Sound Event Detection (SED) and Direction of Arrival (DOA) estimation have attracted increasing attention in smart agriculture for tasks such as pest detection, machine-anomaly monitoring, and intruder localization. By integrating SED and DOA, Sound Event Localization and Detection (SELD) provides richer spatio-temporal acoustic information for advanced agricultural applications, such as targeted pest deterrence, fault-source tracing, and event-aware robotic response. However, SELD systems typically rely on large-scale neural networks to handle frame-wise multi-source detection and localization, which significantly limits their deployment on edge devices in real farming environments. To address this challenge, we propose Efficient SELDNet, a lightweight SELD architecture designed for resource-constrained edge platforms. Efficient SELDNet introduces two key modules: (1) an Efficient Multi-Scale Convolution (MSC) module that enhances spatial-spectral feature extraction while reducing computational load, and (2) a Convolutional Conformer module that replaces the standard linear layers in Transformer blocks with MSC modules, thereby substantially reducing the number of parameters. To evaluate the effectiveness and generalization ability of the proposed architecture, extensive experiments are conducted on two public SELD datasets, TNSSE21 and STARSS23, as well as a newly constructed agriculture-oriented synthetic SELD dataset. Experimental results show that Efficient SELDNet achieves comparable accuracy to the state-of-the-art (SOTA) baseline while reducing model size by up to 86% and floating-point operations by 79%. Furthermore, to evaluate cost-effective deployment feasibility in agricultural edge scenarios, we conducted CPU-only inference experiments on both a server-grade CPU and a low-power ARM CPU platform. The results demonstrate that Efficient SELDNet maintains real-time CPU-only inference with a smaller memory footprint, and the large version reduces both latency and memory usage compared with the SOTA baseline on the ARM CPU platform.
The rapid growth of low-altitude aerial services and applications, driven by uncrewed aerial vehicles (UAVs), calls for a new class of digital infrastructure beyond conventional terrestrial networks. The low-altitude wireless network (LAWN) has been proposed as dynamically reconfigurable three-dimensional architectures that integrate aerial and ground nodes to provide connectivity, sensing, and control in open, safety-critical airspace. This tutorial presents a comprehensive treatment of LAWNs from the joint perspectives of artificial intelligence (AI) and signal processing. We first review the historical evolution and architectural foundations of LAWNs, introducing altitude-based layers and functional planes, and summarizing the regulatory and standardization landscape. Building on this system view, we then discuss signal processing fundamentals for LAWNs, including 3D channel and system models, performance metrics, waveform and receiver design, localization and tracking, and multi-functionality co-design. Next, we survey AI techniques for LAWNs, covering discriminative and generative models for perception, control, resource management, and security, as well as emerging paradigms such as foundation models, large language models, and digital twins for mission planning and closed-loop optimization. To illustrate AI-signal processing integration in practice, we provide a case study of an AI-driven multi-tier LAWN with hybrid satellite, high-altitude, and ground nodes. The tutorial concludes by outlining key research challenges in architecture design, signal processing-AI co-design, safety and security, experimentation, and standardization, and by highlighting opportunities for LAWNs to evolve into dependable, AI-native infrastructure for the intelligent skies.
Reliable and energy-efficient trajectory planning is essential for ensuring stable multi-hop communication in Low-Altitude Wireless Networks (LAWNs) operating in dynamic three-dimensional environments. However, jointly optimizing Unmanned Aerial Vehicle (UAV) trajectories and routing paths is computationally intractable, since feasibility changes abruptly with obstacle mobility, link variations, and partial observations. To address the challenges, we propose Hierarchical Adaptive Reasoning and Planning for LAWN (HARP-LAWN), a hierarchical coordination framework that separates long-horizon global reasoning from short-horizon local feasibility control. The global layer uses a Large Language Model (LLM) to semantically interpret coarse network and obstacle information, generating coordination points and routing structures without solving the full optimization problem. The local execution layer refines these coarse plans through event-triggered, in-context replanning based on each UAV's partial observations using onboard LLMs. This enables real-time collision avoidance and link maintenance without the extensive policy retraining required by conventional learning-based methods. Simulation results show that HARP-LAWN provides robust zero-shot adaptability, outperforming reinforcement learning baselines in dynamic environments, and effectively scales to dense UAV deployments without exacerbating communication overhead. The analysis of local execution time and energy consumption further confirms that the hierarchical design enables efficient adaptation, with different LLM models exhibiting distinct reliability-latency-energy trade-offs.
