Search Results (2,083)

This paper presents the H-CoRE (Heterogeneous Cooperative Multi-Robot Execution) framework designed to enable autonomous multi-robot operations in GNSS-denied environments. Built on an ROS 2-based architecture, H-CoRE enables collaborative, structured task execution through standardized software stacks. Each robot’s stack combines a high-level executive system with an agent-specific motion layer and leverages multi-sensor fusion for localization and mapping. The framework is inherently reconfigurable, allowing individual agents to operate autonomously or as part of a multi-robot team for collaborative missions. In the considered scenario, the system integrates aerial and ground vehicles, a fixed pan–tilt–zoom camera, and a human supervisory interface within a unified, modular infrastructure. The proposed system has been deployed in indoor, GNSS-denied environments, demonstrating autonomous navigation, cooperative area coverage, and real-time information sharing across multiple agents. Experimental results confirm the effectiveness of H-CoRE in maintaining general awareness and mission continuity, paving the way for future applications in search-and-rescue, inspection, and exploration tasks. Full article

Older adults often need guidance when visiting new buildings for the first time. However, indoor navigation remains challenging due to the lack of Global Positioning System (GPS) availability, visually repetitive corridors, and frequent location failures. This article presents a multimodal indoor navigation assistant that combines graph-based route planning with visual landmark verification to provide step-by-step guidance. The environment is modelled as a directed graph whose nodes are annotated with semantic landmarks, and the graph is constructed primarily from a video of the building, reducing the need for 3D scanners, beacons, or other specialised instruments. Routes are calculated using Dijkstra’s shortest-path algorithm over the semantic graph. During navigation, camera frames are analysed using a restricted vision-language recognition strategy that only considers candidate landmarks from the current and next nodes, reducing false detections and improving interpretability. To increase robustness, a temporary voting mechanism was introduced to confirm node transitions, as well as a hierarchical redirection strategy with local and global recovery. The system is implemented in two modes: handheld mode with visual cues using augmented reality arrows, mini map and voice instructions, and hands-free mode with front camera using voice instructions and keywords. Evaluation involved preliminary technical testing in the United Kingdom followed by formal user validation in Spain. During these trials, participants reported high usability, strong confidence and safety, and increased perceived independence. Full article

Indoor CO₂ forecasting supports proactive ventilation control that balances air quality with energy efficiency. While Machine Learning (ML) models have shown strong performance in controlled settings such as schools, their generalization across indoor spaces with diverse occupancy dynamics remains poorly characterized. We present a systematic benchmark of 11 forecasting models spanning simple baselines, statistical methods, classical ML, deep learning, ensembles, and foundation models using 18 weeks of IoT sensor data spanning six real-world use cases: conference rooms, dining halls, hospitals, food markets, offices and student residences. Performance depends strongly on the prediction horizon and on the regularity of occupancy-driven CO₂ patterns. Simple baselines tend to perform best at short horizons (10 min ahead), while ensembles and fine-tuned foundation models provide more robust accuracy at longer horizons (4 h ahead). Remarkably, zero-shot foundation models demonstrate the ability to outperform trained classical models in data-scarce scenarios, challenging the traditional paradigm of localized training. These findings indicate that optimal forecasting strategies are context-dependent and challenge the assumption of universal model superiority. Full article

Assistive mobility devices for small animals require reliable monitoring to ensure safe and comfortable operation without increasing system complexity or invasiveness. This study presents a low-cost monitoring platform that integrates a laser-induced graphene (LIG) contact-temperature sensor into a passive mobility device for small dogs, supported by a lightweight Internet of Things (IoT) architecture. The system combines contact temperature, ambient temperature, speed, and obstacle distance using an energy-aware acquisition strategy and prioritized wireless transmission for near-real-time monitoring. An unsupervised anomaly detection framework based on Isolation Forest identifies potentially unsafe operating conditions without labeled pathological data by leveraging absolute temperature and the differential feature $\Delta T$ between contact and ambient measurements. Experimental validation was conducted under controlled indoor conditions across six independent sessions with a small-breed dog, including static and dynamic phases to ensure repeatability. The system achieved packet delivery ratios of approximately 95%, with typical end-to-end latencies below 500 ms and worst-case delays below 850 ms. The proposed approach detected localized thermal deviations associated with friction or prolonged contact while remaining robust to normal activity- and environment-driven variations. These results demonstrate the feasibility of integrating LIG-based sensing and unsupervised analytics into assistive animal mobility platforms to enhance safety through continuous, non-invasive monitoring. Full article

