The latest advancements in visible light communication (VLC) technology have greatly facilitated the evolution of wireless sensing based on visible light, commonly referred to as visible light sensing (VLS) systems. Similar to other wireless sensing technologies, these systems can be categorized into two primary types: device-based (or active) and device-free (or passive), depending on whether targets, such as people or objects of interest, are equipped with an optical receiver (e.g., a photodiode or camera). By reusing the existing Light-emitting diode (LED) lighting infrastructure as transmitters in indoor environments, device-free VLS (DF-VLS) systems can help address the growing energy demands of ubiquitous sensing and communication in Sixth-Generation Wireless Networks (6G). This survey meticulously explores the landscape of DF-VLS technology, proposing a redefinition of the classification of VLS systems and delving into the existing literature in this domain. We conduct a comprehensive analysis of various DF-VLS sensing techniques, focusing on their diverse applications in the Internet of things (IoT). By closely examining the integration and utility of DF-VLS in IoT environments, we highlight the potential of this technology. Furthermore, we underscore critical open challenges that demand attention for the effective development of efficient DF-VLS systems. Finally, the paper outlines a future roadmap for DF-VLS systems, shedding light on potential directions this field may embark upon.

Private 5G mobile networks are emerging as a platform for wireless connectivity in professional applications across smart industrial sectors such as automated warehousing, logistics, autonomous vehicle deployments in campus environments, mining, and material processing, among others. It is expected that most Machine-to-Machine (M2M) and Industrial Internet of Things (IIoT) communication links will increasingly rely on wireless solutions, as the flexibility they offer provides clear advantages over hard-wired network installations. To gain insight into workers’ exposure to radiofrequency electromagnetic fields (RF EMF) emitted by 5G private mobile networks, an analysis was conducted based on measured and calculated RF EMF levels from various 5G private networks in real-world scenarios across different smart industrial sectors and R&D platforms in three countries. Several exposure scenarios were evaluated, including production facilities, logistics operations, office environments, and research sites. The installations included different configurations: private standalone and non-standalone 5G networks operating at 3.5 GHz and 26 GHz, as well as public networks with private slicing. The results clearly demonstrated that exposure levels in all investigated scenarios were well below existing exposure limits. In a typical indoor industrial environment where pico 5G base stations are deployed, the measured exposure was found to be no greater than 0.006% of the Directive 2013/35/EU action value and 0.03% of the ICNIRP guideline limits for the general public.

Ultra-wideband (UWB) radar positioning plays an important role in non-cooperative personnel positioning or device-free positioning. Recent work has reformulated the time-of-flight (ToF) estimation task as a 2-D image processing problem using residual convolutional neural networks (RCNNs), avoiding intricate procedures of traditional methods. Despite its benefits, the RCNN struggles with fully exploiting feature reutilization, resulting in significant errors in ToF estimation. Although particle filters (PFs) can alleviate this problem, the constant parameters in weight estimation will affect the positioning accuracy. Therefore, we first adopt a dense convolutional network (DenseNet) to replace the RCNN to enhance the feature reutilization and improve the accuracy of ToF estimation. Additionally, we design a method for outlier and anomaly cluster elimination in the ToF time series based on density-based spatial clustering of applications with noise (DBSCAN) clustering, effectively suppressing observation noises. Finally, to address the insufficient adaptability of parameters in PF weight estimation, we improve the loss function in the DenseNet, thereby enabling it to dynamically output the variance of ToF. We verify the effectiveness and generalizability of our proposed method through an open-source dataset collected with low-cost UWB devices. Compared with a classic method, the average root mean square error (RMSE) of the proposed method within the positioning area decreases by 37.1%. Furthermore, through repeated experiments across three distinct scenarios, our method demonstrates RMSE reductions of 20.1%, 36.5%, and 28.5%, respectively, compared to an existing RCNN-based approach.

The sub-terahertz (sub-THz) band offers immense bandwidth for next-generation wireless networks. However, understanding its complex propagation characteristics remains a significant challenge. Furthermore, real multi-user multiple-input multiple-output (MIMO) channel measurements above 100 GHz are currently scarce due to hardware and cost constraints. To bridge this gap, this paper investigates sub-THz multi-user MIMO channel modeling at 140 GHz in an indoor hotspot scenario using a validated ray tracing framework. By incorporating deterministic wave propagation, the Beckmann-Kirchhoff diffuse scattering model, and general environmental modeling, the proposed sub-THz ray tracer is able to generate consistent channel parameters with measurement data. Using an array-of-subarrays (AoSA) hybrid beamforming architecture with 16 radio frequency (RF) chains, we analyze the impact of different antenna element beamwidths on the analog beamforming performance. Results demonstrate that wider antenna element beamwidths are critical for maintaining robust analog beamforming performance. It is found that digital beamforming sum rates also benefit significantly from wide antenna element beamwidths, as well as distributed MIMO configurations. Although important for sub-THz propagation channel modeling, diffuse scattering is shown to have no impact on the beamforming performance under the studied beamforming configuration in LOS scenarios. The findings in this paper provide insights for future wireless system design over 100 GHz.