Precise three-dimensional (3D) tracking of unmanned aerial vehicles (UAVs) trajectory is crucial for the emerging application of low-altitude economy. However, conventional tracking approaches that rely on the base station (BS) are fundamentally constrained by the limited spatial resolution of small antenna arrays. In this paper, we establish a new framework supported by extremely large-scale reconfigurable intelligent surface (XL-RIS) and exploit the virtual RIS observation, i.e., the high-dimensional signal impinging on an XL-RIS, as a far richer sensing paradigm for near-field 3D UAV tracking. Due to the inherent passive-reflection property, the high-dimensional signal $\bm {y}_{r}$ of XL-RIS is not directly observable and must be reconstructed from the low-dimensional BS measurement $\bm {y}$, which appears severely ill-posed. We provide a compact information-theoretic analysis showing that the virtual RIS observation contains no less position information than the compressed BS observation, while VOR exploits offline training data as side information and is benchmarked by a Cramér-Rao-bound-based reconstruction limit. Grounded by this theory, we next develop a DenseNet-based Virtual RIS Observation Reconstruction (VOR) algorithm that learns the low-dimensional near-field channel manifold to recover $\bm {y}_{r}$ effectively. Then, the reconstructed observation is processed by the proposed Spatial-Spectral Feature Extractor (SSFE), which comprises a CNN for spatial encoding, a time-frequency analyser, and a near-field AoA extractor leveraging element-wise Fresnel-region phase variation. Finally, the extracted features are passed to the Bi-LSTM Trajectory Predictor (BiTP) for accurate temporal trajectory prediction. Extensive simulations with random-walk, wave, and spiral UAV trajectories show consistent MSE reductions over BS-only baselines under an estimated-AoA protocol with multiple RIS probing profiles. Pilot-aware large-aperture tests show the benefit of pairing larger RIS panels with sufficient probing observations.
This paper proposes an Affine frequency division multiplexing (AFDM)-empowered integrated sensing and communications (ISAC) design, referred to as AFDM-ISAC. We first design a novel AFDM-ISAC frame structure that consists of both ISAC and pure data symbols. Each ISAC symbol consists of a single chirp subcarrier for both sensing and channel estimation, while the remaining subcarriers are allocated for communication. Building upon this structure, we present an analog-domain sensing receiver that down-mixes the received echo with a local chirp to fully exploit chirp compression gains avoiding the need for full-duplex hardware. In addition, a sensing fusion algorithm, guided by AFDM modulation parameters, is further proposed in the digital domain. Leveraging the distinct features of the proposed AFDM-ISAC frame, we present a low-complexity channel estimation scheme for high mobility channels based on a generalized complex exponential basis expansion model (GCE-BEM), along with an optimal power allocation strategy between pilot and data symbols. Moreover, to support frame-based AFDM communications, a GCE-BEM-based Kalman filter is also employed for robust intra-frame channel estimation.
This paper investigates the integrated sensing and communications (ISAC) design for low-altitude wireless networks (LAWNs) using distributed extremely large antenna arrays (ELAAs). The system employs multiple distributed ELAAs to provide cooperative downlink communication with unmanned aerial vehicles (UAVs) operating within predefined air-corridors, while simultaneously performing sensing tasks in a specified monitoring airspace. Unlike conventional far-field models, short ground-to-air distances and extremely large apertures induce near-field spatial non-stationarity with distinct line-of-sight (LoS)/non-line-of-sight (NLoS) conditions. We exploit a channel knowledge map (CKM) to obtain the LoS visibility information. Accordingly, we formulate a joint UAV trajectory and ELAA beamforming optimization problem to maximize the UAVs' long-term average communication rate, subject to target-point sensing power, ELAA power budgets, sensing thresholds, UAV kinematics, collision, and corridor limits. To tackle this highly non-convex problem, we propose a nested multi-agent proximal policy optimization (MAPPO) framework. Specifically, the MAPPO agents explore the optimal UAV trajectories, while the instantaneous reinforcement learning (RL) reward is evaluated by optimizing the cooperative beamforming. For any given UAV positions, we solve the beamforming subproblem via semidefinite relaxation (SDR) and fractional programming (FP), deriving closed-form updates to provide accurate reward signals for the learning environment. Numerical results demonstrate that the proposed design achieves significant improvements in average communication rates over benchmark schemes, highlighting its potential for scalable, environment-adaptive ISAC in distributed ELAA-enabled low-altitude networks.
