This paper investigates hybrid reconfigurable intelligent surfaces (RIS) integrating sub-connected (SC) active partitions with fully-connected (FC) active, SC-active, or passive designs for multi-user multiple-input multiple-output dual-functional radar-communication (DFRC) systems. We formulate weighted DFRC optimization and energy efficiency maximization under DFRC constraints problems to jointly design DFRC waveforms and RIS beamforming. Efficient alternating optimization algorithms with proven convergence are developed for the proposed hybrid architectures. Numerical results demonstrate substantial performance gains: compared to conventional FC-active RIS, the hybrid SC-active/SC-active design achieves about 80% higher energy efficiency while preserving competitive sumrate and radar-detection performance. Relative to purely passive RIS, hybrid SC-active/passive configurations provide approximately 50–70% energy-efficiency gains and 25–35% higher sum-rate over the simulated distance range. Against no-RIS DFRC baselines, even minimal SC-active partitioning, with 25% of the RIS elements assigned to the active subsurface, yields roughly 30–50% sum-rate enhancement, depending on the base station–user distance. The proposed designs exhibit superior hardware efficiency, with SC-active-based architectures requiring substantially fewer RF chains than FC-active-based alternatives while delivering competitive or superior performance. Results validate that hybrid SC-active RIS-based architectures enable scalable, power-efficient DFRC systems suitable for sustainable next-generation wireless networks.
We consider the downlink of a multi-cell massive multi-input-multi-output wireless network with an intelligent reflecting surface (IRS) in each cell. Each base station (BS) serves its users via its IRS by using non-orthogonal multiple access (NOMA). We derive a closed form spectral efficiency (SE) lower bound for our system by considering spatially-correlated Rician channels. This lower bound is then used to maximize the non-concave global energy efficiency (GEE) metric in two steps. In the first step, we optimize the BS power by leveraging the minorization-maximization (MM) framework, which transforms a non-concave optimization into a sequence of surrogate concave problems. We construct a novel surrogate function to develop the MM framework. In the second step, we design a novel low-complexity accelerated gradient projection algorithm to jointly optimize phases of IRS in all the cells. The algorithm accelerates its convergence by using Nesterov extrapolation. We analytically prove the algorithm convergence. We also show that for uncorrelated Rayleigh channels, SE becomes independent of the IRS phase. It is, therefore, crucial to consider spatial correlation while analyzing and developing IRS-based NOMA systems. We numerically investigate multiple aspects which are crucial for realizing tangible SE and GEE gains while designing multi-cell multi-IRS NOMA systems.
Next-generation networks aiming to support ultra-reliable, low-latency, and high-throughput services often operate in the THz band where high absorption losses necessitate dense gNodeB (gNB) deployment. While this enhances coverage and Quality-of-Service (QoS), it also increases the network energy consumption. A commonly explored solution in this context is the gNB sleep scheduling, which, although effective in conserving energy, introduces additional latency, thereby creating an Energy-Delay Dilemma in time-sensitive, energy-aware communication scenarios. To address this, we propose EnRoute, an intelligent framework for unified routing and scheduling in heterogeneous networks, designed to optimize the route energy efficiency while simultaneously meeting the traffic-specific stringent latency and reliability constraints. We first formulate the Time-Sensitive Energy-Efficient Routing Problem (TSEE-RP), and prove its NP-hardness. Subsequently, EnRoute offers a solution that integrates Deep Q-Network (DQN), k-means clustering, and an adaptive fair scheduling policy to tackle the energy-delay dilemma while preserving the performance of throughput-intensive applications. Our performance analysis shows that EnRoute outperforms state-of-the-art methods with up to 98.4% fewer latency violations, 46.5% higher energy efficiency, and 72.8% lower model energy cost, while ensuring Key Performance Index (KPI) compliance across mixed traffic.
The unmanned aerial vehicle (UAV)-assisted Internet-of-Things (IoT) network architecture has emerged as a key technology for supporting communications in emergency scenarios. The quality and freshness of the data collected by UAVs directly impact the effectiveness of emergency decision-making and the overall responsiveness of the system in time-critical situations. However, limited by the battery capacity of UAVs, a fundamental tradeoff exists between information freshness and energy consumption in UAV-assisted IoT networks, which constrains the effective coverage range of emergency operations. Accordingly, this paper explores the inherent trade-off between Age of Information (AoI) and energy consumption in UAV-assisted emergency IoT systems. To reflect the highly dynamic and uncertain channel conditions typical of post-disaster environments, this work further incorporates an imperfect channel state information (CSI) model. Then, a probabilistic constraint AoI-energy tradeoff problem is formulated to jointly optimizing the time allocation of data collection, UAV trajectory, and duration of time slots. To address the resulting non-convex non-linear problem, we first convert the probabilistic constraint problem into a non-probability one, then employ a block coordinate descent method to iteratively solve the highly coupled multi-variable problem. Finally, comprehensive simulation results validate the effectiveness of the proposed method.