Visual localization plays a critical role for mobile robots to estimate their position and orientation in GPS-denied environments. However, its efficiency, robustness, and generalization are fundamentally undermined by severe viewpoint changes and dramatic appearance variations, which present persistent challenges for image-based feature representation and pose estimation under real-world conditions. Recently, map-free visual relocalization (MFVR) has emerged as a promising paradigm for lightweight deployment and privacy isolation on edge devices, while how to learn compact and invariant image tokens without relying on structural 3D maps still remains a core problem, particularly in highly dynamic or long-term scenarios. In this paper, we propose the Debiased Multiplex Tokenizer as a novel method (termed as DMT-Loc) for efficient and versatile MFVR to address these issues. Specifically, DMT-Loc is built upon a pretrained vision Mamba encoder and integrates three key modules for relative pose regression: First, Multiplex Interactive Tokenization yields robust image tokens with non-local affinities and cross-domain descriptions. Second, Debiased Anchor Registration facilitates anchor token matching through proximity graph retrieval and autoregressive pointer attribution. Third, Geometry-Informed Pose Regression empowers multi-layer perceptrons with a symmetric swap gating mechanism operating inside each decoupled regression head to support accurate and flexible pose prediction in both pair-wise and multi-view modes. Extensive evaluations across seven public datasets demonstrate that DMT-Loc substantially outperforms existing baselines and ablation variants in diverse indoor and outdoor environments. Full article

When is an unknown, confined environment traversable for a specific ground robot using only touch? We answer by (i) giving an environment-anchored definition of traversability, expressed through the max-min value $T^{★}(E;A)={sup}_{\pi \in {\Pi }_{S\to G}}{inf}_{s\in [0,1]}\varphi \left(\pi \left(s\right)\right)$, where the bottleneck margin $\varphi$ aggregates the clearance, curvature ($\rho \ge R_{min}$), slope/step, and friction constraints, and (ii) introducing an on-policy, tactile certificate (TC) that maintains a conservative, monotone lower bound $T_t$ using partial contact histories. The TC fuses pessimistic free-space from contacts and the body envelope, the M3 decaying contact memory as a risk prior, and local bend/FSR proxies; a certificate is issued when $T_t>0$ and the explored corridor graph connects S to G. Relative to Papers 1–2 (tactile traversal; offline software assurance), this work formalizes traversability itself and provides a tactile-only, online certificate computable during runs. In a retrospective analysis of 660 trials across Indoor/Outdoor/Dark lighting environments, (H1) the early TC margin predicts success and traversal time better than contact/dwell heuristics (higher AUC/$R^2$), (H2) the TC predictivity is lighting-invariant, and (H3) speed-gating M3 by a TC margin recovers part of the CB-V speed gap without degrading success. Artifacts include the TC implementation, explored-corridor graphs, and per-trial TC time series added to the Paper-1 log bundle; these materials are available from the corresponding author upon reasonable request. Full article

To address the topological mismatch between signal space and physical space caused by uneven signal feature distribution in indoor non-line-of-sight and complex topological environments, this paper proposes an indoor positioning algorithm based on Energy-domain Fingerprint Manifold Graph Total Variation (EFM-GTV). To mitigate neighborhood distortion caused by uneven high-dimensional signal feature distribution, a UMAP manifold topology graph construction method based on fuzzy simplicial sets is designed to establish a graph basis consistent with physical space topology. To reduce false matching risks in global search, a physical topology pruning strategy combining Jaccard similarity is proposed, effectively eliminating pseudo-connections. Building upon this foundation, we introduced an optimization model based on graph total variation, reformulating the positioning problem as a graph signal recovery task. This approach effectively overcomes signal fluctuation interference in complex topologies like U-shaped corridors, achieving robust position estimation. Experiments demonstrate that this algorithm effectively leverages manifold structure constraints to correct NLOS errors. On real-world field test datasets, compared to traditional weighted algorithms, the average positioning accuracy improves to 1.4267 m, with maximum positioning error reduced by over 50%, achieving high-precision robust positioning. Full article

Localization and navigation techniques have become fundamental for modern lives, while achieving accurate results indoors still remains a significant challenge. The widespread adoption of smart devices, and especially smartphones, has increased the need for accurate and robust pedestrian dead reckoning systems that operate in infrastructure-less environments. Pedestrian dead reckoning’s primary challenge is maintaining accuracy despite varying smartphone placements (attitudes) and the noisy, low-cost inertial measurements units. In this work, a comprehensive pedestrian dead reckoning framework is presented that integrates advanced step counting and heading estimation techniques. For step detection and counting, we propose a robust step counting algorithm that utilizes the optimum fusion of the raw IMU readings, i.e., accelerometer, linear accelerometer, gyroscope, and magnetometer readings, each broken down into three degrees of freedom for different body placements and walking speeds. Furthermore, to address the critical issue of heading estimation, we propose the heading estimation axis mapping (HEAT-MAP) algorithm, which dynamically adjusts the sensor axes in response to the smartphone’s orientation, ensuring a consistent coordinate frame and reducing heading drift. Moreover, to eliminate cumulative pedestrian dead reckoning errors, the system incorporates an adaptive weighted fusion mechanism with Wi-Fi fingerprinting. Experimental results demonstrate that this integrated system significantly improves the overall trajectory accuracy, providing a high-precision, attitude-unconstrained solution for real-time indoor pedestrian navigation. Full article

