Search Results (3,175)

Deploying artificial intelligence models on low-cost edge devices requires balancing predictive accuracy with strict constraints on computational resources, such as inference latency and memory footprint. Despite growing interest in TinyML systems, limited empirical evidence exists on how these factors interact across different embedded hardware platforms. This study presents a systematic multi-objective evaluation of three lightweight AI architectures—multinomial logistic regression (MLR), multilayer perceptron (MLP), and a reduced convolutional neural network (CNN)—implemented natively on three representative platforms: ESP32-S3, Raspberry Pi Pico, and Raspberry Pi Zero W. The models were evaluated on three image classification datasets of increasing complexity (Synthetic Geometric Figures, MNIST, and Fashion-MNIST), measuring classification accuracy, inference latency, and peak memory footprint under real execution conditions. Pareto-front analysis was used to identify efficient model–platform configurations and characterize the trade-offs between predictive performance and computational resources. The results provide quantitative insight into accuracy–resource trade-offs and establish a reproducible framework for evaluating lightweight AI models on resource-constrained edge devices, supporting informed design decisions in TinyML applications. Full article

With the proliferation of edge computing in IoT and smart security, there is a growing demand for large-scale encrypted traffic anomaly detection. However, the opaque nature of encrypted traffic makes it difficult for traditional detection methods to balance efficiency and accuracy. To address this challenge, this paper proposes FMTF, a Multi-Scale Feature Fusion method based on Federated Learning for encrypted traffic anomaly detection. FMTF constructs graph structures at three scales—spatial, statistical, and content—to comprehensively characterize traffic features. At the spatial scale, communication graphs are constructed based on host-to-host IP interactions, where each node represents the IP address of a host and edges capture the communication relationships between them. The statistical scale builds traffic statistic graphs based on interactions between port numbers, with nodes representing individual ports and edge weights corresponding to the lengths of transmitted packets. At the content scale, byte-level traffic graphs are generated, where nodes represent pairs of bytes extracted from the traffic data, and edges are weighted using pointwise mutual information (PMI) to reflect the statistical association between byte occurrences. To extract and fuse these multi-scale features, FMTF employs the Graph Attention Network (GAT), enhancing the model’s traffic representation capability. Furthermore, to reduce raw-data exposure in distributed edge environments, FMTF integrates a federated learning framework. In this framework, edge devices train models locally based on their multi-scale traffic features and periodically share model parameters with a central server for aggregation, thereby optimizing the global model without exposing raw data. Experimental results demonstrate that FMTF maintains efficient and accurate anomaly detection performance even under limited computing resources, offering a practical and effective solution for encrypted traffic identification and network security protection in edge computing environments. Full article

Structural reconstruction helps infer the spatial relationships and object layouts in a scene, which is an essential computer vision task for understanding visual content. However, it remains challenging due to the high complexity of scene structural topologies in real-world environments. To address this challenge, this paper proposes RegionGraph, a novel method for structural reconstruction of buildings from a satellite image. It utilizes a layout region graph construction and graph contraction approach, introducing a primitive (layout region) estimation network named ConPNet for detecting and estimating different structural primitives. By combining structural extraction and rendering synthesis processes, RegionGraph constructs a graph structure with layout regions as nodes and adjacency relationships as edges, and transforms the graph optimization process into a node-merging-based graph contraction problem to obtain the final structural representation. The experiments demonstrated that RegionGraph achieves a 4% improvement in average F1 scores across three types of primitives and exhibits higher regional completeness and structural coherency in the reconstructed structure. Full article

