainty up to a certain level. The performance of each case is demonstrated, and a comparison is made with conventional satellite communication schemes where only one satellite can connect to a user. The results of our study reveal that the proposed constellation with a total of 3168 satellites in orbit can enable a capacity of 800 Mbits/sec through cooperation of 12 satellites with and occupied bandwidth of 500 MHz. In contrast, conventional satellite communication approaches with the same system parameters yield a significantly lower capacity of less than 150 Mbits/sec for the nearest satellite.

This study investigates the downlink transmission of a multiple-input single-output (MISO) communication system with an intelligent reflecting surface (IRS) under channel uncertainty. The problem of maximizing the sum-rate is formulated by incorporating the channel estimation error footprints, and it is solved through joint optimization of the active precoding at the access point (AP) and passive beamforming at the IRS using the penalty dual decomposition (PDD) algorithm. The PDD algorithm is iterative, has guaranteed convergence, and yields closed-form solutions in each iteration, resulting in low computational complexity. Two cases of continuous and discrete phase shifts at the IRS are considered. Simulation results indicate the effectiveness of the proposed method in handling channel uncertainty for both continuous and discrete phase shifts.

In this paper, we investigate the uplink transmission of a single-antenna handheld user to a cluster of satellites. Taking advantage of the inter-satellite links, the satellites can cooperate which each other to jointly detect the received signal. We examine a scenario in which the satellite cluster lacks access to the instantaneous channel state information (CSI). Thus, using only statistical CSI, we design the joint detection by minimizing the mean square error (MSE). We calculate the ergodic capacity using the properties of the Wishart matrix, and then for low-SNR scenarios we provide a closed-form approximation for it. Our numerical results demonstrate the effectiveness of the detection scheme, along with the proximity of the approximation to the actual ergodic capacity. Considering a mega constellation with 3168 satellites in low earth orbit (LEO), we show that a capacity of more than 10 MB/sec can be achieved even if only the statistical CSI is known to the receiver, and a capacity of up to 38 MB/sec can be achieved with perfect instantaneous CSI.

In this paper, we investigate secure wireless commu-nication in an intelligent reflecting surface (IRS)-assisted system where the IRS is used to secure the communication of one legitimate receiver in presence of an eavesdropper. We assume that the IRS is standalone, i.e. the passive beamforming of the IRS is carried out completely on its own. Thus, we design an IRS with several passive elements and only two RF chains that can obtain a partial channel state information (CSI) among each node and the IRS. The partial CSI is then mapped into full CSI by using the correlation information between the channels of different IRS elements. We develop a deep reinforcement learning (DRL)-based framework using the deep deterministic policy gradient (DDPG) algorithm to obtain the IRS beamforming vector resulting in maximizing the secrecy rate. Numerical results demonstrate the ability of this technique to secure the wireless communication system.

The authors consider downlink transmission in a multi‐cell massive multiple‐input and multiple‐output (MIMO) system and focus on decreasing the feedback overhead caused by cell cooperation in constant envelope precoding (CEP). In the literature, the single‐cell case for CEP with perfect channel state information (CSI) has been studied. Here, considering full cooperation among the cells, they develop CEP for the multi‐cell case. They devise a low overhead centralised construction for CEP which employs limited cooperation among the cells, providing higher system throughput. To achieve the minimum feedback overhead, a distributed realisation of CEP is proposed where each cell locally performs CEP. Furthermore, a new optimisation problem is solved to compensate for the effects of pilot contamination. In addition, to reduce the computational complexity, a relaxed iterative form is presented for the limited cooperation and the distributed scenarios. Numerical results show that in the proposed structures, despite a tremendous decrease in the system overhead, the performance confronts merely an insignificant degradation, <10%, compared to the full‐cooperation case. Moreover, in the presence of imperfect CSI, the performance of the distributed structure whose imperfect CSI is compensated for, approaches the performance of this structure in the perfect CSI scenario.