ABSTRACT
Human Activity Recognition (HAR) has emerged as a critical research area due to its extensive applications in various real-world domains. Numerous CSI-based datasets have been established to support the development and evaluation of advanced HAR algorithms. However, existing CSI-based HAR datasets are frequently limited by a dearth of complexity and diversity in the activities represented, hindering the design of robust HAR models. These limitations typically manifest as a narrow focus on a limited range of activities or the exclusion of factors influencing real-world CSI measurements. Consequently, the scarcity of diverse training data can impede the development of efficient HAR systems. To address the limitations of existing datasets, this paper introduces a novel dataset that captures spatial diversity through multiple transceiver orientations over a high dimensional space encompassing a large number of subcarriers. The dataset incorporates a wider range of real-world factors including extensive activity range, a spectrum of human movements (encompassing both micro-and macro-movements), variations in body composition, and diverse environmental conditions (noise and interference). The experiment is performed in a controlled laboratory environment with dimensions of 5 m (width) × 8 m (length) × 3 m (height) to capture CSI measurements for various human activities. Four ESP32-S3-DevKitC-1 devices, configured as transceiver pairs with unique Media Access Control (MAC) addresses, collect CSI data according to the Wi-Fi IEEE 802.11n standard. Mounted on tripods at a height of 1.5 m, the transmitter devices (powered by external power banks) positioned at north and east send multiple Wi-Fi beacons to their respective receivers (connected to laptops via USB for data collection) located at south and west. To capture multi-perspective CSI data, all six participants sequentially performed designated activities while standing in the centre of the tripod arrangement for 5 s per sample. The system collected approximately 300-450 packets per sample for approximately 1200 samples per activity, capturing CSI information across the 166 subcarriers employed in the Wi-Fi IEEE 802.11n standard. By leveraging the richness of this dataset, HAR researchers can develop more robust and generalizable CSI-based HAR models. Compared to traditional HAR approaches, these CSI-based models hold the promise of significantly enhanced accuracy and robustness when deployed in real-world scenarios. This stems from their ability to capture the nuanced dynamics of human movement through the analysis of wireless channel characteristic from different spatial variations (utilizing two-diagonal ESP32 transceivers configuration) with higher degree of dimensionality (166 subcarriers).
ABSTRACT
A miniaturized and wideband four-port multiple-input multiple-output (MIMO) antenna pair for Wi-Fi mobile terminals application is proposed. The proposed antenna pair consists of four multi-branch antenna elements arranged orthogonally, with an overall size of 40 × 40 × 3.5 mm3 and each antenna element size of 15.2 × 3.5 mm × 0.8 mm3. The performance of the proposed antenna shows the advantages of a wide frequency band, low mutual coupling, high efficiency, and a compact structure. The wideband characteristics of the antenna elements are achieved through multi-mode resonance. The suppression of coupling is accomplished by strategically positioning the four compact antenna elements to ensure their maximum radiation directions are orthogonal, thus eliminating the need for an additional decoupling structure. In this paper, the proposed antenna is optimized in terms of the parameters then simulated and measured. The simulated results illustrate that an impedance bandwidth of the antenna is about 15% (5.06~5.88 GHz) with S11 < -10 dB, excellent port isolation exceeds 20 dB between all ports, a high radiation efficiency ranges from 51.2% to 89.9%, the maximum gain is 4.5 dBi, and the ECCs are less than 0.04. The measured results show that the -10 dB impedance bandwidth of the antenna is about 13% (5.13~5.80 GHz), the isolation between the antenna elements is better than 21 dB, the radiation efficiency ranges from 51.8% to 92.3%, the maximum gain is 5.3 dBi, and the ECCs are less than 0.05. The proposed four-port MIMO antenna works on the 5G LTE band 46 and Wi-Fi 6E operating bands. As a mobile terminal antenna, the proposed design scheme demonstrates excellent performance and applicability, fulfilling the requirements for 5G mobile terminal applications.
