Pub Date : 2023-11-22DOI: 10.1109/JISPIN.2023.3335882
Tomas Bravenec;Joaquín Torres-Sospedra;Michael Gould;Tomas Fryza
This article centers on the deeper presentation of a new and publicly accessible dataset comprising Wi-Fi probe requests. Probe requests fall within the category of management frames utilized by the 802.11 (Wi-Fi) protocol. Given the ever-evolving technological landscape and the imperative need for up-to-date data, research on probe requests remains essential. In this context, we present a comprehensive dataset encompassing a one-month probe request capture conducted in a university office environment. This dataset accounts for a diverse range of scenarios, including workdays, weekends, and holidays, accumulating over 1 400 000 probe requests. Our contribution encompasses a detailed exposition of the dataset, delving into its critical facets. In addition to the raw packet capture, we furnish a detailed floor plan of the office environment, commonly referred to as a radio map, to equip dataset users with comprehensive environmental information. To safeguard user privacy, all individual user information within the dataset has been anonymized. This anonymization process rigorously balances the preservation of users' privacy with the dataset's analytical utility, rendering it nearly as informative as raw data for research purposes. Furthermore, we demonstrate a range of potential applications for this dataset, including but not limited to presence detection, expanded assessment of temporal received signal strength indicator stability, and evaluation of privacy protection measures. Apart from these, we also include temporal analysis of probe request transmission frequency and period between Wi-Fi scans as well as a peak into possibilities with pattern analysis.
{"title":"UJI Probes Revisited: Deeper Dive Into the Dataset of Wi-Fi Probe Requests","authors":"Tomas Bravenec;Joaquín Torres-Sospedra;Michael Gould;Tomas Fryza","doi":"10.1109/JISPIN.2023.3335882","DOIUrl":"https://doi.org/10.1109/JISPIN.2023.3335882","url":null,"abstract":"This article centers on the deeper presentation of a new and publicly accessible dataset comprising Wi-Fi probe requests. Probe requests fall within the category of management frames utilized by the 802.11 (Wi-Fi) protocol. Given the ever-evolving technological landscape and the imperative need for up-to-date data, research on probe requests remains essential. In this context, we present a comprehensive dataset encompassing a one-month probe request capture conducted in a university office environment. This dataset accounts for a diverse range of scenarios, including workdays, weekends, and holidays, accumulating over 1 400 000 probe requests. Our contribution encompasses a detailed exposition of the dataset, delving into its critical facets. In addition to the raw packet capture, we furnish a detailed floor plan of the office environment, commonly referred to as a radio map, to equip dataset users with comprehensive environmental information. To safeguard user privacy, all individual user information within the dataset has been anonymized. This anonymization process rigorously balances the preservation of users' privacy with the dataset's analytical utility, rendering it nearly as informative as raw data for research purposes. Furthermore, we demonstrate a range of potential applications for this dataset, including but not limited to presence detection, expanded assessment of temporal received signal strength indicator stability, and evaluation of privacy protection measures. Apart from these, we also include temporal analysis of probe request transmission frequency and period between Wi-Fi scans as well as a peak into possibilities with pattern analysis.","PeriodicalId":100621,"journal":{"name":"IEEE Journal of Indoor and Seamless Positioning and Navigation","volume":"1 ","pages":"221-230"},"PeriodicalIF":0.0,"publicationDate":"2023-11-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10325607","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138822153","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-11-20DOI: 10.1109/JISPIN.2023.3334690
Ali A. Darwish;Arie Nakhmani
This article proposes an architecture for drone navigation and target interception, utilizing a self-supervised, model-free deep reinforcement learning approach. Unlike the traditional methods relying on complex controllers, our approach uses deep reinforcement learning with cascade rewards, enabling a single drone to navigate obstacles and intercept targets using only a forward-facing depth–RGB camera. This research has significant implications for robotics, as it demonstrates how complex tasks can be tackled using deep reinforcement learning. Our work encompasses three key contributions. First, we tackle the challenge of partial observability when employing nonlinear function approximators for learning stochastic policies. Second, we optimize the task of maximizing the overall expected reward. Finally, we develop a software library for training drones to track and intercept targets. Through our experiments, we demonstrated that our approach, incorporating cascade reward, outperforms state-of-the-art deep Q -network algorithms in terms of learning policies. By leveraging our methodology, drones can successfully navigate complex indoor and outdoor environments and effectively intercept targets based on visual cues.
