Search Results (861)

Teleoperated open-pit excavation requires fast and reliable 3D scene modeling under lightweight sensor configurations. To this end, this paper proposes a monocular camera–LiDAR fusion-based fast 3D Gaussian reconstruction method tailored for teleoperated open-pit excavation. The proposed approach uses only two sensors, a monocular camera and LiDAR, and integrates SPNet, a depth completion network, to improve the geometric completeness of the reconstructed scene. It further introduces a stride-aware initialization strategy that leverages the depth–stride correlation to jointly construct the initial Gaussian set and estimate the initial scales. During optimization, scale and color regularization are applied to prevent uncontrolled growth of Gaussians. Experiments in a Carla-simulated open-pit excavation scenario show that, under high-resolution input of 1920 × 1080, the proposed method achieves a stable 3D model update rate of approximately 2.5 Hz. The reconstruction quality under training viewpoints reaches PSNR 30.5388, SSIM 0.9161, and LPIPS 0.1333. Compared with 4DTAM and MonoGS, the proposed method achieves better overall reconstruction quality. It also maintains a much higher update rate than 4DTAM and a comparable update rate to MonoGS. Ablation studies further verify the critical contribution of the depth completion module and the stride-aware initialization strategy to the overall reconstruction performance. In addition, preliminary validation on field data further demonstrates the applicability of the proposed method under real-world open-pit excavation-loading conditions. The proposed method generates stable and usable 3D models of rock-pile working face under a lightweight sensor configuration, providing a reliable geometric basis for remote situational awareness and excavation assistance. Full article

To further explore the energy-saving potential of power-split hybrid electric vehicles, this paper addresses issues in traditional Radial Basis Function (RBF) neural network-based vehicle speed prediction methods, which rely solely on time-varying information from historical speed sequences of the host vehicle, leading to problems such as idle overestimation, large local prediction errors, and low prediction accuracy across different time horizons. An improved RBF neural network-based vehicle speed prediction method that integrates multi-sensor information is proposed. This method identifies the driver’s driving intention through a fuzzy inference system, extracts historical speed sequences within a fixed time window in a rolling manner, and integrates inter-vehicle motion characteristic parameters obtained through fusion of millimeter-wave radar and camera data. These multi-dimensional influencing factors are used as inputs to the RBF neural network for vehicle speed prediction. Based on this, an energy management optimization model for the vehicle is established, with the goal of optimizing fuel economy. The model predictive control (MPC) strategy is employed, and the Dynamic Programming (DP) algorithm is used to solve for the real-time optimal torque distribution among various power sources within a limited time horizon. Finally, simulation validation is conducted on the MATLAB/Simulink platform under the CHTC-B driving cycle, CCBC driving cycle, and actual road driving cycle. The results show that, compared with the traditional method adopting Radial Basis Function (RBF) neural network-based vehicle speed prediction and rule-based energy management, the proposed method improves the vehicle’s fuel economy by 4.11%. Full article

The Internet of Things (IoT) uses networked devices, dispersed sensors, and cameras to create huge, diverse data streams. Concept drift, in which the underlying data distribution shifts over time, is frequently caused by the non-stationary and multimodal character of these streams. Static machine learning models, based on fixed data distributions, reduce forecast accuracy and system reliability since they are unable to adapt to such changes. This paper proposes an Adaptive Multimodal Long Short-Term Memory (AM-LSTM) architecture to address these challenges by combining modality-specific temporal modelling, attention-based dynamic fusion, and drift-aware online learning. An attention mechanism adaptively weights informative streams to mitigate the impact of noisy or missing input, while specialist LSTM encoders capture the temporal correlations of each modality. Concept drift is detected using a sliding-window error monitoring technique, and adaptive learning rate adjustment and selective retraining are started when significant distributional changes occur. The proposed system is tested under synthetic drift conditions using the Edge-IoT and UNSW-NB15 benchmark datasets. Experimental results demonstrate that AM-LSTM achieves 88.7% accuracy and an F1-score of 0.85, adapting to drift within 620 samples while maintaining an average update latency of 47 ms per batch. Compared with static and existing adaptive baselines, the proposed approach provides improved robustness, faster drift adaptation, and computational efficiency suitable for real-time IoT environments. Full article

