The mining of ambiguous weather attributes carried by atmospheric electric fields (AEFs), especially, the accurate estimation of thunderstorm occurrence times, is significantly challenging. In this paper, a method to capture thunderstorm time windows is proposed for the first time, based on fuzzy C-means (FCM) clustering. Information granules, i.e., fuzzy sets characterizing thunderstorms as dominant weather, are constructed by processing the time series AEF signal (AEFS) with first-order difference, statistics, and time attributes as data to be clustered by the FCM algorithm. Combined with labeled AEF cases, the optimal information granularity among multiple granularities is determined. Then, alpha-cut is applied to target granules to capture the temporal start and end points of granule elements representing thunderstorm dominance obtained by this cut. These time points correspond to a possible thunderstorm time window and the corresponding AEFS component. A first-step evaluation of the window's rationality is performed based on expert knowledge and weather variation rules. Only components passing the first-step evaluation and subsequently evaluated as thunderstorm-dominated by the designed expert experience fusion-based method are finally evaluated to be utilized for imaging thunderstorm cloud point charge moving paths in the spatial domain. Results show that the proposed method accurately captures the time window and reliably images thunderstorm moving paths in multiple cases. This interpretable dynamic-capture approach has significant advantages over those equal-time AEFS-division-based weather evaluations, facilitating real-time and high-efficiency thunderstorm monitoring.

To address the challenges of user mobility, task dependency, and restricted aerial trajectory planning in maritime mobile edge computing, a user mobility-aware air-sea collaborative task offloading framework was proposed. A three-tier computing architecture comprising user equipment, aerial auxiliary nodes (AAN), and the edge layer was established, formulating a multi-objective optimization model that minimizes system costs (latency and energy consumption) while considering task dependencies, resource allocation constraints, and AAN 3D spatial safety requirements. A dynamic K-means-based user clustering mechanism that periodically updates mobile device (MD)-AAN associations according to real-time mobility patterns was proposed; a twin-delayed deep deterministic policy gradient based heterogeneous agent offloading (TD3-HAO) algorithm was designed,enabling joint optimization of dependent sub-task offloading, resource allocation, and 3D AAN trajectory planning through heterogeneous multi-agent coordination. Simulation results demonstrate superior performance over baseline methods,such as the local algorithm and the deep deterministic policy gradient (DDPG) algorithm,maintaining less than 3% deviation from optimal solutions when scaling to 25 MDs, with 16.94%~38.34% latency reduction. The proposed solution effectively resolves the resource underutilization caused by ignoring task dependencies and agent homogeneity in existing approaches, providing theoretical guidance for air-sea integrated edge computing systems.

The rapid advancement of Autonomous and Intelligent Transportation Systems (A-ITS) has driven the widespread adoption of vehicular communication technologies to enhance the safety, efficiency, and sustainability of modern transportation networks. Key communication technologies, including Vehicle-to-Vehicle (V2V), Vehicle-to-Infrastructure (V2I), and Vehicle-to-Everything (V2X), are fundamental for the operation of A-ITS, enabling seamless interaction among vehicles, infrastructure, and other road users. This survey provides a comprehensive analysis of these critical communication technologies and explores the integration of edge computing and Artificial Intelligence (AI) to optimize A-ITS performance. Furthermore, we identify and discuss the major challenges facing A-ITS, including network reliability, data privacy, security, and interoperability, which must be addressed for successful deployment. We also examine the role of emerging 5G/6G networks in smart cities and the Internet of Things (IoT), focusing on their potential to advance autonomous transportation systems. This work serves as a valuable resource for understanding the current state and future prospects of vehicular communications in A-ITS.

In this letter, we introduce HPGS-SLAM, a real-time RGB-D SLAM system guided by hybrid point features (combining traditional and learned point features), enabling high-precision tracking and online dense mapping with photorealistic reconstruction. HPGS-SLAM consists of two main components: (1) a lightweight feature-based frontend guided by hybrid points with adaptive learnable feature matching, aiming for accurate pose tracking and 3D landmarks generation; and (2) a backend that leverages 3D Gaussian Splatting for real-time dense mapping and photorealistic rendering, where the spawning of Gaussian primitives is guided by the 3D landmarks and hybrid keypoints shared from the frontend. HPGS-SLAM is designed in a distributed architecture to facilitate practical deployment. We evaluate HPGS-SLAM on the Replica, TUM-RGBD, and EuRoC MAV datasets. Both quantitative and qualitative results demonstrate that HPGS-SLAM outperforms existing systems in tracking accuracy and mapping efficiency, while achieving competitive visual quality for visual rendering.