ABSTRACT
Introduction: Electromagnetic radiation (EMR) is widely used nowadays in various fields due to rapid expansion of technology and affects different organs such as endocrine glands. Antioxidants protect the cells and act as a free radical scavenger. Aim of Work: The aim of the study was to clarify the effect of EMR emitted from Wi-Fi router on the thyroid gland of adult male albino rats and the possible protective role of combined Vitamin C and zinc. Materials and Methods: Thirty adult male albino rats were divided into three groups: Group I (control group), Group II (received combined Vitamin C and Zinc in one tablet called IMMUNO-MASH), and Group III (experimental groups). Group III was divided into two subgroups (A and B) according to the duration of exposure: 6 h and 24 h/day. Each of these groups was divided into two equal subgroups. One was exposed only to EMR while the other was exposed to EMR and received combined Vitamin C and zinc. All rats were weighed at the beginning and at the end of the experiment. The thyroid gland was prepared for general histological, anti-calcitonin immunostaining, and ultrastructural study. Furthermore, measurement of total serum T3, T4, and thyroid-stimulating hormone (TSH) hormone levels and quantitative analysis of immunoreactive C-cells were done. Then, statistical analysis was done on the number of immunoreactive C-cells, data of the body weight, and the hormonal levels. Results: A highly significant increase in the body weight in subgroups exposed to EMR for 24 h/day was observed. Furthermore, they showed a highly significant decline in T3 and T4 levels together with a highly significant increase in TSH level. With increasing period of exposure, there was a variable degree of deterioration in the form of congestion and dilatation of blood vessels, cellular infiltration, follicular disintegration, vacuolar degeneration, and desquamated follicular cells in the colloid. The C-cells showed a significant increase in the mean number compared with the control group. Ultrastructural analysis of follicular cells revealed colloid droplets, deteriorations in rough endoplasmic reticulum, degenerating nuclei, and swollen mitochondria according to the dose of exposure. There was apparent improvement with the use of combined Vitamin C and zinc. Conclusion: Wi-Fi radiation has a very serious effect on thyroid gland morphology and activity. Moreover, experimentally induced hypothyroidism by radiation resulted in increased C-cell number. Combined Vitamin C and zinc could have a protective role against this tissue damage.
ABSTRACT
Large-scale multi-building and multi-floor indoor localization has recently been the focus of intense research in indoor localization based on Wi-Fi fingerprinting. Although significant progress has been made in developing indoor localization algorithms, few studies are dedicated to the critical issues of using existing and constructing new Wi-Fi fingerprint databases, especially for large-scale multi-building and multi-floor indoor localization. In this paper, we first identify the challenges in using and constructing Wi-Fi fingerprint databases for large-scale multi-building and multi-floor indoor localization and then provide our recommendations for those challenges based on a case study of the UJIIndoorLoc database, which is the most popular publicly available Wi-Fi fingerprint multi-building and multi-floor database. Through the case study, we investigate its statistical characteristics with a focus on the three aspects of (1) the properties of detected wireless access points, (2) the number, distribution and quality of labels, and (3) the composition of the database records. We then identify potential issues and ways to address them using the UJIIndoorLoc database. Based on the results from the case study, we not only provide valuable insights on the use of existing databases but also give important directions for the design and construction of new databases for large-scale multi-building and multi-floor indoor localization in the future.
ABSTRACT
In the face of rising population, erratic climate, resource depletion, and increased exposure to natural hazards, environmental monitoring is increasingly important. Satellite data form most of our observations of Earth. On-the-ground observations based on in situ sensor systems are crucial for these remote measurements to be dependable. Providing open-source options to rapidly prototype environmental datalogging systems allows quick advancement of research and monitoring programs. This paper introduces Loom, a development environment for low-power Arduino-programmable microcontrollers. Loom accommodates a range of integrated components including sensors, various datalogging formats, internet connectivity (including Wi-Fi and 4G Long Term Evolution (LTE)), radio telemetry, timing mechanisms, debugging information, and power conservation functions. Additionally, Loom includes unique applications for science, technology, engineering, and mathematics (STEM) education. By establishing modular, reconfigurable, and extensible functionality across components, Loom reduces development time for prototyping new systems. Bug fixes and optimizations achieved in one project benefit all projects that use Loom, enhancing efficiency. Although not a one-size-fits-all solution, this approach has empowered a small group of developers to support larger multidisciplinary teams designing diverse environmental sensing applications for water, soil, atmosphere, agriculture, environmental hazards, scientific monitoring, and education. This paper not only outlines the system design but also discusses alternative approaches explored and key decision points in Loom's development.