Geothermal exploration and monitoring often require dense temperature observations in terrains where wired networks are impractical and battery replacement for in situ sensors is costly. This paper proposes an infrared-assisted, temperature-aware access scheme for a UAV-enabled backscatter IoT network tailored to geothermal hotspot sensing. A rotary-wing UAV equipped with a thermal infrared camera and an RF transceiver first surveys the area to construct a surface temperature map and identify candidate hotspots, and then hovers above a selected hotspot to perform periodic frames consisting of wireless energy transfer followed by backscatter uplink collection. Ground sensors harvest RF energy, measure their local temperature, and autonomously activate only when both the harvested energy exceeds a threshold and the measured temperature falls within a target interval broadcast by the UAV, thereby concentrating channel access on thermally relevant nodes. We develop a system model that couples a geothermal-like thermal field, RF energy harvesting, and framed slotted backscatter access, and introduce hotspot-oriented performance metrics including effective hotspot throughput, task completion time, and energy per hotspot report. The simulation results show that the proposed temperature–energy-gated access significantly increases the fraction of successfully decoded packets originating from hotspot regions and improves the energy efficiency of geothermal monitoring compared with full activation and purely energy-based activation. Full article

Automotive vision systems depend critically on front-view cameras, whose image quality frequently degrades under adverse conditions such as rain, fog, low illumination, and rapid motion. To address this challenge, we propose VMESR, a variable mamba-enhanced super-resolution network that integrates a selective state-space model into a lightweight super-resolution architecture. By serializing 2D feature maps and applying variable-depth mamba blocks, VMESR captures long-range dependencies with linear complexity. A multi-scale feature extractor, enhanced residual modules equipped with a convolutional block attention module, and dense fusion connections work together to improve the recovery of high-frequency details. Extensive experiments demonstrate that VMESR achieves competitive performance in both objective metrics and perceptual quality compared to state-of-the-art methods, while significantly reducing parameter counts and computational cost. VMESR provides a practical balance between efficiency and reconstructive accuracy, offering a deployable super-resolution solution for embedded automotive sensors and enhancing the robustness of autonomous driving perception pipelines. Full article

Although underwater gliders are highly energy-efficient platforms capable of long-duration and large-scale ocean observation, their lack of self-propulsion requires external assistance for recovery upon mission completion. In harsh and dynamic marine environments, reliably detecting the glider and accurately estimating its three-dimensional position are critical to ensuring the recovery operations are safe and efficient. This paper proposes a perception framework based on deep learning to detect underwater glider hulls and estimate their three-dimensional relative positions using camera–sonar multi-sensor fusion. This approach integrates a hierarchical convolutional neural network (CNN) vision encoder and a transformer-based architecture to estimate the glider’s spatial location and heading direction simultaneously. The hierarchical CNN encoder extracts multi-level, semantically rich visual features, thereby improving robustness to visual degradation and environmental disturbances common in underwater settings. Additionally, the transformer incorporates a variable mixture-of-experts (vMoE) mechanism that adaptively allocates expert networks across layers, enhancing representational capacity while maintaining computational efficiency. The resulting pose estimates enable precise, collision-free ROV navigation for automated recovery and onboard sensor inspection tasks. Experimental results, including ablation studies, validate the effectiveness of the proposed components and demonstrate their contributions to accurate glider hull detection and three-dimensional localization. Overall, the proposed framework provides a scalable, reliable perception solution that allows for the safe, autonomous recovery of underwater gliders with an ROV in realistic ocean environments. Full article