User orientation is crucial for many context-aware applications, including interactive museum experiences, smart door access, and intuitive human-environment interaction. However, most existing indoor localization systems focus on estimating position, while body orientation is typically assigned to secondary devices such as inertial measurement units. In this paper, we propose a purely UWB-based approach that predicts yaw orientation directly from UWB Channel Impulse Response (CIR) measurements recorded at fixed anchors as they receive transmissions from a single wearable tag. We use a bidirectional Mamba architecture that captures dependencies across the anchor observations through forward and backward recurrent scans. The model uses per-anchor CIR and a body-part conditioning module to adapt the representation to different tag placements on the body. Two different Kalman filters are used as post-processing stages to exploit temporal continuity: an orientation-based filter that smooths the neural network predictions, and a location-based filter that additionally incorporates position-derived heading corrections. We evaluated the model's performance in different scenarios to ensure generalizability. The proposed Mamba model achieves a mean absolute error of 38.6 degrees in its raw form, outperforming a rule-based baseline of 49.5 degrees. With the location-based Kalman filter, the error is further reduced to 18.9 degrees, corresponding to a 51

This study describes a neural network-based method for estimating exposure levels in industrial environments, without requiring detailed technical inputs, allowing usage of the model by layman people or by workers active in these areas. A pipeline based on Blender environments and MATLAB ray-tracing simulations is created and after defining a set of 11 candidate input parameters for the model, more than 20 000 different wireless configurations are simulated, varying the different environmental and wireless input parameters. A correlation analysis shows that main inputs influencing the exposure levels in the industrial area are the transmit power of the antennas, the density of clutter in the area, the density of transmitters in the area, and the height and location of the transmitters. A multi-layer fully connected neural network regression model is developed to predict median (E50) and 95th percentile (E95) exposure levels in industrial areas. Testing the obtained model on an unseen dataset of environments withE50values between 0 and 3.25 V m-1andE95values between 0 and 7 V m-1, demonstrates the good prediction performance of the model: root-mean-square error values below 0.173 V m-1andR2values above 95% are obtained. Subsequently, the model is validated with measurement data collected in three distinct realistic industrial environments. The average absolute deviation of the model predictions with respect to the measurements is limited to 20.4%. This novel and broadly accessible approach demonstrates that it is possible to reliably estimate exposure levels in realistic environments without having to rely on external experts or on dedicated complex software.

The rising demand for location-based services (LBSs) is accelerating the deployment of indoor positioning systems (IPSs) and driving the need for stricter quality of positioning service (QPoS) requirements. Leveraging existing illumination infrastructure, visible light positioning (VLP) is a promising low-cost candidate for centimeter- to decimetre-level localization, but its real-world QPoS-spanning accuracy, availability, latency, robustness, scalability, energy use, and cost-is not yet well characterized. Existing IPS literature, including VLP studies, mostly prioritizes accuracy, and no comprehensive, literature-based QPoS comparison of positioning technologies currently exists. This article presents a focused review of VLP from a broad QPoS perspective and situates it within the IPS landscape. We propose a reproducible, normalized QPoS framework to enable holistic, fair cross-study comparisons. Through a review of published experimental and simulation studies, we analyze how different classes of VLP approaches perform with respect to QPoS and position them relative to other IPS technologies, including ultrawideband (UWB), Wi-Fi/Bluetooth low energy (Wi-Fi/BLE), and vision/inertial measurement unit (vision/IMU) hybrids. It is shown that VLP can exceed the accuracy of many IPS technologies and can approach UWB-level performance, with a favorable overall QPoS under line-of-sight conditions and a preexisting suitable lighting infrastructure. Nevertheless, its suitability remains application-specific, as no IPS technology-including VLP-optimizes all QPoS dimensions simultaneously, and performance varies with deployment and calibration. Our review highlights several unresolved challenges in VLP research, including nonstandardized QPoS metrics, a lack of reference testbeds and protocols, limited 2.5-D/3-D and mobility evaluations under realistic multipath and industrial lighting, hardware and transmitter limitations, ambient-light interference, shadowing and multipath effects, and unresolved latency-reliability tradeoffs. To address these challenges, we propose future research on exploiting VLP signals of opportunity, optimizing transmitter planning, reducing driver/modulator cost, and characterizing hardware nonidealities, advancing multi-photodiode (PD) receiver studies, hybridizing with other systems, exploring reconfigurable intelligent surfaces (RISs), and developing standardized benchmark suites and protocols.

