Jul 2026· International Conference on Computer Communications and Networks· pp. 1-7· 0 citations· 14 references
Abstract
In large-scale reconfigurable intelligent surface (RIS)-assisted non-orthogonal multiple access (NOMA) systems, achieving energy-efficient communication while obtaining full channel state information (CSI) is challenging due to prohibitive pilot overhead. The joint CSI estimation and RIS phase shift prediction pose a tightly coupled challenge between performance and energy consumption. In this paper, we propose an energy-efficient and deep learning (DL)–based end-to-end framework that jointly performs CSI estimation and phase shift prediction for a large-scale RIS-assisted NOMA system, leveraging only partial CSI obtained from a small subset of active RIS elements (6%). The proposed framework integrates a DL model termed DSRNetV2 for CSI estimation and a lightweight phase shift prediction network termed PhaseNet. Both models are jointly trained using an energy-efficient dynamic hybrid learning strategy, which first minimizes the estimation error and then maximizes the system sum rate. Simulation results demonstrate that the proposed hybrid learning approach outperforms the fixed-weight strategy by 32% in throughput and 37% in energy efficiency, and exceeds the MSE-only training by 59% in throughput and 82% in energy efficiency, while maintaining the fairness over 99% for the weaker user. These results confirm that the proposed framework achieves accurate channel recovery, efficient phase shift prediction, and high energy efficiency, supporting a sustainable RIS-NOMA design for future green 6G systems.
Accurate channel state information (CSI) is essential for multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) systems, yet fast time-varying channels pose significant prediction challenges. Traditional approaches fail under high mobility, while deep learning methods rely heavily on large labeled datasets, limiting generalization with scarce training data. Although large language models (LLMs) show promise, their massive parameter count hinders deployment on resource-constrained edge devices. This paper proposes a lightweight, end-to-end CSI prediction framework built upon a general LLM. A time-frequency dual-domain feature extraction module captures subcarrier correlations and temporal dynamics from historical CSI, overcoming single-domain limitations. The end-to-end design maps historical CSI directly to future states, avoiding error propagation inherent in explicit channel estimation. Parameter efficiency is achieved through low-rank adaptation (LoRA) combined with knowledge distillation from a pre-trained LLM, enabling effective few-shot learning at low computational cost. Simulations demonstrate that the proposed scheme delivers robust prediction accuracy across diverse mobility scenar-ios, maintains strong performance under limited training data, and exhibits zero-shot cross-scenario generalization, significantly outperforming both conventional and deep learning baselines in TDD and FDD modes.
Keywords: Channel prediction, multiple-input multiple-output (MIMO), orthogonal frequency division multiplexing (OFDM), large language model (LLM), knowledge distillation, low-rank adaptation (LoRA)
This paper considers a downlink communication framework comprising a simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS)-aided by orthogonal time frequency space (OTFS) and non-orthogonal multiple access (NOMA) technologies. Further, delay-Doppler mobility in such frameworks renders classical alternating optimization impractical for per-coherence interval reconfiguration. To mitigate such issues, the STAR-RIS phase-shift and energy-splitting design is formulated as a constrained, non-convex sum-rate maximization problem with closed-form maximum ratio transmission beamforming and fixed NOMA power allocation. To circumvent the per-interval re-optimization burden, a deep reinforcement learning (DRL) approach is adopted that maps observed channel realizations to STAR-RIS configurations through a single forward pass. Specifically, Beta-Space Soft Actor-Critic (SAC-BSE), a maximum entropy DRL agent, is proposed. Simulation results, with two NOMA-multiplexed users on each STAR-RIS branch, confirm rapid convergence, limit the sum-rate degradation to roughly 10\% across a 128-fold user-speed range, and yield consistent gains over OTFS-only, NOMA-only, STAR-RIS-only, fixed-split, and mode-switching baselines as transmit power and the number of STAR-RIS elements increase.
Rais J. Gachaba, Manobendu Sarker, Anirban Bhowal· 0 citations
A reinforcement learning (RL)-driven control framework that operates over a bank of pretrained multi-rate AEs, each corresponding to a distinct compression ratio (CR), aiming to dynamically optimize the trade-off between reconstruction fidelity and signaling overhead.