Proper agricultural practices, and appropriate edaphological and environmental conditions ensure a substantial agri-productions. However, the uncertainty and variability in these factors, often leads to inefficient and low productivity. To address this issue, data-driven intelligence-assisted informed decision making can be considered a prominent tool that would enhance the performance and sustainability in agriculture domain. This work proposes a Predictive Digital Twin that utilizes Edge-Cloud-assisted collaborative intelligence built upon soil and weather data for providing on-time and contextual analytics related to Soil Health Alerts, Soil Type Identification, Weather Forecasting and Crop Recommendations, enabling concerned stakeholders for enhanced agriculture outputs. The work utilizes a layered architecture for Digital Twin that employs Edge-assisted Kernel K-Means Clustering for Nutrient-based Edaphological Variability Analysis and XGBoost for Soil Type Identification, and Cloud-assisted Temporal Fusion Transformers (TFT) for Weather Forecasting and TabNet for Crop Recommendations. The work also proposes an adaptive information-aware transmission strategy for the energy conservation of sensor nodes, required in the agriculture field settings. The experimental evaluation clearly depicts the effectiveness of the proposed Digital Twin approach for Crop Recommendation and Energy Conservation of sensor node.
Active reconfigurable intelligent surface (RIS)-assisted low-altitude unmanned aerial vehicle (UAV) networks have emerged as a promising paradigm for enabling simultaneous wireless information and power transfer (SWIPT) in next-generation wireless systems. However, the joint optimization of communication and energy harvesting (EH) in such systems remains highly challenging due to the coupled temporal and spatial resource allocation, the presence of mixed discrete-continuous control variables, and the inherently non-convex nature of the problem. Existing works typically consider either time-switching or spatial resource allocation independently, thereby limiting the achievable system performance. In this paper, we propose a dual-domain resource allocation framework, referred to as RIS-assisted adaptive hybrid time-switching (RAHTS), which jointly integrates temporal scheduling and RIS element-level spatial allocation within a unified optimization model. To efficiently solve the resulting high-dimensional and constrained optimization problem, we develop a softmax deep double deterministic policy gradient (SD3)-based deep reinforcement learning (DRL) approach under a constrained Markov decision process (CMDP) formulation. The proposed framework jointly optimizes time allocation, RIS element scheduling, transmit power, and amplification coefficients while ensuring system constraints are satisfied. Simulation results demonstrate that the proposed SD3-based RAHTS framework significantly outperforms conventional DRL baselines, such as deep deterministic policy gradient (DDPG), twin delayed deep deterministic policy gradient (TD3), and proximal policy optimization (PPO), in terms of system utility and energy efficiency. These results highlight the effectiveness of the proposed dual-domain optimization strategy for enabling efficient SWIPT in active RIS-assisted UAV networks.
This paper proposes a transmission time interval (TTI)-level video surveillance scheme to support aerial surveillance services in low-altitude wireless networks (LAWNs) with multiple uncrewed aerial vehicles (UAVs). Specifically, a base station (BS) and multiple UAVs perform radio resource management (i.e., MCS selection and RB allocation) and video bitrate adaptation over multiple TTIs. Specifically, a dynamic priority weight is designed to balance RB allocation among UAVs in each TTI, followed by video bitrate adaptation at each UAV, thereby improving overall video quality. Furthermore, we formulate a long-term video quality maximization problem via jointly optimizing RB allocation, MCS selection, and bitrate adaptation. Due to the coupled decision variables and non-convex objective function, we design a learning-based two-layer scheduling algorithm to solve the problem. In the outer layer, a deep reinforcement learning algorithm is adopted to determine the appropriate priority weight in each TTI. In the inner layer, given the determined priority weight, the RB allocation, MCS selection, and bitrate adaptation can be derived in the closed form in each TTI, thereby enabling low-complexity scheduling. Extensive simulation results demonstrate that the proposed algorithm can averagely improve overall video quality by up to 12.02% as compared to commercial off-the-shelf 5G schedulers.