This paper explores the use of a commercial-off-the-shelf impulse-radio ultra-wideband transceiver with 0.5 GHz bandwidth for ranging and respiration rate monitoring of a moving target. Respiration causes the chest to change in shape, which alters the target’s radar cross section, which causes periodical changes in the channel impulse response. A median filter, Viterbi algorithm and a particle filter are used to determine the range of the target, band-pass filtering and fourier transforming are used for respiration rate estimation. The results demonstrate that accurate localization and respiration rate monitoring is possible under controlled settings, but not in a realistic scenario.

5G private mobile networks are becoming a platform for ‘wire-free’ networking for professional applications in smart industry sectors, such as automated warehousing, logistics, autonomous vehicle deployments in campus environments, mining, material processing, and more. It is expected that most of these Machine-to-Machine (M2M) and Industrial Internet of Things (IIoT) communication paths will be realized wirelessly, as the advantages of providing flexibility are obvious compared to hard-wired network installations. Unfortunately, the deployment of private 5G networks in smart industries has faced delays due to a combination of high costs, technical challenges, and uncertain returns on investment, which is reflected in troublesome access to fully operational private networks. To obtain insight into occupational exposure to radiofrequency electromagnetic fields (RF EMF) emitted by 5G private mobile networks, an analysis of RF EMF due to different types of 5G equipment was carried out on a real case scenario in the production and logistic (warehouse) industrial sector. A private standalone (SA) 5G network operating at 3.7 GHz in a real industrial environment was numerically modeled and compared with in situ RF EMF measurements. The results show that RF EMF exposure of the workers was far below the existing exposure limits due to the relatively low power (1 W) of indoor 5G base stations in private networks, and thus similar exposure scenarios could also be expected in other deployed 5G networks. In the analyzed RF EMF exposure scenarios, the radio transmitter—so-called ‘radio head’—installation heights were relatively low, and thus the obtained results represent the worst-case scenarios of the workers’ exposure that are to be expected due to private 5G networks in smart industries.

Ultrawideband (UWB) is a high-precision positioning and navigation technology, it faces significant challenges due to the abundance of non-line-of-sight (NLOS) conditions in complex indoor environments. In this study, we introduce the bidirectional encoder representations from transformers (BERTs) to identify and mitigate the impact of NLOS paths using the channel impulse response (CIR). We derive three new CIR features that comprise both the time and energy characteristics of CIR sequences. These proposed features are fused with fuzzy probabilities into BERT (F-BERT), in order to identify the NLOS paths. Based on the NLOS identification results from F-BERT, a ranging classification and mitigation strategy with another BERT is further designed to enhance the ranging and positioning accuracy. The experimental results indicate that F-BERT outperforms state-of-the-art algorithms such as least-squares support vector machine (LS-SVM), convolutional neural network (CNN), and CNN with long short-term memory (CNN-LSTM) by 12.5%, 13.9%, and 14.9%, respectively, in terms of NLOS identification accuracy with LOS and NLOS recall. The proposed BERT also outperforms the existing algorithms by 36.2% in ranging error reduction in an NLOS environment. Furthermore, our proposed algorithms similarly outperform existing algorithms in mean positioning accuracy by 37.9%. Finally, our BERT algorithms achieve generality as, although they were trained in one environment, they are shown to still work well in another unknown environment.

This paper presents D-band channel modeling campaigns for short-range line-of-sight (LOS) communications in representative data center and industrial environments. An accurate multipath estimation framework, exploiting measured antenna radiation patterns, is utilized to extract multipath components (MPCs) with 0.1° angular resolution for precise large- and small-scale channel characterization. In the data center, directional inter-rack communication links exhibit a high path loss exponent (PLE) of 2.39 and a low mean delay spread of 1.32 ns, with cable obstructions increasing the PLE to 3.24 but further decreasing the mean delay spread to 0.59 ns. The omnidirectional measurement in the data center shows significant clustering effects in spatial-temporal domain. In contrast, industrial machine-to-machine (M2M) communication links, characterized through omnidirectional measurements, exhibit a lower PLE of 1.67. Mean values for delay spread, angular spread, and K-factor in industrial environments are moderate, which are 4.35 ns, 16.46° , and 8.11 dB, respectively, with weak clustering effects. The measurement results are compared and discussed with the literature across different environments and frequency bands for an overall understanding of unique propagation characteristics in these two environments. The quantitative measurement results contribute to D-band channel standardization.