In this paper, a large-scale field trial campaign is conducted for the passive reconfigurable intelligent surfaces (RIS) in Chongqing, China, a mountainous megacity characterized by complex terrain, to validate scenario-adaptive coverage enhancement in the existing 5G commercial networks. Operating RIS at 2.6 GHz (512 elements) and 4.9 GHz (1,024 elements), the study systematically evaluates eight real-world scenarios, including mountain tunnels, river crossing bridges, riverside roads, iconic landmarks, business districts, residential areas, rural settlements, and scenic areas. These trials incorporate real-world uncertainties, extending the analysis from downlink coverage to comprehensively evaluating uplink / downlink channel quality, throughput, and regional interference metrics. The key results demonstrate significant performance gains, such as a maximum increase in received signal strength by 6.14 dB in residential areas and a maximum gain in channel quality 351% on riverside roads, along with a maximum throughput improvement of up to 89% in uplink transmissions. The study further proposes a scenario applicability evaluation framework that includes five dimensions, considering deployment constraints and performance gains, to guide the deployment of RIS in future networks.

Traditional single-architecture models struggle to meet the performance requirements for atmospheric electric field (AEF) prediction. This article proposes a novel hybrid architecture that integrates multimodal radar features for AEF prediction. Specifically, given the strong correlation between real-time radar charts and AEF signals, this study determines the usable regions of radar charts based on the detection range of the AEF apparatus, then designs handcrafted radar chart features and concatenates them with convolutional neural network-extracted features to form a multimodal feature set. Leveraging a cross-modal attention module, it temporally aligns the asynchronous low-frequency radar chart features and high-frequency AEF signals. Finally, the aligned data are input to a temporal convolutional network (TCN) to predict AEF signals. Experimental results demonstrate that our method achieves the most competitive performance, reducing the root mean square error by 42% compared to the long short-term memory-based method. The proposed method provides a feasible solution for asynchronous data fusion in meteorological environment monitoring.

The integration of Reconfigurable Intelligent Surfaces (RIS) into commercial networks is a pivotal step towards realizing the vision of the sixth generation (6G) network. However, field trials in large-scale, complex urban environments remain scarce, leaving a gap in understanding their practical performance under real-world constraints. This paper presents field trial results from a commercial network deployed in the challenging, mountainous city of Chongqing, focusing on three critical transportation infrastructure scenarios: tunnels, river-crossing bridges, and riverside roads. Contrary to idealized setups, our trials demonstrate that RIS delivers substantial performance gains even under non-optimal deployment conditions with large incident angles and long distances. This work goes beyond conventional downlink-centric trials by providing a comprehensive uplink-downlink joint analysis and introducing fine-grained physical-layer metrics, the channel quality indicator (CQI) and the modulation and coding scheme (MCS), which collectively reveal the performance gains of RIS in dynamic environments. Our work offers invaluable empirical evidence and deployment insights for integrating RIS into future wireless networks.

The security of maritime communication for transmitting sea clutter signals is of critical importance from both information-theoretic and electromagnetic signal processing perspectives. Existing prediction models face significant challenges in accurately capturing the non-stationary and chaotic characteristics of sea clutter, a typical electromagnetic scattering signal influenced by complex maritime environments. To address this, we propose a novel maritime covert communication scheme based on information-theoretic secrecy metrics, leveraging communication relay unmanned aerial vehicles to minimize the monitor's detection probability. A hybrid neural network model is developed for sea clutter prediction by integrating convolutional neural networks (CNN), bidirectional long short-term memory (BiLSTM), and an attention mechanism. Initially, phase space reconstruction is applied to sea clutter signals measured by IPIX radar, exploiting their spatiotemporal correlation—a key property in electromagnetic signal analysis. The CNN extracts spatial features from the reconstructed signals, while the BiLSTM models temporal dependencies, effectively mitigating overfitting in long-sequence prediction. The attention mechanism further enhances performance by dynamically weighting salient features. Experimental results demonstrate superior prediction accuracy and robust target detection capability based on prediction errors. This work bridges deep learning-based electro magnetic signal processing and covert communication design, providing insights for the design of secure maritime information systems.