Maryam Ansarifard, M. Sharma, Georgios Exarchakos et al.· arXiv.org· 0 citations
: Massive multiple-input multiple-output (MIMO) technology is a key enabler for 5G and beyond wireless networks, offering significant improvements in spectral efficiency and link reliability. However, conventional beamforming techniques such as Zero Forcing (ZF) and Minimum Mean Square Error (MMSE) require complex matrix computations and fail to adapt efficiently to dynamic channel variations. To address these challenges, this paper proposes a Deep Deterministic Policy Gradient (DDPG)-based beamforming framework that formulates beamforming optimization as a continuous-action deep reinforcement learning problem. The proposed model directly generates complex-valued beamforming weight vectors to jointly maximize spectral efficiency (SE) and energy efficiency (EE) while minimizing the bit error rate (BER) and decision latency. An adaptive state representation incorporating channel state information (CSI), previous beamforming vectors, and performance metrics enables real-time policy learning under time-varying channel conditions. Simulation results demonstrate that the proposed method outperforms conventional and heuristic beamforming schemes, achieving up to 25% higher SE, 45% lower BER, 20% improvement in EE, and 30% latency reduction. The results validate the effectiveness of the proposed framework for energy-efficient, low-latency beamforming in next-generation massive MIMO wireless networks.
Nilakshee Rajule, Mithra Venkatesan, Harshada Magar et al.· Proceedings of the 1st Inter...· 0 citations
Reliable channel estimation (CE) in unmanned aerial vehicles (UAVs)-assisted orthogonal frequency-division multiplexing (OFDM) systems is fundamentally challenged by mobility-induced Doppler dynamics and frequency-dependent beam squint, which jointly distort pilot observations and reduce channel coherence across subcarriers. These impairments limit the effectiveness of conventional model-based estimators and increase retransmissions, thereby degrading spectral and energy efficiency. This paper develops a structured UAV-assisted OFDM framework that explicitly captures Doppler-induced phase evolution and wideband spatial distortion. Least-squares (LS) estimation and a genie-aided minimum mean square error (MMSE) equalizer are employed as analytical references, where the latter serves as an ideal upper performance bound. Building upon this foundation, we introduce a physics-informed contextual deep learning formulation that refines LS estimates by exploiting cross-subcarrier frequency correlation. The proposed hybrid architecture, termed CRDBA-Net, integrates dilated residual convolution, bidirectional sequential modeling, and attention-based subcarrier weighting to capture multi-scale frequency structure and mobility-driven variation. Extensive bit error rate (BER) evaluations demonstrate that the proposed framework consistently outperforms classical and representative learning-based baselines across a wide signal-to-noise ratio range and under severe Doppler and beam-squint conditions, while approaching genie-aided MMSE performance without requiring prior channel statistics or matrix inversion. The results highlight the potential of structured deep learning to enhance reliability and computational efficiency in high-mobility green UAV communication networks.
Muhammad Usman, I. Hameed, Md Habibur Rahman et al.· IEEE Transactions on Green C...· 0 citations
With advancements in long-distance wireless power transfer (WPT) and space-based energy technologies, the integration of WPT into non-terrestrial networks (NTNs), hereafter referred to as NTN-WPT, is emerging as a promising approach for next-generation wireless networks. This paper proposes an energy-scheduling approach to jointly optimize energy efficiency, task completion rate, and task waiting time for power transfer from low Earth orbit satellites to terrestrial mobile user devices (UDs). To address the significant energy-scheduling challenges arising from satellite and UD mobility and further exacerbated by channel uncertainty due to stochastic propagation effects, we decompose the problem into three subproblems corresponding to a three-layer predictive framework: 1) a state prediction layer forecasts UD and satellite states; 2) an interaction mapping layer, employing a graph neural network (GNN), models the energy transfer efficiency between them; and 3) a decision-making layer determines the optimal energy allocation plan. We employ distinct machine learning (ML) methods within this framework, tailored to the specific requirements of each layer. Furthermore, balancing these competing objectives presents a challenging multi-objective optimization problem (MOP). We address this by adopting a key multi-objective reinforcement learning (MORL) technique: scalarizing the objectives into a single weighted-sum reward function. This scalarization transforms the MOP into a tractable, single-objective problem for the agents to solve. To help the agents balance these competing objectives effectively, we introduce a multi-agent deep learning model that integrates a self-attention mechanism with multi-agent proximal policy optimization (MAPPO). This approach provides a robust and efficient solution for WPT in NTNs, particularly for mission-critical scenarios. Simulation results show that the proposed approach can achieve a better overall trade-off than the baseline methods, maintaining competitive task completion rates and energy efficiency while reducing task waiting times. It also demonstrates robust performance under highly variable conditions.