Over-the-air computation (AirComp) has emerged as a promising approach for massive data aggregation, which is yet challenged by the channel variations, task distributions, and inherent energy limitation of the computation nodes. In this paper, we propose an unmanned aerial vehicle (UAV)-assisted Aircomp system to serve multi-cluster computation tasks over time, where the UAV mobility-facilitated spatial and time diversity is exploited for efficient and accurate data computation. Specifically, we aim for the minimization of AirComp aggregation error and the energy consumption by jointly optimizing the transceiver beamforming, normalizing factors, sensor scheduling, and UAV trajectory. To solve the formulated problem, we decompose it into two layers where the inner layer addresses the optimization-based AirComp transceiver design, and the outer layer focuses on the deep reinforcement learning (DRL)-based scheduling and trajectory design. In particular, a pointer network actor-critic learning is developed to tackle the binary scheduling problem, and a soft actor-critic DRL algorithm is employed to determine the UAV trajectory. Simulation results validate the convergence of the proposed hierarchical learning framework and demonstrate its significant performance gains in terms of AirComp aggregation error and energy consumption as compared with baseline schemes.
Near-field user tracking is an important problem in low-altitude wireless networks (LAWNs). Conventional fixed-position antennas (FPAs) lack the deployment flexibility to guarantee robust tracking. When the user is nearly collinear with the array, this geometry may lead to ill-conditioned estimation and degraded Fisher information. Movable antennas (MAs) mitigate this limitation by dynamically repositioning array elements to introduce additional spatial degrees of freedom and enrich the observation geometry. However, this capability raises a new signal processing challenge of jointly optimizing the antenna placements and estimating the user's position. To address this challenge, we first derive the posterior Cramér–Rao lower bound (PCRLB) for near-field tracking and analyze the Fisher information matrix (FIM). This analysis motivates a decomposition of the Bayesian inference problem into two sub-problems: 1) proactively optimizing the placement of MAs to sculpt a favorable log-likelihood landscape; 2) estimating the user's position within the resulting well-conditioned landscape. Based on this insight, we propose a hierarchical tracking algorithm that tackles these sub-problems. First, the sequential convex programming (SCP) is employed to obtain the optimized MA placement, aided by a user position function from the channel map. Then, the offline reinforcement learning (RL) is proposed to distill the MA control policy that maps user position function to near-optimal dynamic MA placements. Finally, leveraging the favorable log-likelihood landscape induced by the learned MA placement policy, we design a two-stage tracking algorithm that achieves both rapid convergence and high estimation accuracy. Simulation results show that the proposed solution substantially outperforms existing benchmarks and closely approaches the PCRLB.
The low-altitude unmanned aerial vehicle (UAV) communication network is crucial for future sixth-generation wireless communications. However, it is susceptible to the severe spectrum scarcity with the expectation of massive user access and high-capacity communications. Moreover, the cognitive UAV air-to-ground payload spectrum decision is susceptible to potential interference issues due to the open-space signal exposure and broadcast nature of the wireless channels. Furthermore, UAV flight control and command heavily depend on the non-payload communication link between the UAV and the remote pilot ground control station for information exchange, which leads to significant decision latency and resource wastage. To tackle these issues, we propose a novel adversarial spectrum embodied intelligence paradigm, within which a spectrum sharing UAV network confronted with multiple intelligent jammers is investigated. Specifically, an adversarial embodied intelligent resource allocation framework is developed to enable autonomous and adaptive resource decision within a closed perception-decision-action loop. Exploiting this framework, two game-theoretic embodied intelligent spectrum allocation and power optimization schemes are designed by integrating multiple embodied agents into the formulated explainable game-theoretic architectures. Specifically, tailored Stackelberg game formulations are derived to capture the competitive interactions between the jamming UAVs and the legitimate users. Theoretically, we prove that multi-agent subgame constitutes an exact potential game that admits a Nash equilibrium, ensuring that the embodied coordination process converges to a stable Stackelberg equilibrium. Based on this, a cross-layer hierarchical training algorithm is designed to solve the coupled optimization problems defined within the Stackelberg game model by exploiting the proposed deep Q network and multi-agent proximal policy optimization based networks. Simulation results show that our proposed schemes are superior to the benchmark schemes in terms of the learning efficiency and utility improvement. Moreover, this extension of embodied intelligence to adversarial spectrum domain shows the distinctive applicability of embodied intelligence in handling non-stationary and resource-constrained wireless environments.