In the context of a single-input multiple-output (SIMO) system, this work focuses on the joint three-dimensional (3D) localization and synchronization problem for a user equipment (UE) and a hybrid reconfigurable intelligent surface (HRIS) in a far-field scenario. A multi-stage estimation approach, which consists of constructing a maximum likelihood (ML) framework aimed at accurate channel parameter estimation is proposed. Initially, root-MUSIC algorithm is utilized to ascertain the time-of-arrival (TOA) information from both line-of-sight and non-line-of-sight paths as detected at the base station (BS) and HRIS. These TOAs are used as preliminary estimates and converted in the form of delay steering vectors to be integrated into the channel model, eliminating the delay term in the received signal. Subsequently, the signal received at the HRIS terminal is exploited to formulate an ML estimator. The angle-of-arrival (AOA) at the HRIS is computed through a 2D search. Since the position and state of HRIS are unknown, we need to find extra angular parameters than the traditional model. Hence we sequentially perform a multi-stage estimation for cascade and other angles within the channel. Acknowledging the constraints of a 2D search, the channel parameters derived from the coarse estimation are globally refined to achieve more accurate estimates. The position parameters and clock deviations of the UE, HRIS, as well as the 3D rotation matrix of the HRIS, are deduced from geometric relationships and channel parameters. This multi-stage estimation of channel parameters, combined with the application of geometric relationships to derive position and state parameters, enhances positioning and synchronization accuracy within the SIMO system through global refinement. The proposed methodology significantly improves the system’s capability to accurately estimate the spatial orientation and attain timing synchronization between the HRIS and UE in farfield scenarios. Simulation results demonstrate that the mean square error performance of the estimated channel and position parameters approaches the Cram´er-Rao bound.
Wireless power transfer (WPT) using dedicated radio-frequency sources presents a compelling solution for energizing battery-less communication nodes. However, in indoor environments, energy transfer is limited by propagation losses and multipath fading, leading to significant variability in the received power and charging time. This paper analyzes the statistical distribution of charging time for single-antenna and two-antenna transmission schemes under Rician fading. Transmission strategies that do not require channel state information are investigated to ensure reliable operation over spatially distributed nodes. The results show that transmit diversity significantly improves energy transfer coverage, defined as the area within which multiple nodes can be charged within a target time, while reducing charging-time variability. The proposed analytical models are validated through simulations and experimental measurements conducted in a representative indoor environment, showing good agreement between theoretical and measured results.
With the exponential increase in Internet of Things (IoT) deployments, advanced resource scheduling is needed to address massive connectivity and strict quality of service requirements. In this paper, we propose a hybrid sensitivity and residual subspace refinement (HSRSR) method for protocol-aware coordination between uplink control information reporting and downlink control information scheduling. HSRSR integrates sensitivity driven subspace selection for feedback compression and residual based subspace determination for dynamic scheduling. Sensitivity analysis extracts dominant eigenmodes from channel covariance matrices to preserve mutual information for modulation and coding scheme adaptation. Residual based prioritization adapts resource block group allocations and beamforming by selecting active constraints according to their Lagrange multiplier magnitudes. Compared with codebook based scheduling and heuristic interference coordination based scheduling, HSRSR achieves a sum rate gain of up to 46% under moderate delay constraints with a channel quality indicator quantization granularity of zero point one bits per hertz. Additionally, HSRSR reduces computational complexity from cubic to linear scaling with the number of devices, enabling real-time scheduling in massive IoT networks. Furthermore, it outperforms these baselines while maintaining protocol-compliant resource allocation.