To address the urgent need for enhancing landslide risk monitoring in reservoir embankment slopes, a core component of water conservancy projects, this paper proposes the LRES-YOLO algorithm for real-time landslide detection on reservoir embankments. In LRES-YOLO, we first integrate coordinate attention into basic feature extraction convolutional blocks to form the CACBS attention module, which enhances the model’s ability to identify and locate landslide targets in complex reservoir terrain, overcoming positional information insensitivity in deep networks. Second, we add novel downsampling DP modules and ELAN-W modules to the backbone network, improving feature recognition efficiency for embankment slopes with diverse hydrological and topographical interference. Third, we optimize the feature fusion network with targeted concatenation and pooling operations, balancing semantic information enhancement with computational load reduction to mitigate overfitting in variable reservoir environments. Finally, we adopt Focal Loss and EIoU Loss to accelerate training convergence and strengthen target feature representation for small or obscured landslides on embankments. Experimental results show that LRES-YOLO outperforms traditional algorithms in detecting landslides across diverse reservoir embankment scenarios: it achieves an average improvement of 8.4 percentage points in mean mAP over the best-performing baseline across five independent trials, a detection speed of 8.2 ms per image, and memory usage of 139 MB. This lightweight design makes it suitable for edge computing devices, providing robust technical support for intelligent monitoring systems in water conservancy projects. Full article

Eigenvector-based centrality captures recursive notions of importance in networks. While the direct problem computes centrality from given edge weights, the inverse eigenvector centrality problem seeks edge weights that reproduce a prescribed centrality profile; for directed multigraphs, this inverse task is typically non-unique and depends on the admissible arc structure. We study the direct and inverse problems on directed multigraphs and derive an explicit linear characterization of the set of admissible edge-weight vectors that are compatible with a given centrality target. On this feasible set, we formulate a generalized Nash equilibrium problem with shared centrality constraints, in which multiple agents select edge weights to maximize economically interpretable payoffs that incorporate arc-level competition effects. We provide conditions under which the induced game admits a concave potential function, yielding equilibrium existence and, under standard strict concavity assumptions, uniqueness. Finally, we illustrate the model on an airport network where nodes represent airports and parallel arcs represent airline-specific routes, showing that equilibrium selection produces a feasible and interpretable weight configuration that preserves the prescribed centrality. Full article

In intelligent excavator applications, traditional excavator posture recognition methods face two major challenges: limited recognition accuracy and insufficient computing resources on edge devices. To address these issues, this study proposes an excavator posture recognition method based on an improved Real-Time Detection Transformer (RT-DETR). First, a new backbone network is designed based on the Reparameterized Vision Transformer to improve feature utilization efficiency while reducing computational demands. Next, the overall architecture is optimized by introducing lightweight Dynamic Upsamplers, which reduce information loss during upsampling and enhance multi-scale feature fusion. In addition, a Cross-Attention Fusion Module is adopted to strengthen local feature extraction while retaining the global modeling capability of the Transformer, thereby improving the discrimination between foreground and background. Finally, a Multi-Scale Fusion Network is introduced to further enhance the multi-scale feature representation ability of RT-DETR. Experimental results show that the proposed method achieves a mean average precision (mAP) of 94.29% for small object detection, which is 7.96% higher than that of the baseline RT-DETR, while reducing the number of model parameters by 34.95%. Compared with YOLO-series models, the proposed method improves mAP by 8.62% to 12.75%. These results indicate that the proposed method outperforms existing methods in both detection accuracy and computational efficiency and provides an efficient and feasible solution for real-time excavator posture recognition. Full article

Millimeter-wave radar has gradually gained popularity as a sensor mode for Human Activity Recognition (HAR) in recent years because it preserves the privacy of individuals and is resistant to environmental conditions. Nevertheless, the fast inference of high-dimensional and sparse 4D radar data is still difficult to perform on low-resource edge devices. Current models, including 3D Convolutional Neural Networks and Transformer-based models, are frequently plagued by extensive parameter overhead or quadratic computational complexity, which restricts their applicability to edge applications. The present paper attempts to resolve these issues by introducing RadarSSM as a lightweight spatiotemporal hybrid network in the context of radar-based HAR. The explicit separation of spatial feature extraction and temporal dependency modeling helps RadarSSM decrease the overall complexity of computation significantly. Specifically, a spatial encoder based on depthwise separable 3D convolutions is designed to efficiently capture fine-grained geometric and motion features from voxelized radar data. For temporal modeling, a bidirectional State Space Model is introduced to capture long-range temporal dependencies with linear time complexity $O\left(T\right)$, thereby avoiding the quadratic cost associated with self-attention mechanisms. Extensive experiments conducted on public radar HAR datasets demonstrate that RadarSSM achieves accuracy competitive with state-of-the-art methods while substantially reducing parameter count and computational cost relative to representative convolutional baselines. These results validate the effectiveness of RadarSSM and highlight its suitability for efficient radar sensing on edge hardware. Full article