ABSTRACT
Multi-link operation (MLO) is a new and essential mechanism of IEEE 802.11be Extremely High Throughput (Wi-Fi 7) that can increase throughput and decrease latency in Wireless Local Area Networks (WLANs). The MLO enables a Multi-Link Device (MLD) to perform Simultaneous Transmission and Reception (STR) in different frequency bands. However, not all MLDs can support STR due to cross-link or in-device coexistence interference, while an STR-unable MLD (NSTR-MLD) can transmit multiple frames simultaneously in more than two links. This study focuses on the problems when NSTR-MLDs share a link with Single-Link Devices (SLDs). We propose a Contention-Less Synchronous Transmission (CLST) mechanism to improve fairness between NSTR-MLDs and SLDs while increasing the total network throughput. The proposed mechanism classifies links as MLD Dominant Links (MDLs) and Heterogeneous Coexistence Links (HCLs). In the proposed mechanism, an NSTR-MLD obtains a Synchronous Transmission Token (STT) through a virtual channel contention in the HCL but does not actually transmit a frame in the HCL, which is compensated for by a synchronous transmission triggered in the MDL. Moreover, the CLST mechanism allows additional subsequent transmissions up to the accumulated STT without further contention. Extensive simulation results confirm the outstanding performance of the CLST mechanism in terms of total throughput and fairness compared to existing synchronous transmission mechanisms.
ABSTRACT
Wi-Fi is an integral and invaluable part of our media practices. Wireless networks are blended into our media environment and, in terms of infrastructural importance, have become comparable with electricity or water. This article offers a new transnational perspective on the underexplored history of IEEE 802.11 standards by focusing on the tensions between the United States and Europe in terms of development trajectories of wireless technology. The goal is to analyze the standardization of wireless networking through a transnational lens and to contribute to enhanced understanding of the global proliferation of Wi-Fi technology. Four particular aspects of the transnational development of Wi-Fi technology are discussed: the rivalry between US and European standards, the constitutive choice to focus on data transmission, radio spectrum availability, and the peculiarities of network authentication.
ABSTRACT
With the escalating demand for Radio Frequency Identification (RFID) technology and the Internet of Things (IoT), there is a growing need for sustainable and autonomous power solutions to energize low-powered devices. Consequently, there is a critical imperative to mitigate dependency on batteries during passive operation. This paper proposes the conceptual framework of rectenna architecture-based radio frequency energy harvesters' performance, specifically optimized for low-power device applications. The proposed prototype utilizes the surroundings' Wi-Fi signals within the 2.4 GHz frequency band. The design integrates a seven-stage Cockroft-Walton rectifier featuring a Schottky diode HSMS286C and MA4E2054B1-1146T, a low-pass filter, and a fractal antenna. Preliminary simulations conducted using Advanced Design System (ADS) reveal that a voltage of 3.53 V can be harvested by employing a 1.57 mm thickness Rogers 5880 printed circuit board (PCB) substrate with an MA4E2054B1-1146T rectifier prototype, given a minimum power input of -10 dBm (0.1 mW). Integrating the fabricated rectifier and fractal antenna successfully yields a 1.5 V DC output from Wi-Fi signals, demonstrable by illuminating a red LED. These findings underscore the viability of deploying a fractal antenna-based radio frequency (RF) harvester for empowering small electronic devices.