Near-field communication is a key scenario in 6G, especially in indoor environments equipped with large-scale reconfigurable intelligent surfaces (RIS), where its effects are often significant. In this paper, we evaluate RIS performance through system-level simulation in a near-field InH-Office setting. A typical case is considered, in which the RIS-to-mobile station (MS) link falls within the Rayleigh distance, modeled according to InH-Office channel conditions. To evaluate reference signal received power (RSRP) and signal-to-interference-plus-noise ratio (SINR) performance, we focus on path loss and large-scale fading, analyzing the long-term received power via the two-hop RIS-relayed transmission. System-level simulations are carried out using RIS panels of varying sizes, with comparisons made among three configurations: ideal near-field beamforming, farfield beamforming, and no RIS.

With the capability to configure wireless propagation environment, reconfigurable intelligent surface (RIS) has attracted wide attention from both academia and industry, in which measurement campaign of RIS in commercial networks is of great importance for performance evaluation. However, existing RIS configuration schemes in practical environments are generally angle-based which require accurate angle information and mainly consider single incident beam, or statistical methods with high sampling overhead. In this paper, we propose a phase shift design scheme for RIS referred to as multiple incident beam superposition (MIBS) scheme, which can be applied in scenarios with incident signals on the RIS from multiple directions. Instead of utilizing random sampling, the proposed scheme firstly employs an incident beam scanning process to extract the incident beam information by pre-calculated codebook to reduce sampling cost, and requires no complex channel estimation or prior channel information. Then the proposed scheme concentrates the incident beams toward the desired reflection direction through a phase shift superposition algorithm. Numerical simulations verify the excellent performance and a 90% sample reduction of the proposed scheme compared with existing methods under the condition of 1-bit RIS hardware for practical applications. Furthermore, a measurement campaign in 5G commercial networks is conducted to validate the advantages of the proposed MIBS scheme in optimizing crucial signal metrics, yielding a 6.76 dB RSRP gain, a 4.9 dB SINR gain and a 37% throughput improvement, showing great potential of RIS for coverage enhancement.

To address the challenges of user mobility, task dependency, and restricted aerial trajectory planning in maritime mobile edge computing, this study proposes a user mobility-aware air-sea collaborative task offloading framework. We establish a three-tier computing architecture comprising user equipment, aerial auxiliary nodes (AANs), and edge servers, formulating a multi-objective optimization model that minimizes system costs (latency and energy consumption) while considering task dependencies, resource allocation constraints, and AAN 3D spatial safety requirements. The technical contributions include: 1) A dynamic K-means-based user clustering mechanism that periodically updates MD-AAN associations according to real-time mobility patterns; 2) A Twin-delayed Deep Deterministic Policy Gradient based Heterogeneous Agent Offloading (TD3-HAO) algorithm enabling joint optimization of dependent sub-task offloading, resource allocation, and 3D AAN trajectory planning through heterogeneous multi-agent coordination. Simulation results demonstrate superior performance over baseline methods, maintaining less than 3% deviation from optimal solutions when scaling to 25 MDs, with 16.94%-38.34% latency reduction. The proposed solution effectively resolves the resource underutilization caused by ignoring task dependencies and agent homogeneity in existing approaches, providing theoretical guidance for air-sea integrated edge computing systems.

Recently, reconfigurable intelligent surfaces (RIS) technology has been researched to realize holographic communications, by utilizing its property of ultra-dense element spacing. However, the position-fixed RIS can only take advantage of phase-shift domain and reconfigurable holographic surfaces (RHS) can only take advantage of the amplitude domain. Furthermore, most of the previous researches only realize the holographic beamforming, which is only part of the holography theory. Therefore, to leverage the full potential of holography theory, a Fluid-RIS-enabled system is proposed to utilize holographic parameter reconstruction process to circumvent the challenging RIS channel estimation process. Furthermore, ideal and practical communication scenarios are considered, in which the two types of interference are different from the optical holography scenarios. The first type of interference stems from the background noise in the communication environment, and the second type of interference arises from the hardware impairment or phase shift error of the RIS system. To address these challenges, algorithms are proposed to eliminate the influence of second type of interference, even in the simultaneous presence of both interference sources. Consequently, we can perfectly reconstruct the channels’ information as well as the interferences’ information. Finally, simulation results verify the effectiveness and necessity of the proposed algorithms.