Forest fires are a major concern due to their significant impact on the environment, economy, and wildlife habitats. Efficient early detection systems can significantly mitigate their devastating effects. This paper provides a comprehensive review of forest fire detection (FFD) techniques and traces their evolution from basic lookout-based methods to sophisticated remote sensing technologies, including recent Internet of Things (IoT)- and Unmanned Aerial Vehicle (UAV)-based sensor network systems. Historical methods, characterized primarily by human surveillance and basic electronic sensors, laid the foundation for modern techniques. Recently, there has been a noticeable shift toward ground-based sensors, automated camera systems, aerial surveillance using drones and aircraft, and satellite imaging. Moreover, the rise of Artificial Intelligence (AI), Machine Learning (ML), and the IoT introduces a new era of advanced detection capabilities. These detection systems are being actively deployed in wildfire-prone regions, where early alerts have proven critical in minimizing damage and aiding rapid response. All FFD techniques follow a common path of data collection, pre-processing, data compression, transmission, and post-processing. Providing sufficient power to complete these tasks is also an important area of research. Recent research focuses on image compression techniques, data transmission, the application of ML and AI at edge nodes and servers, and the minimization of energy consumption, among other emerging directions. However, to build a sustainable FFD model, proper sensor deployment is essential. Sensors can be either fixed at specific geographic locations or attached to UAVs. In some cases, a combination of fixed and UAV-mounted sensors may be used. Careful planning of sensor deployment is essential for the success of the model. Moreover, ensuring adequate energy supply for both ground-based and UAV-based sensors is important. Replacing sensor batteries or recharging UAVs in remote areas is highly challenging, particularly in the absence of an operator. Hence, future FFD systems must prioritize not only detection accuracy but also long-term energy autonomy and strategic sensor placement. Integrating renewable energy sources, optimizing data processing, and ensuring minimal human intervention will be key to developing truly sustainable and scalable solutions. This review aims to guide researchers and developers in designing next-generation FFD systems aligned with practical field demands and environmental resilience. Full article

A Digital Twin is a virtual environment that simulates, predicts, and optimizes the performance of its physical counterpart. Digital Twin models hold great potential in wireless networking testing and development. This paper aims to envision our concept of simulating the operation of different sensors in vehicle test-track conditions. Vehicle parameters are embedded into the edge computing entity, which uses them to generate a test configuration for the Digital Twin. This configuration is then applied in simulated sensor-output prediction, ultimately producing event data for the vehicle entity. The sensor suite—comprising radar, cameras, GPS and LiDAR—is modeled to provide the multi-modal input required for generating simulated perception data in the Digital Twin. To ensure realistic perception behavior, the physical vehicle is represented within a digital environment that reproduces the actual test track. This allows LiDAR occlusions to be attributed to genuine environmental structures (e.g., trees, buildings, other vehicles) rather than simulation artifacts. Within the Digital Twin, the objective is to evaluate how sensor signals—such as radar waves and LiDAR light pulses—propagate through the environment and how real-world obstacles may weaken or distort them. Historical datasets are used to calibrate and validate the Digital Twin, ensuring that the simulated sensor behavior aligns with real-world observations; the data collected during previous test runs can be used for visualization and analysis. Weather conditions are modeled to evaluate how rain, fog and snow impact sensor performance within the Digital Twin environment, to learn about the effects and predict sensor operation in different weather conditions. In this article, we examine the Digital Twin of our test track as a development environment for designing, deploying and testing ITS-enhanced road-weather services and warnings. These services integrate real-world road-weather observations, forecast data, roadside sensors and on-board vehicle measurements to support safe driving and optimize vehicle trajectories for both passenger and autonomous vehicles. This research is expected to benefit stakeholders involved in automotive testing, simulation and road-weather service development. Full article

Human activity recognition (HAR) is essential in many applications, such as smart homes, assisted living, healthcare monitoring, rehabilitation, physiotherapy, and geriatric care. Conventional methods of HAR use wearable sensors, e.g., acceleration sensors and gyroscopes. However, they are limited by issues such as sensitivity to position, user inconvenience, and potential health risks with long-term use. Optical camera systems that are vision-based provide an alternative that is not intrusive; however, they are susceptible to variations in lighting, intrusions, and privacy issues. The paper uses an optical method of recognizing human domestic activities based on pose estimation and deep learning ensemble models. The skeletal keypoint features proposed in the current methodology are extracted from video data using PoseNet to generate a privacy-preserving representation that captures key motion dynamics without being sensitive to changes in appearance. A total of 30 subjects (15 male and 15 female) were sampled across 2734 activity samples, including nine daily domestic activities. There were six deep learning architectures, namely, the Transformer (Transformer), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Multilayer Perceptron (MLP), One-Dimensional Convolutional Neural Network (1D CNN), and a hybrid Convolutional Neural Network–Long Short-Term Memory (CNN–LSTM) architecture. The results on the hold-out test set show that the CNN–LSTM architecture achieves an accuracy of 98.78% within our experimental setting. Leave-One-Subject-Out cross-validation further confirms robust generalization across unseen individuals, with CNN–LSTM achieving a mean accuracy of 97.21% ± 1.84% across 30 subjects. The results demonstrate that vision-based pose estimation with deep learning is a useful, precise, and non-intrusive approach to HAR in smart healthcare and home automation systems. Full article