ABSTRACT
Most facilities are structured in a repetitive manner. In this paper, we propose an algorithm and its partial implementation for a cellular guide in such facilities without GPS use. The complete system is based on iBeacons-like components, which operate on BLE technology, and their integration into a navigation application. We assume that the user's location is determined with sufficient accuracy. Our main goal revolves around leveraging the repetitive structure of the given facility to optimize navigation in terms of storage requirements, energy efficiency in the cellular device, algorithmic complexity, and other aspects. To the best of our knowledge, there is no prior experience in addressing this specific aim. In order to provide high performance in real time, we rely on optimal saving and the use of pre-calculated and stored navigation sub-routes. Our implementation seamlessly integrates iBeacon communications, a pre-defined indoor map, diverse data structures for efficient information storage, and a user interface, all working cohesively under a single supervision. Each module can be considered, developed, and improved independently. The approach is mainly directed to places, such as passenger ships, hotels, colleges, and so on. Because of the fact that there are "replicated" parts on different floors, stored once and used for multiple routes, we reduce the amount of information that must be stored, thus helping to reduce memory usage and as a result, yielding a better running time and energy consumption.
ABSTRACT
Microstrip antennas usually suffer from high losses, gain, and efficiency degradation. It is a challenging task to miniaturize the patch antenna without degrading the performance parameters. To mitigate the above problems, a microstrip patch antenna loaded with stubs and printed on the ground plane loaded with dumbbell meta-atom is presented in this paper. The proposed double dumbbell meta-atom consists of two complementary split ring resonator (CSRR) cells loaded with rectangular rings. This exhibits the Double Negative (DNG) characteristic at 2.45 GHz. The devised meta-atom possesses dimensions of 0.05λ x 0.03λ at lower giga-hertz range. The meta-atom is further analyzed in CST-Microwave Studio and the corresponding S-parameters are extracted in MATLAB using the Nicolson Ross Weir (NRW) method. The electrical model of the meta-atom is also analyzed using Agilent ADS simulator. Further, two models of the proposed antenna with FR-4 and RT/Duroid-5880 are designed and compared. The proposed patch antenna resonates at three different frequency bands i.e. 2.445 GHz with a 3-dB bandwidth of 110 MHz (2.4 GHz-2.51 GHz), at 5.85 GHz with a 3-dB bandwidth of 730 MHz (5.13 GHz-5.86 GHz), and at 8.83 GHz with a 3-dB bandwidth of 1.83 GHz (7.7 GHz-9.53 GHz). This exhibit peak gains of 2.75dBi, 3.53dBc and 4.36dBi with low cross polarization levels at the said frequencies of operation. Further, the antenna possesses circular polarization in the frequency band (5.15 GHz-5.63 GHz). This antenna is used for Wi-Fi, ISM and X-band communications. The designed prototype is fabricated and tested and bears resemblance to the simulated results.
ABSTRACT
This paper introduces an innovative non-contact heart rate monitoring method based on Wi-Fi Channel State Information (CSI). This approach integrates both amplitude and phase information of the CSI signal through rotational projection, aiming to optimize the accuracy of heart rate estimation in home environments. We develop a frequency domain subcarrier selection algorithm based on Heartbeat to subcomponent ratio (HSR) and design a complete set of signal filtering and subcarrier selection processes to further enhance the accuracy of heart rate estimation. Heart rate estimation is conducted by combining the peak frequencies of multiple subcarriers. Extensive experimental validations demonstrate that our method exhibits exceptional performance under various environmental conditions. The experimental results show that our subcarrier selection method for heart rate estimation achieves an average accuracy of 96.8%, with a median error of only 0.8 bpm, representing an approximately 20% performance improvement over existing technologies.
