We address comprehensive optimization of uplink spectral efficiency (SE) and resource allocation in multiple reconfigurable intelligent surfaces (RIS)-aided cell-free massive MIMO (CF-mMIMO) systems. While integrating CF-mMIMO with RISs enhances SE, existing solutions assume ideal conditions or separately optimize access point (AP) clustering, large-scale fading decoding (LSFD), RIS phase-shifts, and power allocation. To bridge these gaps, we propose a unified statistical channel state information (CSI)-based optimization (SCOP) framework that jointly optimizes AP clustering, LSFD, RIS phase-shift control, and uplink power allocation to maximize the minimum uplink SE. A closed-form SE expression is derived for maximum ratio (MR) combining, accounting for both direct and cascaded channels with spatially correlated Ricean fading. Leveraging statistical CSI significantly reduces real-time acquisition overhead while enabling robust and efficient uplink transmission design. SCOP is solved via a multi-step strategy: (i) for fixed power, an iterative algorithm computes joint optimization parameters (JOP) vector representing a combination of AP clustering, LSFD, and RIS phase-shift parameters; (ii) a closed-form solution updates power allocation; and (iii) an alternating optimization jointly refines both. We also introduce a novel method to extract the optimal system parameters from the JOP. Simulation results show that, in a representative 60 APs, 30 users, and 4 RISs scenario, the proposed SCOP framework lifts the median uplink SE from 2.4 bit/s/Hz to 3.9 bit/s/Hz (+65%) and more than doubles the bottom 5% rate, with similar 60–120% gains in other setups.

This paper proposes a secure wireless system that integrates rate-splitting multiple access (RSMA), intelligent reflecting surfaces (IRS), and full-duplex relaying (FDR) to enhance secrecy performance against multiple colluding eavesdroppers. Closed-form expressions for the secrecy outage probabilities (SOPs) and average secrecy capacities (ASCs) of both common and private messages are derived. Comparative analysis with a baseline RSMA-FDR system (without IRS) demonstrates the benefits of IRS in improving secrecy. Numerical results reveal that RSMA-IRS-FDR achieves superior secrecy performance, with SOPs and ASCs strongly influenced by transmission power, particularly differing between message types. Moreover, the adverse effect of residual self-interference (RSI) is substantially mitigated in the IRS-aided system. The secrecy performance is further enhanced by increasing the number of IRS elements. The study also investigates the impact of key parameters such as power allocation, target secrecy rate, fading order, Wi-Fi frequency, and number of eavesdroppers, offering practical insights for secure RSMA-IRS-FDR system design.

This article proposes a novel QoS multicast routing (QSMR) protocol that leverages cross-layer design for IoT-enabled mobile ad hoc networks (MANETs) within a reconfigurable intelligent surface (RIS)-aided cell-free massive MIMO (CF-mMIMO) environment, especially when multiple active eavesdroppers (Eves) are present. The proposed QSMR protocol employs a cross-layer approach to establish a QoS multicast mesh (QMM) route, considering spectrum efficiency (SE), achievable secrecy rate (ASR), end-to-end (E2E) delay, route stability, and hop count requirements. Initially, we formulate the SE expressions for each IoT device in both uplink and downlink, and the downlink ASR expressions for users under attack. Second, we propose two optimization problems: 1) the phase-shift control (PSC) problem, aimed at maximizing the minimum uplink SE and 2) the power control (PC) problem, focused on maximizing the minimum ASR. Additionally, a deep neural network is designed to reduce the computation time required for solving the PSC and PC problems. Third, leveraging the cross-layer synergy, we propose the QSMR protocol to establish the QMM route from a source to multiple destinations by seamlessly integrating information from both the physical and network layers. Finally, the simulation results evaluate the secure performance and provide comparisons for different scenarios. Moreover, in the presence of Eves, the proposed QSMR protocol outperforms other routing protocols in terms of routing delay, control overhead, and packet delivery ratio.

This article introduces an innovative cell-free massive MIMO (CF-mMIMO) architecture integrating a low-Earth orbit satellite with unmanned aerial vehicles (UAVs) as flying access points (FAPs), referred to as a CF-mMIMO satellite-UAV network, aimed at enhancing uplink spectrum efficiency (SE) for ground users (GUs). Expressions are derived to estimate both direct and cascaded channels involving satellites, UAVs, and GUs, while proposing local combining strategies for the network. A straightforward dynamic clustering framework is developed to mitigate interference among GUs effectively, leveraging the inherent characteristics of UAV-based networks to optimize user service quality. An optimization model for FAP trajectory and energy consumption is proposed to effectively manage energy in limited-resource scenarios and enhance network performance. Additionally, max-min fairness power control issue is explored, with a closed-form solution being presented that ensures the equitable SE distribution among GUs. We also tackle a nonconvex optimization challenge to maximize the network’s sum-rate by refining GU power levels, incorporating large-scale fading decoding (LSFD) to further elevate SE. Extensive simulations validate the effectiveness of our proposed methods, illustrating significant improvements over traditional configurations. These approaches not only advance CF-mMIMO satellite-UAV network performance but also lay foundational strategies for future aerial communication systems.

In this article, we propose an efficient deep-neural-network-based relay selection (DNS) scheme to evaluate and improve the end-to-end throughput in wireless-powered cognitive Internet-of-Things (IoT) networks. In this system, multiple energy harvesting (EH) relays are deployed randomly to assist data transmission from a source node to multiple users under practical nonlinearity of the EH circuits. We first design an incremental relaying protocol, where a selected user will request the help from relays if the direct transmission is not favorable. In such a protocol, we develop a deep neural network framework for relay selection and throughput prediction with high accuracy, less channel feedback amount, and short execution time. Simulation results show that the proposed DNS scheme achieves higher throughput than the conventional relay selection methods, while it considerably reduces computational complexity, suggesting a real-time configuration for IoT systems under complex scenarios. Moreover, the proposed DNS scheme achieves the root-mean-square error (RMSE) of 6.6×10-3on the considered dataset, which exhibits the lowest RMSE as compared to the state-of-the-art machine learning approaches.