We investigate wheat lodging segmentation from UAV RGB imagery acquired over real production fields rather than controlled experimental sites. Besides pixel-level accuracy, our evaluation also emphasizes robustness under heterogeneous farmland conditions and deployment-oriented efficiency. We propose CASGNet, an edge-oriented segmentation network with a content-aware spatial gating mechanism that reweights intermediate features according to local structural variation. Instead of uniformly aggregating features, the module suppresses responses in homogeneous regions while preserving activation in structurally complex areas. In practice, this improves the continuity of irregular lodging shapes and reduces spurious responses in relatively homogeneous backgrounds. The dataset spans 46 farms across Jiaozuo, Jiyuan, and Luoyang, covering progressively fragmented farmland. Under a stricter mission-level data-isolation protocol, CASGNet achieves 94.4% mIoU and 90.38% IoU for the lodging class on the combined dataset. Under sequential regional adaptation, performance remains relatively stable in continuous parcels, and degradation is less severe than most compact baselines in highly fragmented landscapes. On Jetson Nano, CASGNet achieves 1.94 FPS embedded inference under the 5 W mode. Smaller networks achieve higher speed but show reduced structural continuity in complex scenes. The results indicate that CASGNet provides a favorable balance between structural fidelity and computational cost, while its robustness remains constrained by scene complexity. Full article

NetworkGuard is a modular edge-based virtual network sensing framework designed for residential smart home security. The system interprets network telemetry—such as DNS queries, firewall events, VPN latency, and connection establishment delay—as structured sensing signals for gateway-level monitoring. Implemented on a Raspberry Pi 4 and managed via an Android interface, NetworkGuard integrates DNS filtering (Pi-hole), firewall enforcement (UFW), encrypted VPN tunneling (WireGuard), and an AI-assisted advisory layer for contextual log interpretation. During a six-week residential deployment, DNS blocking efficiency improved from 81.2% to 97.0% following blocklist refinement, while VPN connection establishment time decreased from approximately 3012 ms to 2410 ms after configuration tuning. ICMP-based measurements indicated a stable tunnel latency under moderate traffic conditions. Controlled validation scenarios—including DNS manipulation attempts, port scanning, and VPN interruption testing—confirmed consistent firewall enforcement and tunnel containment. The results demonstrate that layered security principles can be adapted into a lightweight, reproducible edge architecture suitable for small-scale residential IoT environments without a reliance on enterprise infrastructure. Full article