High noise levels in industrial workplaces pose significant challenges to occupational safety, particularly with hearing protection and effective communication. Traditional hearing protection devices, while effectively attenuating harmful noise, often compromise situational awareness by excessively isolating workers from the acoustic environment and preventing the perception of critical auditory cues (e.g., emergency alarms), thereby introducing additional safety risks. This paper presents a smart industrial safety system that integrates Internet of Things (IoT) and artificial intelligence (AI) and is based on intelligent hearing protection devices to (a) selectively attenuate hazardous industrial noise while (b) preserving human speech and (c) reproduce targeted audio notifications to workers near malfunctioning or hazardous machinery. A real-time voice activity detection (VAD) model is employed to distinguish vocal components from background noise to adaptively control digital signal processing filters. Furthermore, indoor localization enables the delivery of targeted audio messages to workers in proximity to relevant events. Experimental evaluations on embedded hardware demonstrate that the selected VAD model operates well within real-time constraints and effectively supports dynamic noise filtering. Objective evaluation of the filtering stage using Mean Opinion Score (MOS), signal-to-noise ratio (SNR), and Harmonics-to-Noise Ratio (HNR) shows consistent quality improvements across all tested conditions, with MOS gains up to +118%, SNR increases between +10.4 and +29.0 dB, and HNR improvements up to +6.22 dB, indicating enhanced speech intelligibility and preservation of voice harmonic structure even under high-noise scenarios. Robustness validation of the VAD module across varying acoustic conditions confirms reliable speech detection performance, achieving perfect classification at +10 dB SNR, very high accuracy at 0 dB (98.3%, ROC AUC 0.998), and stable operation even at $-7$ dB SNR (79.8% accuracy, ROC AUC 0.878). The proposed architecture achieves a balanced trade-off between hearing protection and speech intelligibility while enhancing the effectiveness of safety communications in noisy industrial environments. Full article

V2X communication is crucial for intelligent connected vehicles, but suffers from multipath fading. In existing studies, V2X channel modeling primarily employs 2D models or follows the models specified in 3GPP TR36.885. The reliability has not been verified, and it cannot reflect the small-scale fading of multipath fading. There are also problems, such as easy local optimality and slow convergence in model parameter optimization. Therefore, based on the V2V 3D channel model, this paper uses the K-Means++ algorithm to obtain the main category data, takes the main category data as the input, and then uses the genetic algorithm (GA) to perform multi-parameter optimization of the reflection point M and reflection coefficient $\mu$ of six scenarios to obtain the optimal parameters. Based on the China Communications Standards Association (CCSA)’s white paper, the Rice factor K was determined. In-chamber and on-road comparison tests were designed to verify the rationality of the parameters optimized by the GA. The experiments show that this model and method can accurately reproduce the characteristics of V2V channels, support the setting of indoor V2X test parameters, and provide a standardized solution for the verification of V2V communication performance. Full article

Indoor pedestrian navigation tasks, as a key part of smart cities and navigation services, face dual challenges of accuracy and cost under complex building environments. Currently, neural inertial navigation is at the vanguard of current research in indoor pedestrian navigation, and existing related studies have achieved positive results. However, the exploration of deep learning solutions is still not sufficient, mainly reflected in the lack of explorations of model training configurations. Based on testing results under different deep learning schemes, this paper proposes EDIN, an enhanced deep inertial navigation approach. This method benefits from a proprietary neural network based on ResNeXt with Convolutional Block Attention Module (CBAM) to predict the relationship between inertial data and motion trajectory. Compared to existing projects, this paper also makes improvements in the model training process, thereby improving the predictive effect of the trained model. Specifically, this paper innovatively uses Logcosh as the loss function and combines data rotation and additional noise as data augment methods. To assess EDIN’s performance, extensive tests were conducted using three publicly available datasets: RoNIN, OXIOD, and RIDI. The results clearly indicate EDIN’s superior performance relative to other neural inertial navigation systems. Notably, localization accuracy improved significantly, with an average enhancement of 16.06% compared to the RoNIN-ResNet method. Full article