Monocular camera localization in known environments is a critical task for applications like autonomous navigation, augmented reality, and robotic positioning, requiring precise spatial awareness. Unlike localization in unknown environments, which builds maps in real time, this leverages pre-existing data for higher accuracy. This review comprehensively analyzes monocular camera localization methods in known environments, categorizing them into 2D-2D feature matching, 2D-3D feature matching, and regression-based approaches. It consolidates foundational techniques and recent advancements, providing inter-class and intra-class performance comparisons on mainstream datasets. Key findings show that 2D-3D methods generally offer the highest accuracy, especially in structured outdoor environments, due to robust use of 3D spatial information. However, recent scene coordinate regression methods, such as ACE and ACE++, achieve comparable or superior performance in indoor scenes with more efficient pipelines. This review highlights challenges and proposes future directions: (1) synthetic data generation to meet deep learning demands, while addressing domain gaps; (2) improving generalization to unseen scenes and reducing retraining; (3) multi-sensor fusion for enhanced robustness; (4) exploring transformer-based and graph neural network architectures; (5) developing lightweight models for real-time performance on resource-constrained devices. This review aims to guide researchers and practitioners in method selection and identify key research directions. Full article

Traditional post-earthquake structural health monitoring (SHM) methods based on dedicated sensors lack scalability due to installation and maintenance demands, leaving most buildings unmonitored. This study investigates the use of existing in-building surveillance cameras to infer structural demand by tracking earthquake-induced building motion. The proposed methodology repurposes ceiling-mounted surveillance cameras to estimate the inter-story drift (IDR) which is directly correlated with structural damage using FEMA guidelines. Shake-table experiments spanning a wide range of excitation intensities and dominant frequencies demonstrate that off-the-shelf surveillance cameras can estimate displacement with accuracy similar to dedicated vision-based SHM setups. To establish operating limits, we quantify how temporal sampling (frame rate) and spatial sampling (video resolution) affect drift estimation accuracy. We also evaluate peak drift/IDR estimation accuracy and peak timing sensitivity under reduced temporal sampling. The results highlight the potential of widely available camera networks as a low-cost, scalable, and rapidly deployable sensing network for post-earthquake assessment. Full article

This research significantly improves the link reliability and robustness of optical camera communications (OCC) by leveraging deep learning for light source modulation filtering, reflection filtering, and precise light source localisation. By using image sensors as receivers in OCC, data transmission is not only enabled, but other applications are also facilitated, such as detecting objects and humans, making OCC highly attractive in healthcare, intelligent transport systems, and indoor positioning. However, the position of the desired signal in the received image frame must be tracked in dynamic scenarios (i.e., nonstationary applications), in order to maintain the communication link. Moreover, as sixth-generation (6G) wireless networks envision highly autonomous systems that rely on seamless integration of communication and sensing, deep learning is key to enabling robust and adaptive light source localisation and sensing in OCC, which enables vision-based autonomy in dynamic environments. It should be noted that a deep learning-based approach provides more accuracy even when there are multiple noise sources in the environment, reflections, and complex backgrounds, and under mobility conditions, in which traditional light source detection/tracking methods are not effective. Hence this study investigates the use of a deep learning-based approach by analysing the detection accuracy under different configurations and unseen images. The results obtained demonstrate consistently high detection performance with average precision (at an intersection-over-union threshold of 0.70 of 0.84 to 0.97. These results pave the way for autonomous receivers that will be able to select signals intelligently and decode them. Full article