Fluid antennas (FAs) offer dynamic and flexible reconfiguration capabilities, including free positioning on a surface or array, surpassing the limitations of conventional solid-state antennas. Building upon this paradigm, fluid antenna arrays (FAAs) hold significant potential to enhance the performance of wireless communication and radar systems. This paper investigates the design of an FAA-based MIMO radar system aimed at improving target detection in complex interference environments. Specifically, we leverage the free positioning capability of FAA elements by introducing the antenna position vector (APV) as an additional degree of freedom, jointly optimized with transmit waveforms and receive filters. The design problem is formulated as the maximization of the signal-to-interference-plus-noise ratio (SINR), subject to practical constraints such as waveform unimodularity (for power amplifier efficiency) and minimum inter-element spacing (to mitigate mutual coupling). To solve the resulting nonconvex optimization problem, we develop an efficient iterative algorithm based on block majorization-minimization (B-MM). Simulation results demonstrate significant performance gains achieved through the joint optimization of antenna positions and waveforms. Notably, the proposed approach automatically balances angular resolution and sidelobe suppression via optimal antenna placement. Furthermore, the benefits of FAA optimization are particularly evident in resource-constrained scenarios, where optimizing antenna positions alone can yield SINR improvements comparable to those achieved through waveform design, highlighting the strong practical potential of FAAs in next-generation radar systems.
To address the pressing need for efficient multimodal data transmission in low-altitude wireless networks, this paper proposes a Multimodal Semantic Compression and Reconstruction Model (MSCRM). This framework achieves semantic-aware transmission in bandwidth-constrained environments by exchanging only semantic information and model feature memory between the onboard encoding terminal and ground decoding terminal. The architecture employs dual multimodal generative adversarial networks at both the onboard encoding end and ground decoding end. These networks share a specialized Multimodal Feature Fusion Encoding Module (MFFEM) discriminator optimized for low-altitude communication, while their generators can be flexibly adapted to UAV data characteristics. At the onboard encoding terminal, the model enhances the feature discrimination capability of a lightweight MFFEM discriminator through dynamic environmental constraints, enabling real-time conversion of multimodal sensor data into structured semantic labels. The optimized feature memory from the MFFEM network is then fused with semantic labels to form joint information suitable for low-altitude wireless transmission. At the ground decoding terminal, the pre-trained MFFEM network serves as a fixed discriminator, working with received semantic labels to maintain semantic integrity and task usability under limited bandwidth conditions. Comprehensive experiments on multimodal datasets comprising remote-sensing and natural images demonstrate that the proposed model maintains SSIM $\geq 0.8$ and FID $\leq 80$ at bit-rates below 0.3 bpp, markedly surpassing current benchmarks in both compression efficiency and reconstruction fidelity.
The rapid development of the low-altitude economy (LAE) has created growing demand for reliable aerial communication systems. Extremely large-scale multiple-input multiple-output (XL-MIMO) is a promising enabler for such systems due to its high spatial resolution and robust connectivity. However, three-dimensional (3D) mobility together with near-field propagation makes it difficult to obtain dedicated high-fidelity wireless datasets, hindering systematic algorithm development and evaluation. To address this issue, we develop LAETwin-XL, a digital twin (DT)-based toolchain and dataset for XL-MIMO research in LAE scenarios. Built on the Sionna ray-tracing (RT) module, the proposed toolchain simulates near-field and far-field channels with diverse wireless labels for practical environments. Building on this dataset, we further develop a conditional denoising diffusion implicit model (CDDIM)-based generative foundation model that is pretrained to learn transferable XL-MIMO channel representations from incomplete channel observations. Unlike conventional task-specific or foundation models that rely on relatively complete channel inputs, the proposed model can generatively infer informative channel representations from partially observed channels. Experimental results demonstrate that the proposed framework achieves effective zero-shot channel extrapolation performance. Furthermore, using lightweight task heads and limited training data, it enables parameter-efficient transfer to various downstream tasks (e.g., channel estimation, classification, and localization), delivering high accuracy and robustness even under sparse antenna observations. The codes and dataset are available at https://github.com/Lmyxxn/LAETwin-XL.