Climate change is a critical global challenge, and the building sector accounts for nearly 30% of global greenhouse gas (GHG) emissions, remaining a key target for mitigation. Indoor environments contribute significantly to GHG emissions, primarily through heating, cooling, lighting, and occupant-driven energy use. Indoor mapping, serving as the foundation for Digital Twins (DTs), provides a spatiotemporal framework that integrates sensor data with Building Information Modelling (BIM), Geographic Information Systems (GIS), and Internet of Things (IoT) to support energy-efficient, low-carbon building operations. This review examined the role of indoor mapping in understanding, modelling, and reducing GHG emissions in buildings. It synthesized current advancements in indoor spatial data acquisition, ranging from Light Detection And Ranging (LiDAR) and Simultaneous Localization and Mapping (SLAM) to deep learning-based floor plan extraction, and evaluated their contribution to improved indoor environmental analysis. The review highlighted emerging techniques, challenges, and gaps, particularly the limited integration of physical indoor spaces with virtual layers representing assets, occupants, and equipment. Addressing this gap requires embedding spatial modelling as an intermediate analytical layer that structures and contextualizes sensor data to support spatiotemporal decision-making. Overall, this review demonstrated that indoor mapping plays a critical role in transforming spatial information into actionable insights, enabling more accurate energy modelling, enhanced real-time building management, and stronger data-driven strategies for GHG mitigation in the built environment. Full article

The evaluation of the seismic behavior of masonry aggregates, which characterize Italian historic centres, is a challenging and widely debated topic in the field of structural engineering. These constructions, composed of several adjacent structural units, tend to exhibit both global and local damage when subjected to horizontal seismic actions—loads that were not considered at the time of their original construction. Developed over centuries of unplanned urban growth, they are based on empirical construction rules and locally sourced materials. Due to their poor thermal properties, these buildings are also affected by significant heat losses, resulting in reduced indoor comfort. In this context, the present study aims to evaluate the seismic performance of a masonry aggregate and two of its constituent structural units located in Visso, in the province of Macerata, an area severely affected by the 2016 Central Italy seismic sequence, both before and after the application of an innovative integrated retrofitting solution. The proposed strengthening system combines aluminium alloy exoskeleton with insulating sandwich panels, simultaneously addressing seismic vulnerability and energy inefficiency. The assessment is carried out through numerical analyses, including nonlinear static and dynamic approaches, to achieve a comprehensive understanding of the structural response. Moreover, a comparative analysis between the masonry aggregate and the two individual structural units, modelled as isolated buildings, is performed to investigate the influence of structural interaction among adjacent units. The results demonstrate the effectiveness of the proposed retrofitting strategy, highlighting a significant improvement in global stability. Furthermore, the comparison confirms the critical role of inter-unit interaction and underscores the necessity of modelling historic masonry aggregates rather than isolated buildings to obtain a more realistic seismic performance evaluation. Full article

Inertial localization is particularly valuable in GPS-denied environments such as indoors. However, localization using only Inertial Measurement Units (IMUs) suffers from drift caused by motion-process noise and sensor biases. This paper introduces Uncertainty-aware Map-constrained Inertial Localization (UMLoc), an end-to-end framework that jointly models IMU uncertainty and map constraints to achieve drift-resilient positioning. UMLoc integrates two coupled modules: (1) a Long Short-Term Memory (LSTM) quantile regressor, which estimates the specific quantiles needed to define $68\%$, $90\%$ and $95\%$ prediction intervals serving as a measure of localization uncertainty and (2) a Conditioned Generative Adversarial Network (CGAN) with cross-attention that fuses IMU dynamic data with distance-based floor-plan maps to generate geometrically feasible trajectories. The modules are trained jointly, allowing uncertainty estimates to propagate through the CGAN during trajectory generation. UMLoc was evaluated on three datasets, including a newly collected 2-h indoor benchmark with time-aligned IMU data, ground-truth poses and floor-plan maps. Results show that the method achieves a mean drift ratio of $5.9\%$ over a $70\mathrm{m}$ travel distance and an average Absolute Trajectory Error (ATE) of $1.36\mathrm{m}$, while maintaining calibrated prediction bounds. Full article

Crowdsensing has been recently investigated as an incorporation of Human-Machine intelligence in which contribution of users is crucial. Indoor localization is one of the significant applications among divers applications that have been introduced in this area. Considering the slight infiltration of GPS signals in indoor environments crowdsensing and its promising indoor localization schemes have been utilized for providing precise localization services. Precision of crowdsensing indoor localization schemes and elimination of erroneous data collection is strongly dependent on the underlying task allocation mechanism. In this work, we have approached the localization precision as a consequence of task allocation mechanism of crowd-powered indoor localization schemes. Hence, we have proposed to tackle this issue by applying GWO (Gray Wolf Optimizer) algorithm on participants of crowdsensing scheme. It is expected that the GWO algorithm implicitly performs the task allocation procedure in account of its crowd-powered nature. Accordingly, we have applied GWO algorithm on a proposed indoor localization scenario to undertake the requirements for discrete task allocation mechanism. Implementation results demonstrated that the population-centric structure of the GWO algorithm significantly increments the accuracy of fingerprint collection mechanism which maintains an exceptional localization precision. Full article