Emerging mobile IoT, augmented/virtual reality (AR/VR) and robotic applications require continuous wireless power delivery together with accurate spatial awareness. Millimeter-wave (mmWave) array beamforming can enhance power focusing and angular resolution, but it usually requires stringent beam control, phase alignment, and feedback signaling. To address these challenges, this paper proposes a resonant beam-based power and direction-sensing system (RB-PDS), which employs retrodirective antenna arrays (RAAs) and closed-loop resonant propagation to establish a self-aligned focused beam between the base station (BS) and an RF-passive user device (UD). The same resonant beam is reused as both the energy-transfer carrier and the direction-sensitive sensing signal, thereby enabling simultaneous wireless power transfer (WPT) and direction-of-arrival (DoA) estimation without independent localization waveforms or active RF transmission from the UD. To improve DoA accuracy in the Fresnel range, a two-step modified multiple signal classification (M-MUSIC) algorithm is developed to compensate for the wavefront curvature of the converged beam. Analytical modeling and system-level simulations show that under the considered short-range operating conditions, RB-PDS can deliver watt-level net DC power while achieving centimeter-level lateral localization accuracy. These results demonstrate the potential of RB-PDS as a low-overhead framework for integrated wireless powering and direction sensing in short-range intelligent systems.
Simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) can reconstruct indirect links when the direct link is blocked for millimeter wave (mmWave) communications. Moreover, the spectral efficiency of non-orthogonal multiple access (NOMA) can benefit from the highly directional transmission of mmWave. To capitalize on the complementarity of STAR-RIS, mmWave, and NOMA technologies, in STAR-RIS-assisted mmWave-NOMA downlink communication systems we propose the spectral efficiency optimization problem, considering the differences of the paired users’ channels. The problem is subject to maximum available transmission power, the constraints of the amplitude and phase-shift coefficients of each STAR-RIS element’s transmission and reflection, and so on. To solve the non-convex problem, the alternating iterative optimization algorithms (AIOAs) based on the successive convex approximation (SCA) and the semidefinite programming (SDP) are proposed, respectively. For the SCA-based AIOA algorithm, the effective channel gain is written as the sum of the square of the variable’ real part and the square of the variable’ imaginary part, which is further transformed into a convex expression after proving that the constraint on the sum of the amplitude coefficients of each STAR-RIS element’s transmission and reflection can be relaxed. For the SDP-based AIOA algorithm, the original problem is transformed into a SDP problem by introducing the auxiliary variables. Moreover, simulation results demonstrate that the two proposed algorithms improve spectral efficiency compared to existing algorithms, and NOMA is superior to orthogonal multiple access (OMA) in terms of spectral efficiency.
This paper proposes a non-terrestrial network (NTN) assisted hybrid Terahertz (THz)–acoustic communication framework for reliable data transmission to underwater autonomous vehicles (AUVs). In the proposed architecture, the NTN node communicates with a floating vessel (FV) using THz carrier, while the FV forwards the information to multiple underwater users via the acoustic channel. To efficiently serve multiple AUVs under limited underwater bandwidth conditions, downlink rate-splitting multiple access (RSMA) technique is employed at the FV, enabling improved spectral efficiency and robust interference management. The FV operates as a decode-and-forward relay, ensuring reliable signal reception from the NTN node and subsequent transmission to underwater users. The THz wireless link connecting NTN node to the FV is modeled using the α–μ fading distribution and accounts various impairments such as free-space-path-loss, molecular absorption, and pointing loss, which are prevalent at THz frequencies. Further, the underwater acoustic link from the FV to the AUVs is modeled using the α–F fading distribution, which captures the harsh and highly variable nature of the underwater acoustic channel. To comprehensively assess system performance under realistic channel conditions, we present novel mathematical expressions for the outage probability, sum ergodic capacity, and throughput. The theoretical plots of the performance metrics have been verified through extensive Monte-Carlo simulations. Further, we have demonstrated the effectiveness of the proposed system compared to power-domain non-orthogonal multiple access under both perfect and imperfect channel state information.
Long Range Wide Area Network (LoRaWAN) enables scalable, long-range Internet of Things (IoT) connectivity, but physical obstructions often reduce coverage and link reliability. Gateway mesh architectures mitigate this limitation by allowing relay gateways to forward packets to border gateways over multi-hop paths. However, existing approaches are largely oblivious to environmental obstructions, which can lead to unstable connectivity, coverage gaps, and insufficient redundancy in critical zones, such as operationally important buildings and key open areas. To address this, we propose an obstruction-aware gateway mesh framework that integrates environmental modeling with gateway placement. The framework constructs a 3D representation of the deployment area from satellite imagery, shadow geometry, and semantic segmentation, and uses a multi-loss propagation model to estimate device-to-gateway and inter-gateway link strength under obstruction-induced attenuation. Using these estimates, the framework determines relay and border gateway placements to improve coverage, ensure reliable relay-to-border gateway connectivity, and provide redundancy in critical zones. Real-world campus evaluations show that the proposed solution achieves higher coverage, reliability, and network efficiency than baseline strategies under a fixed gateway budget in obstruction-prone environments.