With the rapid development of advanced networking and computing technologies such as the Internet of Things, network function virtualization, and 5G infrastructure, new development opportunities are emerging for Maritime Meteorological Sensor Networks (MMSNs). However, the increasing number of intelligent devices joining the MMSN poses a growing threat to network security. Current Artificial Intelligence (AI) intrusion detection techniques turn intrusion detection into a classification problem, where AI excels. These techniques assume sufficient high-quality instances for model construction, which is often unsatisfactory for real-world operation with limited attack instances and constantly evolving characteristics. This paper proposes an Adaptive Personalized Federated learning (APFed) framework that allows multiple MMSN owners to engage in collaborative training. By employing an adaptive personalized update and a shared global classifier, the adverse effects of imbalanced, Non-Independent and Identically Distributed (Non-IID) data are mitigated, enabling the intrusion detection model to possess personalized capabilities and good global generalization. In addition, a lightweight intrusion detection model is proposed to detect various attacks with an effective adaptation to the MMSN environment. Finally, extensive experiments on a classical network dataset show that the attack classification accuracy is improved by about 5% compared to most baselines in the global scenarios.

With the advancement of contemporary technology and the evolution of industrial paradigms, industrial intelligence has emerged as a critical factor in facilitating the enhancement of industrial processes. As core components in electronic manufacturing, printed circuit boards (PCBs)' quality directly determines product reliability. However, automated defect detection faces two major challenges: (1) the coexistence of small targets and complex background textures and (2) the inherent trade-off between high-accuracy detection and real-time processing requirements. To address these challenges, we propose an enhanced lightweight framework based on YOLOv5s, featuring two technical innovations: (1) the HS-LSKA module combines hierarchical split architecture with separable large kernel attention, enabling efficient multi-scale feature fusion with 17.1% fewer parameters than standard YOLOv5s and (2) the global-efficient multi-scale attention mechanism integrates global contextual information through hybrid attention paths, which is particularly effective for small defect detection in cluttered industrial environments. Experimental results demonstrate that the improved model achieves state-of-the-art performance on multiple PCB datasets, with mAP50 of 94.8% and 98.8% on industrial and public datasets, respectively.

Addressing the issues of decreased accuracy and detection failures caused by blurred edges of target objects and reduced image resolution under foggy conditions, a vehicle and pedestrian target detection algorithm combining SPD-Conv structure and lightweight CBAM-FE attention mechanism was proposed. Firstly, the CBAM-FE attention mechanism was introduced into the backbone network to reduce fog interference and enhance foggy edges through local standard deviation pooling and learnable edge detection. Secondly, SPD-Conv is employed to minimize the loss of critical information caused by downsampling and to improve the effect of distant target detection and multi-scale feature fusion. The experimental evaluation reveals that our novel methodology attains the mAP of 76.4% on the RTTS dataset and 44.5% on the Foggy Cityscapes dataset, which is 1.6% and 1.2% higher than those of the original model. Experimental results demonstrate that the proposed algorithm enhances detection precision without compromising its real-time processing capabilities.

Wireless communication technology is one of the key technologies to achieve the key requirements of high reliability, low latency and mass connectivity in the Internet of Things (IoT). To meet these requirements, we apply the reconfigurable intelligent surface (RIS) in the wireless IoT networks. Specifically, we study a scheme that combines the joint beamforming design with RIS selection strategy for multi-RIS aided multicell wireless networks. In the joint beamforming design, we propose the controlling mode of each base station (BS) controlling its local RISs for nonfree propagation scenarios in wireless IoT networks, such as smart city and intelligent transportation system. Specifically, we give the details of algorithm for this control mode, and prove that it can achieve the performance comparable to the in the nonfree propagation scenario, with lower algorithm complexity and control signaling overhead. Moreover, based on the joint beamforming design, we further propose a genetic algorithm-based RIS selection strategy to maximize the sum-rate under the constraint of limited RIS deployment. Finally, simulation results show that the proposed RIS selection strategy is effective and outperforms the existing schemes.