In the rapidly evolving Internet of Things (IoT) landscape, accurate indoor positioning is increasingly vital. The proposed algorithm synergizes an ultrawideband (UWB) sensor with an inertial measurement unit (IMU) and artificial intelligence to obtain precise positioning in nonline-of-sight (NLOS) scenarios. In the proposed UWB module, a large language model (LLM) such as bidirectional encoder representations from transformers (BERT) algorithm is designed to utilize the channel impulse response (CIR) for effective NLOS identification and UWB ranging trustworthiness evaluation. Concurrently, the IMU module is also designed with BERT to recognize various pedestrian activity states, thereby optimizing positioning. BERT's self-attention mechanism and deep learning (DL) bidirectional training efficiently extract essential features from sequential data, capturing both local and global information. The integration of both UWB and IMU through a proposed tightly coupled algorithm significantly boosts positioning performance. Experimental campaigns demonstrate an average NLOS identification accuracy, line-of-sight (LOS), and F2 of 98.8%, 99.4%, and 0.9926, respectively. These performances surpass the state-of-the-art least-squares support vector machine (LS-SVM), convolutional neural network (CNN), and CNN with long short-term memory (CNN-LSTM) up to 17.66% in NLOS identification. In terms of pedestrian activity recognition using BERT, the BERT algorithm achieves a precision (recall) of 99.3% (99.4%), notably outperforming the CNN and CNN-LSTM by 17.9% (16.2%) and 11.9% (10.9%), respectively. Finally, the UWB-IMU algorithm significantly enhances positioning accuracy by 80.5%, outperforming Kalman, LSTM-EKF, and particle filter (PF) methods by 68.1%, 48.3%, and 45.0%, respectively. The proposed approach presents a robust solution for indoor positioning for IoT applications, particularly in challenging NLOS environments.

UWB ranging measurements are prone to positive bias errors in NLOS conditions. Existing algorithms often use a uniform model to correct various ranging measurement errors, which has poor performance and fails to fully utilize the advantages of UWB technology. This article proposes a credibility evaluation system for assessing the quality of ranging measurement and the error fine-grained classification. For different sources of ranging measurement errors, we develop optimization models that integrate BERT with LSTM, focusing on both temporal and energy information, and extending the correction range to LOS measurements. In multi-scenario experiments, the credibility evaluation error was 0.047, with LOS and NLOS ranging measurement errors reduced by 37.9% and 80.94%, respectively. This approach outperforms LS-SVM, CNN, and CNN-LSTM by 57.21%, 43.2%, and 18.99%. After optimization, centimeter-level positioning accuracy improved by 30.2%, and the average error was reduced by 69.68%, surpassing advanced algorithms by 42.53%, 43.61%, and 44.00%.

This study assesses the exposure to 5G radio frequency electromagnetic fields (RF EMF) across four European countries. Spot measurements were conducted indoor and outdoor in both public spaces and educational institutions, encompassing urban and rural environments. In total, 146 measurements were performed in 2023, divided over Belgium (47), Switzerland (38), Hungary (30) and Poland (31). At 34.9% of all measurement locations a 5G connection to 3.6 GHz was established. The average cumulative incident power density (Savg) and maximum cumulative incident power density (Smax) were determined, for both "background" exposure (no 5G user equipment; No UE) and worst-case exposure (maximum downlink with 5G user equipment; Max DL). Furthermore, 3.6 GHz 5G-specific average Savg,5G and maximum Smax, 5G incident power density are considered as well. For the No UE scenario, the highest Smax is 17.6 mW/m2, while for the Max DL, the highest Smax is 23.3 mW/m2. Both values are well within the ICNIRP guidelines. The highest Smax, 5G measured over all countries and scenarios was 10.4 mW/m2, which is 3.2% of the frequency-specific ICNIRP guidelines. Additionally, a comparison was made between big cities, secondary cities, and villages for all four countries. The ratio of power density measured in rural areas was significantly lower than in urban areas (-4.8 to -10.4 dB). Under LOS conditions, the average incident power density was 2.3 mW/m2, whereas under NLOS conditions, the average incident power density decreases to 0.9 mW/m2. Furthermore, the relative variation increases under NLOS scenarios. Lastly, an analysis was performed regarding the power density in educational institutions compared to all other measurement locations, both indoors and outdoors for the different city types. The measured incident power density is not extensively lower in or around schools compared to public places, neither in the big cities, secondary cities, or the villages.