IMT-2030/6G aims for ultra-dense connectivity, making always-on, per-message authentication a scalability and energy bottleneck. Digital signatures offer strong authenticity and non-repudiation but impose high signing and verification costs on battery-powered IoT nodes and edge verifiers, while MACs are much cheaper but their security degrades significantly if symmetric keys are compromised. We propose GreenAuth, a risk-adaptive authentication framework that uses MACs by default and escalates to digital signatures for (i) application-defined critical messages (fraction ϕ) and (ii) a bounded post-alert signing window of duration τw triggered by an anomaly detector. We develop a unified model that captures authentication-related compute energy, communication overhead, and edge verification latency using both M/M/c and deterministic-service (M/D/c) queueing abstractions, and validate it with a reproducible discrete-event simulator at scales up to 10,000 devices. Across four hardware profiles—including a post-quantum Dilithium-2 setting—GreenAuth preserves signature protection for all critical traffic while reducing authentication energy to approximately 10% of always-sign, revealing an explicit (ϕ, τw) energy–security trade-off surface that supports operator-driven configuration.
Reliable channel estimation (CE) in unmanned aerial vehicle (UAV)-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.
The growing smart devices (SDs) in the Industrial Internet of Things (IIoT) generate complex computations that strain the performance and energy of local processing. Mobile Edge Computing (MEC) addresses this by providing nearby computing resources for low-latency offloading. However, achieving efficient computation offloading under massive device concurrency and densely distributed computation offloadings remains a key challenge. To address this, this paper constructs a multi-server MEC system model for IIoT and introduces Mean-Field Game (MFG) theory to model the offloading competition among SDs. This effectively reduces the dimensionality and complexity of multi-agent interactions. A novel Mean-Field Computation Offloading (MFCO) algorithm is proposed, which combines MFG with Rainbow Deep Q-Network under a Multi-Agent Deep Reinforcement Learning framework. By incorporating advanced components such as distributional value estimation, prioritized experience replay, multi-step learning, and dueling architecture, each SD acts as an autonomous agent, optimizing its policy based on local observations and mean-field approximations. Further enhancements include Boltzmann exploration, adaptive learning rates, and a mean Q-network structure, which improve convergence speed and training stability. Extensive simulations on a large-scale IIoT platform (100 SDs, 9 MEC servers) demonstrate that MFCO reduces computation latency and improves long-term rewards while maintaining robust server performance.
Wireless sensor networks (WSNs) for sustainable soil monitoring face critical green challenges: stringent energy constraints, dynamic topologies, and the efficiency-lifetime trade-off. This paper proposes Unidirectional spatial compression-Multi-view feature mining-Masked attention-Deep Q-network-Imitation Learning (UMMDI), a low-carbon routing framework that integrates knowledge-driven imitation learning with deep Q-networks for energy-optimal routing. Its core innovations are knowledge-guided feature enhancement with an efficient training strategy, and energy-constrained masked attention pruning. Technically, UMMDI employs a preprocessing-encoder-decoder architecture consisting of three key functional components: Unidirectional Spatial Compression (USC), which reduces spatial redundancy and achieves 16× compression; Multi-view Feature Mining (MVFM), which fuses energy and spatial knowledge for service-aware clustering and enhances feature distinctiveness by 87.9%; and Masked Attention (MA), which prunes the action space by up to 1059 times and drastically lowers decision complexity. This enables self-organizing dynamic tree-based data routing. Visualizations and ablation studies clarify the intrinsic optimization mechanisms. Extensive experiments show that UMMDI outperforms state-of-the-art schemes across all network survivability metrics, extending network lifetime by 13.76%/11.46% and improving energy efficiency by 26.64%/20.99% under normal and 10% node failure scenarios, respectively. Thus, it significantly reduces the operational carbon footprint of soil monitoring WSNs, establishing a robust knowledge-driven deep reinforcement learning paradigm for green and sustainable WSNs.