Direct antenna modulation (DAM) enhances spectral efficiency and simplifies transmitter design, benefiting the sixth generation (6 G) communications and Internet of Things (IoT) network. Traditional DAM techniques have been constrained by hardware complexity, limited modulation orders, and reliance on cooperative signal sources, which restrict their applicability in wireless communications. To address these limitations, we propose a metasurface-enabled architecture with a dual-antenna receiver that facilitates high-order modulation by dynamically controlling the noncooperative incident electromagnetic waves without any radio frequency (RF) power consumption. Theoretical analysis and numerical simulations demonstrate the feasibility of this approach, showcasing its ability to achieve higher data rates. This work not only advances the field of direct antenna modulation but also highlights the transformative potential of information metasurface in modern wireless communications.

Significant breakthroughs in the Internet of Things (IoT) and 5G technologies have driven several smart healthcare activities, leading to a flood of computationally intensive applications in smart healthcare networks. Mobile Edge Computing (MEC) is considered as an efficient solution to provide powerful computing capabilities to latency or energy sensitive nodes. The low-latency and high-reliability requirements of healthcare application services can be met through optimal offloading and resource allocation for the computational tasks of the nodes. In this study, we established a system model consisting of two types of nodes by considering nondivisible and trade-off computational tasks between latency and energy consumption. To minimize processing cost of the system tasks, a Mixed-Integer Nonlinear Programming (MINLP) task offloading problem is proposed. Furthermore, this problem is decomposed into task offloading decisions and resource allocation problems. The resource allocation problem is solved using traditional optimization algorithms, and the offloading decision problem is solved using a deep reinforcement learning algorithm. We propose an Online Offloading based on the Deep Reinforcement Learning (OO-DRL) algorithm with parallel deep neural networks and a weight-sensitive experience replay mechanism. Simulation results show that, compared with several existing methods, our proposed algorithm can perform real-time task offloading in a smart healthcare network in dynamically varying environments and reduce the system task processing cost.

Compared to terrestrial networks, maritime networks lack centralized infrastructure, and tasks within maritime ultrareliable applications are interdependent, complicating task offloading and resource allocation. Additionally, in the public area of maritime networks, task offloading processes are vulnerable to malicious attacks. To ensure real-time and reliable performance while mitigating the risks of miscalculations, this paper proposes a multi-agent deep reinforcement learning algorithm for task offloading. The proposed algorithm features a real-time alerting mechanism and reliability constraints, focusing on minimizing system costs, including latency and energy consumption. Maritime applications are modeled as directed acyclic graphs, and the offloading problem is framed as a decentralized Markov decision process. We have developed an enhanced actor-critic architecture by transforming conventional fully connected networks into tailored parallel sequence-to-sequence networks, which improves task feature extraction. This approach allows the system to adapt more effectively to the dynamic challenges of marine environments. Experimental results show that the proposed algorithm outperforms existing solutions, demonstrating significant improvements in task offloading and resource allocation for interdependent tasks.

Satellite communication technology has emerged as a key solution to address the challenges of data transmission in remote areas. By overcoming the limitations of traditional terrestrial communication networks, it enables long-distance data transmission anytime and anywhere, ensuring the timely and accurate delivery of water level data, which is particularly crucial for fishway water level monitoring. To enhance the effectiveness of fishway water level monitoring, this study proposes a multi-task learning model, AS-SOMTF, designed for real-time and comprehensive prediction. The model integrates auxiliary sequences with primary input sequences to capture complex relationships and dependencies, thereby improving representational capacity. In addition, a novel time-series embedding algorithm, AS-SOM, is introduced, which combines generative inference and pooling operations to optimize prediction efficiency for long sequences. This innovation not only ensures the timely transmission of water level data but also enhances the accuracy of real-time monitoring. Compared with traditional models such as Transformer and Long Short-Term Memory (LSTM) networks, the proposed model achieves improvements of 3.8% and 1.4% in prediction accuracy, respectively. These advancements provide more precise technical support for water level forecasting and resource management in the Diqing Tibetan Autonomous Prefecture of the Lancang River, contributing to ecosystem protection and improved operational safety.