Interactive teleoperation offers an intuitive pathway for human–robot interaction, yet many existing systems rely on dedicated sensors or wearable devices, limiting accessibility and scalability. This paper presents a vision-based teleoperation framework that enables real-time control of an articulated robotic arm (five joints plus a gripper actuator) using human hand tracking from a single, typical laptop camera. Hand pose and gesture information are extracted using a real-time landmark estimation pipeline, and a set of compact kinematic descriptors—palm position, apparent hand scale, wrist rotation, hand pitch, and pinch gesture—are mapped to robotic joint commands through a calibration-based control strategy. Commands are transmitted over a lightweight network interface to an embedded controller that executes synchronized servo actuation. To enhance stability and usability, temporal smoothing and rate-limited updates are employed to mitigate jitter while preserving responsiveness. In a human-in-the-loop evaluation with 42 participants, the system achieved an 88% success rate (37/42), with a completion time of 53.48 ± 18.51 s, a placement error of 6.73 ± 3.11 cm for successful trials (n = 37), and an ease-of-use score of 2.67 ± 1.20 on a 1–5 scale. Results indicate that the proposed approach enables feasible interactive teleoperation without specialized hardware, supporting its potential as a low-cost platform for robotic manipulation, education, and rapid prototyping. Full article

Human activity recognition (HAR) lies at the core of digital healthcare applications that monitor different types of physical activity. Traditional HAR methods often struggle to adapt to variable-length, real-world activity data and to generalise across cohorts (e.g., from young to old cohorts). Thus, the aim of this study was to investigate HAR using wearable sensor data, with a particular focus on cross-cohort evaluation. Each dataset included two accelerometers (right thigh and lower back) sampling at 50 Hz, capturing a range of daily-life activities that were annotated using video recordings from chest-mounted cameras synchronised with the accelerometers. Neural networks were trained on young cohorts’ data and tested on old cohorts’ data. The effects of network architecture, sampling frequency and sensor location on classification performance were investigated. Network performance was evaluated using accuracy, recall, precision, F1-score and confusion matrices. The gated recurrent unit architecture achieved the best performance when trained solely on young cohorts’ data, with weighted F1-score of 0.95 ± 0.05 and 0.93 ± 0.05 for young and old cohorts, respectively, resulting in a highly generalizable method. Classification performance across multiple sampling frequencies was comparable. The thigh-mounted sensor consistently achieved higher performance than the lower back sensor across activities except lying. Furthermore, combining datasets significantly improved performance on the old cohort (weighted F1-score: 0.97 ± 0.02) due to increased variability in the training data. This study highlights the importance of network architecture and dataset composition in HAR and demonstrates the potential of neural networks for robust, real-world activity recognition across age-defined cohorts, specifically between young and old cohorts. Full article

Eye movement classification, particularly the identification of fixations and saccades, plays a vital role in advancing our understanding of neurological functions and cognitive processing. Conventional modalities of data, such as RGB webcams, often face limitations such as motion blur, latency and susceptibility to noise. Neuromorphic Vision Sensors, also known as event cameras (ECs), capture pixel-level changes asynchronously and at a high temporal resolution, making them well suited for detecting the swift transitions inherent to eye movements. However, the resulting data are sparse, which makes them less well suited for use with conventional algorithms. Spiking Neural Networks (SNNs) are gaining attention due to their discrete spatio-temporal spike mechanism ideally suited for sparse data. These networks offer a biologically inspired computational paradigm capable of modeling the temporal dynamics captured by event cameras. This study validates the use of Spiking Neural Networks (SNNs) with event cameras for efficient eye movement classification. We manually annotated the EV-Eye dataset, the largest publicly available event-based eye-tracking benchmark, into sequences of saccades and fixations, and we propose a convolutional SNN architecture operating directly on spike streams. Our model achieves an accuracy of 94% and a precision of 0.92 across annotated data from 10 users. As the first work to apply SNNs to eye movement classification using event data, we benchmark our approach against spiking baselines such as SpikingVGG and SpikingDenseNet, and additionally provide a detailed computational complexity comparison between SNN and ANN counterparts. Our results highlight the efficiency and robustness of SNNs for event-based vision tasks, with over one order of magnitude improvement in computational efficiency, with implications for fast and low-power neurocognitive diagnostic systems. Full article