ABSTRACT
Respiratory problems are common amongst older people. The rapid increase in the ageing population has led to a need for developing technologies that can monitor such conditions unobtrusively. This paper presents a novel study that investigates Wi-Fi and ultra-wideband (UWB) antenna sensors to simultaneously monitor two different breathing parameters: respiratory rate, and exhaled breath. Experiments were carried out with two subjects undergoing three breathing cases in breaths per minute (BPM): (1) slow breathing (12 BPM), (2) moderate breathing (20 BPM), and (3) fast breathing (28 BPM). Respiratory rates were captured by Wi-Fi sensors, and the data were processed to extract the respiration rates and compared with a metronome that controlled the subjects' breathing. On the other hand, exhaled breath data were captured by a UWB antenna using a vector network analyser (VNA). Corresponding reflection coefficient data (S11) were obtained from the subjects at the time of exhalation and compared with S11 in free space. The exhaled breath data from the UWB antenna were compared with relative humidity, which was measured with a digital psychrometer during the breathing exercises to determine whether a correlation existed between the exhaled breath's water vapour content and recorded S11 data. Finally, captured respiratory rate and exhaled breath data from the antenna sensors were compared to determine whether a correlation existed between the two parameters. The results showed that the antenna sensors were capable of capturing both parameters simultaneously. However, it was found that the two parameters were uncorrelated and independent of one another.
Subject(s)
Body Fluids , Respiration , Humans , Aged , Exhalation , Respiratory Rate , AgingABSTRACT
In high-density network environments with multiple access points (APs) and stations, individual uplink scheduling by each AP can severely interfere with the uplink transmissions of neighboring APs and their associated stations. In congested areas where concurrent uplink transmissions may lead to significant interference, it would be beneficial to deploy a cooperative scheduler or a central coordinating entity responsible for orchestrating cooperative uplink scheduling by assigning several neighboring APs to support the uplink transmission of a single station within a proximate service area to alleviate the excessive interference. Cooperative uplink scheduling facilitated by cooperative information sharing and management is poised to improve the likelihood of successful uplink transmissions in areas with a high concentration of APs and stations. Nonetheless, it is crucial to account for the queue stability of the stations and the potential delays arising from information exchange and the decision-making process in uplink scheduling to maintain the overall effectiveness of the cooperative approach. In this paper, we propose a Lyapunov drift-plus-penalty framework-based cooperative uplink scheduling method for densely populated Wi-Fi networks. The cooperative scheduler aggregates information, such as signal-to-interference-plus-noise ratio (SINR) and queue status. During the aggregation procedure, propagation delays are also estimated and utilized as a value of expected cooperation delays in scheduling decisions. Upon aggregating the information, the cooperative scheduler calculates the Lyapunov drift-plus-penalty value, incorporating a predefined model parameter to adjust the system accordingly. Among the possible scheduling candidates, the proposed method proceeds to make uplink decisions that aim to reduce the upper bound of the Lyapunov drift-plus-penalty value, thereby improving the network performance and stability without a severe increase in cooperation delay in highly congested areas. Through comprehensive performance evaluations, the proposed method effectively enhances network performance with an appropriate model parameter. The performance improvement is particularly notable in highly congested areas and is achieved without a severe increase in cooperation delays.
ABSTRACT
Positioning in indoor scenarios using signals of opportunity is an effective solution enabling accurate and reliable performance in Global Navigation Satellite System (GNSS)-obscured scenarios. Despite the availability of numerous fingerprinting datasets utilizing various wireless signals, the challenge of device heterogeneity and sample density remains an unanswered issue. To address this gap, this work introduces TUJI1, an anonymized IEEE 802.11 Wireless LAN (Wi-Fi) fingerprinting dataset collected using 5 different commercial devices in a fine-grained grid. The dataset contains the matched fingerprints of Received Signal Strength Indicator (RSSI) measurements with the corresponding coordinates, split into training and testing subsets for effortless and fair reproducibility.
ABSTRACT
RF-based gesture recognition systems outperform computer vision-based systems in terms of user privacy. The integration of Wi-Fi sensing and deep learning has opened new application areas for intelligent multimedia technology. Although promising, existing systems have multiple limitations: (1) they only work well in a fixed domain; (2) when working in a new domain, they require the recollection of a large amount of data. These limitations either lead to a subpar cross-domain performance or require a huge amount of human effort, impeding their widespread adoption in practical scenarios. We propose Wi-AM, a privacy-preserving gesture recognition framework, to address the above limitations. Wi-AM can accurately recognize gestures in a new domain with only one sample. To remove irrelevant disturbances induced by interfering domain factors, we design a multi-domain adversarial scheme to reduce the differences in data distribution between different domains and extract the maximum amount of transferable features related to gestures. Moreover, to quickly adapt to an unseen domain with only a few samples, Wi-AM adopts a meta-learning framework to fine-tune the trained model into a new domain with a one-sample-per-gesture manner while achieving an accurate cross-domain performance. Extensive experiments in a real-world dataset demonstrate that Wi-AM can recognize gestures in an unseen domain with average accuracy of 82.13% and 86.76% for 1 and 3 data samples.