Integrated Sensing and Communication (ISAC), as a key enabling technology for 6G sensing networks, offers significant potential for enhancing Uncrewed Aerial Vehicle (UAV) monitoring capabilities. However, existing ISAC-based sensing methods for low-altitude UAVs focus primarily on localization and tracking, lacking high-resolution imaging capabilities. Although Inverse Synthetic Aperture Radar (ISAR) enables 2D imaging of moving targets, phase shifts induced by UAV maneuverability degrade image quality. To address this, we propose a Content Adaptive Iterative Shrinkage Thresholding Algorithm Network (CA-ISTANet) for UAV imaging using ISAC signals. By unfolding the ISTA into a deep neural network and reconstructing images solely from signal amplitude, we effectively mitigate phase-induced blurring. However, recovering a sparse image from only amplitude measurements poses a highly non-convex optimization problem, leading to unstable convergence and poor performance with conventional iterative solvers. To address this, we further design a Gradient Update Network (GUN) for adaptive step size generation and a Proximal Mapping Network (PMN) for nonlinear sparse mapping, enhancing convergence stability and adaptability to diverse SNRs, target velocities, and signal sparsity without manual tuning or retraining. Extensive simulations demonstrate that CA-ISTANet achieves superior imaging quality, with a root mean square error (RMSE) of 0.032 and a correlation coefficient (CC) of 0.765, outperforming state-of-the-art methods. Validation on a hardware testbed with real UAV echoes further confirms its practical efficacy, producing clean, well-focused images with clear target contours and effectively suppressed background clutter.
This paper explores the utilization of a rotary-wing UAV equipped with a vertically deployed uniform linear array (ULA) antenna as an airborne base station, enabling simultaneous downlink data transmission to ground communication users and sensing of the potential ground target. The goal is to jointly optimize the UAV trajectory and beamforming schemes to achieve minimal total energy consumption, under constraints related to communication performance, sensing accuracy, transmission power, and mobility. Due to the presence of an array antenna on the UAV, the trajectory design variables are embedded within the exponential terms of the array steering vectory, leading to strong coupling with the beamforming parameters and causing significant non-convexity in the optimization problem. To simplify the solution process, the original formulation is divided into two tractable subproblems: 1) the UAV transmit beamforming design; 2) the UAV trajectory design. A combination of mathematical tools, including semidefinite relaxation (SDR) and successive convex approximation (SCA), is employed within an alternating optimization framework to derive a computationally efficient suboptimal solution. Specifically, SDR is utilized to recast the beamforming subproblem as a semidefinite program, and SCA is further applied to convexify the resulting formulation. For the trajectory subproblem, a sequence of variable transformations is performed to decouple the trajectory variables from the exponential terms of the array steering vectory, followed by the application of SCA to obtain a convex approximation. The effectiveness of the proposed method is substantiated by simulation experiments, which demonstrate both rapid convergence and significant power saving improvements over three benchmark schemes.