The rapid growth of wireless services has intensified spectrum scarcity, yet paradoxically, studies indicate that spectrum remains underutilized. To address this inefficiency, Dynamic Spectrum Access (DSA) systems have emerged as promising solutions. In particular, the use of TV White Space (TVWS) systems has shown potential for extending connectivity, e.g., via IEEE 802.22 standard-based networks. However, there is limited literature specifically examining the coexistence of IEEE 802.22 networks with Digital Terrestrial Multimedia Broadcast (DTMB) served areas. This paper seeks to address this gap to establish a reference framework for deploying IEEE 802.22 services alongside DTMB broadcasts. The study focuses on determining the necessary protection ratios for optimal performance of both standards and identifying transmission modes that minimize interference between them.

Visible Light Positioning (VLP) has emerged as a promising technology for next-generation indoor positioning systems (IPS), particularly within the scope of sixth-generation (6G) wireless networks. Its attractiveness stems from leveraging existing lighting infrastructures equipped with light-emitting diodes (LEDs), enabling cost-efficient deployments and achieving high-precision positioning accuracy in the centimeter-todecimeter range. However, widespread adoption of traditional VLP solutions faces significant barriers due to the increased costs and operational complexity associated with modulating LEDs, which consequently reduces illumination efficiency by lowering their radiant flux. To address these limitations, recent research has introduced the concept of unmodulated Visible Light Positioning (uVLP), which exploits Light Signals of Opportunity (LSOOP) emitted by unmodulated illumination sources such as conventional LEDs. This paradigm offers a cost-effective, lowinfrastructure alternative for indoor positioning by eliminating the need for modulation hardware and maintaining lighting efficiency. This paper delineates the fundamental principles of uVLP, provides a comparative analysis of uVLP versus conventional VLP methods, and classifies existing uVLP techniques according to receiver technologies into intensity-based methods (e.g., photodiodes, solar cells, etc.) and imaging-based methods. Additionally, we propose a comprehensive taxonomy categorizing techniques into demultiplexed and undemultiplexed approaches. Within this structured framework, we critically review current advancements in uVLP, discuss prevailing challenges, and outline promising research directions essential for developing robust, scalable, and widely deployable uVLP solutions.

Indoor localization based on Bluetooth Low Energy (BLE) is traditionally implemented by matching Received Signal Strength (RSS) fingerprints of nearby BLE nodes. Depending on the node density, pure RSS-based BLE localization only provides room-level or zone-level accuracies. For pedestrian tracking, BLE-based RSS fingerprinting is often fused with Pedestrian Dead Reckoning (PDR) using Inertial Measurement Units (IMU), which can provide up to 1-2 m accuracy. Recently, a phase-based BLE ranging method has been developed, which can accurately measure the distance between two BLE devices. Initial experiments on a moving platform showed promising results for accurate localization. In this work, the feasibility of this technology for pedestrian tracking is evaluated. On-body phase-based BLE and IMU measurements are performed in an industrial lab environment with four BLE anchors. A hybrid Particle Filter (PF) algorithm is designed, which fuses phase-based BLE ranging, PDR, and a range correction algorithm. Our proposed PF algorithm achieves a median and p75 error of 0.70 m and 1.03 m respectively, which outperforms traditional RSS-based (hybrid) BLE localization algoritms.

Knowledge of chicken density in different zones in the barn is an essential aspect of poultry welfare management, providing information on preferred zones, mobility, and influencing chicken growth, health, and economic returns for farmers. Although image and video processing techniques for chicken density estimation are commonly employed nowadays, they are insufficient for poultry environments due to challenges such as occlusions in crowded areas and the high computing costs of real-time analysis. This study aims to propose a novel, low-cost, and low-complexity visible light sensing (VLS) system specifically designed for chicken sensing, counting, and density estimation. Through the utilization of photodiodes (PDs) co-located with the luminaire and a device-free sensing (DFS) approach, the proposed system accurately identifies, counts, and estimates chicken density by capturing reflections from their bodies. The dataset used in this study consists of received signal strength (RSS) data collected from real farms, including recordings under various lighting conditions and scenarios with different chicken densities (e.g., up to 13 large and 25 small white chickens). In real-farm scenarios, the VLS system achieves a counting accuracy of more than 90% within a small detection zone of 0.07 m(2) under the light-emitting diode (LED). Over a larger area of 1.32 m(2), the system achieved a counting accuracy of 87% and 82% and a one-off accuracy of 95% and 90% for small and large chickens, respectively.