Grid-forming (GFM) battery energy storage system (BESS) inverters are becoming a cornerstone of resilient microgrids, where severe voltage sags and abrupt operating shifts can challenge both voltage regulation and controller stability. Finite-set model predictive control (FS-MPC) offers fast transient response and multi-objective coordination, yet conventional designs rely on static cost-function weights that are typically tuned offline and may become suboptimal under disturbance-driven regime changes. This paper proposes a forecast-guided KAN-adaptive FS-MPC framework that (i) formulates the inner-loop predictive control in the stationary αβ frame, thereby avoiding PLL dependency and mitigating loss-of-lock risk under extreme sags, and (ii) introduces an Operating Stress Index (OSI) that fuses load forecasts with reserve-margin or percent-operating-reserve signals to quantify grid vulnerability and trigger resilience-oriented control adaptation. A lightweight Kolmogorov–Arnold Network (KAN), parameterized by learnable B-spline edge functions, is embedded as an online weight governor to update key FS-MPC weighting factors in real time, dynamically balancing voltage tracking and switching effort. Experimental validation under high-frequency microgrid scenarios shows that, under a 50% symmetrical voltage sag, the proposed controller reduces the worst-case voltage deviation from 0.45 p.u. to 0.16 p.u. (64.4%) and shortens the recovery time from 35 ms to 8 ms (77.1%) compared with static-weight FS-MPC. In the islanding-like transition case, the proposed method restores the PCC voltage within 18 ms, whereas the static baseline fails to recover within 100 ms. Moreover, the deployed KAN governor requires only 6.2 μs per inference on a 200 MHz DSP, supporting real-time embedded implementation. These results demonstrate that forecast-guided adaptive weighting improves transient resilience and power quality while maintaining DSP-feasible computational complexity. Full article

Neurodegenerative diseases (NDs), including Alzheimer’s disease, Parkinson’s disease, Huntington’s disease, and amyotrophic lateral sclerosis (ALS), represent a growing global health challenge characterized by progressive neuronal loss and a lack of definitive disease-modifying treatments. This review explores the emerging potential of targeting non-coding RNAs (ncRNAs), such as microRNAs (miRNAs), long non-coding RNAs (lncRNAs), and exosomal RNAs, to modulate pathogenic molecular pathways and address the underlying molecular origins of neurodegeneration. We evaluate the integration of advanced computational techniques for RNA structure prediction and gene regulatory network analysis, alongside chemical engineering strategies—such as Locked Nucleic Acids (LNAs) and phosphorothioate modifications—aimed at enhancing the stability and specificity of RNA-based molecules. Furthermore, we analyze cutting-edge delivery and editing technologies, including nanotechnology-driven solutions for precise neuronal targeting and the CRISPR/Cas13 system for direct ncRNA manipulation.The findings indicate that while challenges in delivery efficiency and long-term efficacy persist, the synergy of chemical engineering and computational modeling significantly improves the therapeutic profile of ncRNAs, with exosomal pathways offering a novel route for intercellular signaling modulation and biomarker discovery. Therapeutic interventions directed at specific clinical targets, such as miR-34a and BACE1-AS, demonstrate the capacity to influence protein aggregation and neuroinflammatory cascades. Although ncRNA-based therapies are currently in nascent stages, ongoing technological advancements in RNA editing and nanotechnology offer a transformative framework that could redefine the future of ND treatment and successfully halt disease progression rather than merely managing symptoms. Full article

The rapid proliferation of mobile devices, particularly smartphones and tablets, has transformed digital entertainment, with mobile gaming emerging as one of the fastest-growing digital segments. Such applications are inherently latency-sensitive and require effective resource management and seamless mobility support. To overcome these issues, this paper suggests a four-layered infrastructure that combines edge, fog, and cloud computing with Software-Defined Networking (SDN) and is assisted by a lightweight proximity-aware heuristic placement strategy and mobility management. The suggested structure follows a microservices contained breakdown of the gaming functionality and uses clustering algorithms to permit coordinated access to resources by edge and fog nodes. A dynamic lightweight proximity-aware virtual machine placement algorithm is presented to deploy application modules nearer to the users depending on the availability and mobility of the resources. The proposed work is simulated using IFogSim2. The proposed model reduces the latency by up to 73 percent and the rate of task completion by 25 percent relative to baseline configurations in the case of dynamic mobility of users. These results indicate that the suggested strategy can be effective in improving the latency-sensitive mobile gaming applications performance in the edge-fog networks. Full article