Rapid channel variations in Low Earth Orbit (LEO) satellite networks require frequent updates of the digital precoder. Conventional iterative hybrid beamforming (HBF) schemes may therefore incur substantial online computational cost, whereas learning-based alternatives may require extensive environment interactions. This paper proposes a graph world model for resource allocation (GWM-RA), a model-based learning framework for energy-efficient, surrogate-assisted non-iterative digital precoding in a two-timescale HBF architecture. Specifically, analog beam steering and scheduler-provided user association are configured on a long timescale using slowly varying statistical channel information, while the short-timescale digital precoder is generated through a single forward pass conditioned on instantaneous channel state information (CSI). Under the considered independent and identically distributed (i.i.d.) block-fading setting, the GWM-RA world model is trained using environment-generated rate and power labels to approximate the immediate action–performance relation conditioned on the current graph state and a candidate digital-precoding action. This differentiable performance surrogate provides one-step estimates of the reward and constraints for policy optimization, thereby reducing the number of true-environment evaluations required under the adopted training protocol. Numerical results show that GWM-RA achieves higher energy efficiency than the considered conventional optimization algorithms and standard model-free reinforcement learning baselines. Under the same training protocol, GWM-RA reaches the target energy-efficiency threshold with over 60% fewer environment interactions than the matched GNN-MFRL counterpart.
Industrial Internet of Things (IIoT) has been an essential technology in industry development, which brings the industry an interconnected ecosystem. However, massive connected sensors, actuators, and industrial devices produce large amount of data to deal with and massive resources to allocate. Traditional communications in IIoT transmit bulk raw data to the edge servers, which increases the burden on the networks, especially when large number of nodes transmit information simultaneously. Semantic Communication (SemCom) transmits semantics instead of raw data, which can increase the transmission efficiency and improve the resource utilization. However, semantic transmission still faces severe conflicts when large number of nodes initialize at the same time. In this article, we propose a priority-aware contention resolution approach for semantic transmission in IIoT. Specifically, considering the applications in industry, we divide the downstream tasks into high-priority and low-priority tasks, and we assume that the nodes are arranged with a Priority IDentification (PID) to recognize corresponding tasks. Based on the PIDs, we design a contention scheduling approach for SemCom in IIoT. Furthermore, we formulate the models of Age of Information (AoI) and minimize the Age of Incorrect Information (AoII) in the system. Then, we estimate AoII and semantic performance to illustrate the impacts of the parameters on the system performance. The experimental results show that the proposed approach is a promising solution to allocate the massive nodes for semantic transmission in IIoT.
Recently, the need for monitoring industrial and agricultural areas has led to the increased adoption of low power wide area network (LPWAN) sensor networks, which utilize a primary group of data transmission technologies. Long Range Wide Area Network (LoRaWAN) networks, which operate using a specific transmission protocol within the LPWAN framework, have gained significant importance in these applications due to their energy efficiency and cost-effectiveness. As monitored areas grow and become more remote, the traditional LoRaWAN protocol faces challenges in scalability and wide-area coverage. This issue is typically managed by adding additional gateways, though this solution increases both infrastructure costs and gateway-to-server connectivity complexity. Recent research has explored multihop schemes to address these limitations; however, these schemes currently support only periodic transmissions, as relay nodes require prior knowledge of transmission times to wake up and forward packets. This reliance on periodic transmissions limits the protocol’s flexibility, making it less responsive to non-periodic or event-driven data needs. This work introduces a novel approach that leverages wake-up radio (WuR), supporting not only periodic but also asynchronous and event-based data retransmissions while maintaining high energy efficiency. Therefore, this work overcomes the limitations of asynchronous and event-driven transmissions without requiring continuously active receivers, achieving high energy efficiency and supporting a theoretically unlimited number of hops. Extensive simulations demonstrate that the network capacity of the proposed LoRa multihop protocol outperforms that of a three-gateway topology , while also achieving higher energy efficiency than an Adaptive Data Rate (ADR)-enabled single-gateway LoRaWAN setup.
Next-generation wireless communication requires ultra-low latency, high data rate, and superior energy efficiency (EE). UAV-aided short-packet visible light communication (VLC) emerges as a promising paradigm to meet these requirements. This paper investigates the joint optimization of UAV trajectory, transmit power, and blocklength to maximize the EE of a UAV-aided short-packet VLC system. We establish the UAV motion and short-packet VLC transmission models, and formulate an EE maximization problem subject to practical constraints. To tackle the resulting fractional and non-convex optimization problem, we adopt the Dinkelbach iterative framework and decompose the problem into three subproblems: the trajectory optimization subproblem, power allocation subproblem, and blocklength optimization subproblem. Each subproblem is transformed into a convex form via cyclic maximization and successive convex approximation. Based on this, we propose a Dinkelbach-based iteration (DBI) algorithm and a low-complexity fixed power (FP) algorithm. Theoretical analysis shows that both algorithms are convergent and computationally efficient. Numerical results demonstrate that the proposed DBI algorithm achieves the best EE performance, while the FP algorithm obtains comparable performance with significantly reduced complexity. Both proposed algorithms consistently outperform existing benchmarks.