Subject(s)
Gestures , Pattern Recognition, Automated , Humans , Recognition, Psychology , Information Technology , Intelligence , AlgorithmsABSTRACT
The Ministry of Health, Labor and Welfare mandated the creation of the business continuity plan (BCP) for disaster key hospitals on March 31, 2017. Supposing the hospital information system (HIS) failure occurred, the picture archiving and communication system (PACS) also suffers obstacles, we assumed building a new network was necessary for radiological examination images. The purpose of this study was to investigate whether building a new network for radiological examination images is necessary in an emergency. Using wireless fidelity (Wi-Fi), the new network consisting of one image server and two tablet terminals A and B was constructed. The study measured the portable image transfer time for various stages of the network. The results were as follows: Transfer time from the mobile X-ray unit to the image server was 4.12±0.86 s, that from the image server to the tablet device A was 5.14±0.71 s, and that from the image server to the tablet device B was 7.32±1.66 s. Therefore, the new network configuration can provide a reliable means of accessing radiological images during emergency situations when the HIS and PACS may experience obstacles or failures.
Subject(s)
Radiology Information Systems , Disasters , Hospital Information Systems , Disaster Planning/methods , HumansABSTRACT
The recognition of human activity is crucial as the Internet of Things (IoT) progresses toward future smart homes. Wi-Fi-based motion-recognition stands out due to its non-contact nature and widespread applicability. However, the channel state information (CSI) related to human movement in indoor environments changes with the direction of movement, which poses challenges for existing Wi-Fi movement-recognition methods. These challenges include limited directions of movement that can be detected, short detection distances, and inaccurate feature extraction, all of which significantly constrain the wide-scale application of Wi-Fi action-recognition. To address this issue, we propose a direction-independent CSI fusion and sharing model named CSI-F, one which combines Convolutional Neural Networks (CNN) and Gated Recurrent Units (GRU). Specifically, we have introduced a series of signal-processing techniques that utilize antenna diversity to eliminate random phase shifts, thereby removing noise influences unrelated to motion information. Later, by amplifying the Doppler frequency shift effect through cyclic actions and generating a spectrogram, we further enhance the impact of actions on CSI. To demonstrate the effectiveness of this method, we conducted experiments on datasets collected in natural environments. We confirmed that the superposition of periodic actions on CSI can improve the accuracy of the process. CSI-F can achieve higher recognition accuracy compared with other methods and a monitoring coverage of up to 6 m.
Subject(s)
Internet of Things , Movement , Humans , Motion , Doppler Effect , EnvironmentABSTRACT
Digital Twin (DT), a concept of Healthcare (4.0), represents the subject's biological properties and characteristics in a digital model. DT can help in monitoring respiratory failures, enabling timely interventions, personalized treatment plans to improve healthcare, and decision-support for healthcare professionals. Large-scale implementation of DT technology requires extensive patient data for accurate monitoring and decision-making with Machine Learning (ML) and Deep Learning (DL). Initial respiration data was collected unobtrusively with the ESP32 Wi-Fi Channel State Information (CSI) sensor. Due to limited respiration data availability, the paper proposes a novel statistical time series data augmentation method for generating larger synthetic respiration data. To ensure accuracy and validity in the augmentation method, correlation methods (Pearson, Spearman, and Kendall) are implemented to provide a comparative analysis of experimental and synthetic datasets. Data processing methodologies of denoising (smoothing and filtering) and dimensionality reduction with Principal Component Analysis (PCA) are implemented to estimate a patient's Breaths Per Minute (BPM) from raw respiration sensor data and the synthetic version. The methodology provided the BPM estimation accuracy of 92.3% from raw respiration data. It was observed that out of 27 supervised classifications with k-fold cross-validation, the Bagged Tree ensemble algorithm provided the best ML-supervised classification. In the case of binary-class and multi-class, the Bagged Tree ensemble showed accuracies of 89.2% and 83.7% respectively with combined real and synthetic respiration dataset with the larger synthetic dataset. Overall, this provides a blueprint of methodologies for the development of the respiration DT model.