The low-altitude wireless networks (LAWN) are envisioned to drive substantial economic benefits by offering integrated sensing, communication, computing and control services. However, the existing terrestrial cellular network cannot provide reliable LAWN coverage for unmanned aerial vehicles (UAVs). The near-space airships present unique advantages in addressing the low-altitude coverage problem and enabling low-latency transmission. Nevertheless, designing communication schemes that support massive connectivity for numerous UAVs and diverse services with heterogeneous requirements in airship-borne LAWNs remains a key challenge. Hence, in this paper, we design a grant-free non-orthogonal multiple access (GF-NOMA) scheme for airship-borne massive multiple-input-multiple-output (MIMO) LAWN. Specifically, we propose a codebook-based non-coherent GF-NOMA scheme with flexible time-frequency resource mapping strategy. The proposed scheme has two operation modes: grant-free unsourced connectivity (GFUC) and grant-free sourced connectivity (GFSC) respectively tailored for data-centric and identity-centric services in the LAWN. Next, we model data detection problems in GFUC and GFSC as a compressive sensing (CS) problem. Furthermore, we design an iterative detection algorithm named cross-domain orthogonal approximate message passing for generalized multiple measurement vector (CD-OAMP-GMMV), where the strong sparsity of wideband massive MIMO channel in the angular-delay domain is exploited. Additionally, we develop a deep unfolding algorithm named cross-domain orthogonal approximate message passing with fixed point network (CD-OAMP-FPN) for reduced computational complexity and enhanced channel estimation performance. Finally, our simulation results show the superiority of the proposed scheme over state-of-the-art GF-NOMA scheme in the airship-borne LAWN. Particularly, the proposed CD-OAMP-GMMV iterative algorithm has higher generalization ability in the GFUC case, while the proposed deep unfolding CD-OAMP-FPN has faster convergence speed and better channel estimation performance in the GFSC case.
Integrated sensing and communication (ISAC) is envisioned as a promising technology to empower the emerging low-altitude economy (LAE). However, current networks primarily serve terrestrial users, resulting in insufficient low-altitude coverage. Furthermore, standalone ISAC systems exhibit restricted sensing ranges, as fixed antenna orientations suffer from significant gain loss at large angles. In view of this, we propose cooperative ISAC enabled by rotatable antenna (RA) technology. RA aligns its boresight with target directions through antenna rotation, effectively mitigating gain loss at large angles. This compact and low-complexity design provides robust coverage and high-precision sensing, making it well-suited for deployment in the LAE. Specifically, multiple RA-equipped base stations (BSs) employ coordinated beamforming to provide robust connectivity for authorized unmanned aerial vehicles (UAVs), while concurrently monitoring distributed critical locations within the target airspace for potential intrusions. To implement this scheme, we maximize the system sum-rate subject to various practical constraints by jointly optimizing BS-UAV associations, transmit beamforming, RA orientation, and UAV trajectories. This nonconvex problem is addressed by an efficient alternating optimization (AO) algorithm that integrates semidefinite relaxation (SDR) and successive convex approximation (SCA). Finally, numerical results demonstrate that the proposed RA-enhanced ISAC system outperforms baselines in line-of-sight (LoS)-dominated, large-angle low-altitude scenarios.
Orthogonal time frequency space (OTFS) modulation has emerged as a transformative technology for enabling reliable wireless communications in high-mobility scenarios (such as low-altitude wireless networks) by leveraging signal processing in the delay-Doppler (DD) domain. Furthermore, integrating OTFS with multiple-input multiple-output (MIMO) technology unlocks spatial flexibility, significantly enhancing system performance. This paper introduces the novel precoding design for multi-user MIMO-OTFS systems, aimed at maximizing the weighted sum rate (WSR) while addressing critical computational complexity challenges. We first develop a comprehensive transceiver framework for joint delay-Doppler-spatial (DDS) domain multiplexing and formulate the WSR maximization problem. Next, we systematically explore various precoding algorithms, starting with a weighted minimum mean-square error (WMMSE)-based benchmark that reveals significant computational bottlenecks in practical implementations. Then, we propose an innovative low-complexity singular value decomposition (SVD)-based algorithm to effectively mitigate inter-user interference. To reduce the computational complexity associated with performing SVD on high-dimensional channel state information (CSI) matrices, we focus on an approximated dominant line-of-sight (LoS) channel. The SVD of the LoS channel can be derived analytically by leveraging the structural properties of the DD and spatial domains, thereby significantly lowering the computational burden. Extensive simulations demonstrate that the proposed low-complexity SVD-based algorithm not only achieves competitive WSR performance but also reduces computational overhead by orders of magnitude compared to conventional methods. Moreover, it requires only partial (LoS-only) CSI, making it highly suitable for practical deployment in next-generation high-mobility wireless systems.