As the scale and complexity of Internet of Things (IoT) deployments increase, maintaining resilience in resource-constrained mesh networks becomes a significant challenge. Frequent node failures due to battery depletion, environmental interference, or hardware degradation can disrupt data flows and lead to operational downtime. To address this, we propose EdgeRescue, a novel lightweight AI-driven framework for self-healing in energy-constrained IoT mesh environments. EdgeRescue enables each node to perform local anomaly detection using compact 1D Convolutional Neural Networks (1D-CNNs) and initiates distributed, energy-aware routing reconfiguration when faults are detected. Unlike cloud-dependent methods, EdgeRescue operates entirely at the edge, requiring minimal computation, memory, and communication overhead. Extensive simulations on a 100-node testbed demonstrate that EdgeRescue improves packet delivery by 13.2%, reduces recovery latency by 57%, and lowers average node energy consumption by 18.8% compared to state-of-the-art baselines. These results establish EdgeRescue as a scalable and practical solution for achieving real-time resilience in next-generation IoT mesh networks. Full article

This paper presents a dual-core RISC-V architecture designed for energy-efficient AI acceleration at the edge, featuring dynamic MAC unit sharing, frequency scaling (DFS), and FIFO-based resource arbitration. The system comprises two RISC-V cores that compete for shared computational resources—a single Multiply–Accumulate (MAC) unit and a shared external memory subsystem—governed by a channel-based arbitration mechanism with CPU-priority semantics, while each core maintains private instruction and data caches. The architecture implements a tightly coupled Neural Processing Unit (NPU) with CONV, GEMM, and POOL operations that execute opportunistically in the background when the MAC unit is available. Dynamic frequency scaling (DFS) with three levels (100/200/400 MHz) is applied to the shared MAC unit, allowing the dynamic acceleration of CNN workloads. The arbitration mechanism uses SystemC sc_fifo channels with CPU-priority polling, ensuring that CPU execution is minimally impacted by background AI processing while the NPU makes progress during idle MAC slots. The NPU supports 3 × 3 convolutions, matrix multiplication (GEMM) with 10 × 10 tiles, and pooling operations. The implementation is cycle-accurate in SystemC, targeting FPGA deployment. Experimental evaluation demonstrates that the dual-core architecture achieves 1.87× speedup with 93.5% efficiency for parallel workloads, while DFS enables 70% power reduction at low frequency. The system successfully executes simultaneous CPU and AI workloads, with CPU-priority arbitration ensuring no CPU starvation under contention. The proposed design offers a practical solution for embedded AI applications requiring both general-purpose computation and neural network acceleration, validated through comprehensive SystemC simulation on modern FPGA platforms. Full article

The deep image prior (DIP) suggests that it is possible to train a randomly initialized network with a suitable architecture to solve inverse imaging problems by simply optimizing its parameters to reconstruct a single degraded image. However, the prior knowledge exploited by vanilla DIP relies on basic local convolutions, which inevitably limits the performance of inverse imaging tasks to the generative capacity of the model. Furthermore, image information is often not only related to neighboring pixels but also dependent on global color features and spatial distribution. Simple local convolutions used in inverse imaging cannot capture precise fine-grained details. Moreover, DIP is an unsupervised process but requires iterations to learn inverse imaging, consuming computational power and limiting the adaptation of global attention. To solve these problems, this article explores an efficient global prior module—a tri-directional multi-head self-attention mechanism—aiming to learn pixel-wise correlations along three directions: horizontal, vertical, and channel-wise. Our observations found that global learning can effectively enhance the detail information of edge pixels, making images more vivid and textures clearer. In addition, tri-directional multi-head self-attention can efficiently replace the global perception ability of pixel-level self-attention. Finally, we demonstrate that global learning can effectively improve the imaging effect of inverse imaging problems and enhance the information of texture edge pixels. Moreover, tri-directional multi-head self-attention can effectively alleviate the computation redundancy of pixel-level self-attention, thus achieving efficient and high-quality inverse imaging tasks. The principle of this method lies in global feature capture and efficient attention modeling, striking a balance between detail fidelity and computational practicality. Full article