Subject(s)
Respiration , Respiratory Insufficiency , Humans , Time Factors , Machine Learning , AlgorithmsABSTRACT
Unobtrusive sensing (device-free sensing) aims to embed sensing into our daily lives. This is achievable by re-purposing communication technologies already used in our environments. Wireless Fidelity (Wi-Fi) sensing, using Channel State Information (CSI) measurement data, seems to be a perfect fit for this purpose since Wi-Fi networks are already omnipresent. However, a big challenge in this regard is CSI data being sensitive to 'domain factors' such as the position and orientation of a subject performing an activity or gesture. Due to these factors, CSI signal disturbances vary, causing domain shifts. Shifts lead to the lack of inference generalization, i.e., the model does not always perform well on unseen data during testing. We present a domain factor-independent feature-extraction pipeline called 'mini-batch alignment'. Mini-batch alignment steers a feature-extraction model's training process such that it is unable to separate intermediate feature-probability density functions of input data batches seen previously from the current input data batch. By means of this steering technique, we hypothesize that mini-batch alignment (i) absolves the need for providing a domain label, (ii) reduces pipeline re-building and re-training likelihood when encountering latent domain factors, and (iii) absolves the need for extra model storage and training time. We test this hypothesis via a vast number of performance-evaluation experiments. The experiments involve both one- and two-domain-factor leave-out cross-validation, two open-source gesture-recognition datasets called SignFi and Widar3, two pre-processed input types called Doppler Frequency Spectrum (DFS) and Gramian Angular Difference Field (GADF), and several existing domain-shift mitigation techniques. We show that mini-batch alignment performs on a par with other domain-shift mitigation techniques in both position and orientation one-domain leave-out cross-validation using the Widar3 dataset and DFS as input type. When considering a memory-complexity-reduced version of the GADF as input type, mini-batch alignment shows hints of recuperating performance regarding a standard baseline model to the extent that no additional performance due to weight steering is lost in both one-domain-factor leave-out and two-orientation-domain-factor leave-out cross-validation scenarios. However, this is not enough evidence that the mini-batch alignment hypothesis is valid. We identified pitfalls leading up to the hypothesis invalidation: (i) lack of good-quality benchmark datasets, (ii) invalid probability distribution assumptions, and (iii) non-linear distribution scaling issues.
ABSTRACT
Wi-Fi-based human activity recognition has attracted significant attention. Deep learning methods are widely used to achieve feature representation and activity sensing. While more learnable parameters in the neural networks model lead to richer feature extraction, it results in significant resource consumption, rendering the model unsuitable for lightweight Internet of Things (IoT) devices. Furthermore, the sensing performance heavily relies on the quality and quantity of data, which is a time-consuming and labor-intensive task. Therefore, there is a need to explore methods that reduce the dependence on the quality and quantity of the dataset while ensuring recognition performance and decreasing model complexity to adapt to ubiquitous lightweight IoT devices. In this paper, we propose a novel Lightweight-Complex Temporal Convolution Network (L-CTCN) for human activity recognition. Specifically, this approach effectively combines complex convolution with a Temporal Convolution Network (TCN). Complex convolution can extract richer information from limited raw complex data, reducing the reliance on the quality and quantity of training samples. Based on the designed TCN framework with 1D convolution and residual blocks, the proposed model can achieve lightweight human activity recognition. Extensive experiments verify the effectiveness of the proposed method. We can achieve an average recognition accuracy of 96.6% with only 0.17 M parameter size. This method performs well under conditions of low sampling rates and a low number of subcarriers and samples.