{"title":"Drone Navigation and Target Interception Using Deep Reinforcement Learning: A Cascade Reward Approach","authors":"Ali A. Darwish;Arie Nakhmani","doi":"10.1109/JISPIN.2023.3334690","DOIUrl":"https://doi.org/10.1109/JISPIN.2023.3334690","url":null,"abstract":"This article proposes an architecture for drone navigation and target interception, utilizing a self-supervised, model-free deep reinforcement learning approach. Unlike the traditional methods relying on complex controllers, our approach uses deep reinforcement learning with cascade rewards, enabling a single drone to navigate obstacles and intercept targets using only a forward-facing depth–RGB camera. This research has significant implications for robotics, as it demonstrates how complex tasks can be tackled using deep reinforcement learning. Our work encompasses three key contributions. First, we tackle the challenge of partial observability when employing nonlinear function approximators for learning stochastic policies. Second, we optimize the task of maximizing the overall expected reward. Finally, we develop a software library for training drones to track and intercept targets. Through our experiments, we demonstrated that our approach, incorporating cascade reward, outperforms state-of-the-art deep \u0000<italic>Q</i>\u0000-network algorithms in terms of learning policies. By leveraging our methodology, drones can successfully navigate complex indoor and outdoor environments and effectively intercept targets based on visual cues.","PeriodicalId":100621,"journal":{"name":"IEEE Journal of Indoor and Seamless Positioning and Navigation","volume":"1 ","pages":"130-140"},"PeriodicalIF":0.0,"publicationDate":"2023-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10323488","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138633905","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-11-20DOI: 10.1109/JISPIN.2023.3334697
Aviad Etzion;Itzik Klein
Mobile robots are used in a variety of applications indoors and outdoors. In real-world scenarios, frequently, the navigation solution relies only on the inertial sensors. Consequently, the navigation solution drifts in time. In this article, we propose the mobile robot pure inertial framework (MoRPI). Instead of travelling in a straight line trajectory, the robot moves in a periodic motion trajectory to enable peak-to-peak estimation. Two types of MoRPI approaches are suggested, one is based on both accelerometer and gyroscope readings while the other requires only the gyroscopes. Closed form analytical solutions are derived to show that MoRPI produces lower position error compared to the classical pure inertial solution. In addition, field experiments were made with a mobile robot equipped with two different types of inertial sensors. The results show the benefits of using our approach.
{"title":"MoRPI: Mobile Robot Pure Inertial Navigation","authors":"Aviad Etzion;Itzik Klein","doi":"10.1109/JISPIN.2023.3334697","DOIUrl":"https://doi.org/10.1109/JISPIN.2023.3334697","url":null,"abstract":"Mobile robots are used in a variety of applications indoors and outdoors. In real-world scenarios, frequently, the navigation solution relies only on the inertial sensors. Consequently, the navigation solution drifts in time. In this article, we propose the mobile robot pure inertial framework (MoRPI). Instead of travelling in a straight line trajectory, the robot moves in a periodic motion trajectory to enable peak-to-peak estimation. Two types of MoRPI approaches are suggested, one is based on both accelerometer and gyroscope readings while the other requires only the gyroscopes. Closed form analytical solutions are derived to show that MoRPI produces lower position error compared to the classical pure inertial solution. In addition, field experiments were made with a mobile robot equipped with two different types of inertial sensors. The results show the benefits of using our approach.","PeriodicalId":100621,"journal":{"name":"IEEE Journal of Indoor and Seamless Positioning and Navigation","volume":"1 ","pages":"141-150"},"PeriodicalIF":0.0,"publicationDate":"2023-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10323471","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138633867","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-11-20DOI: 10.1109/JISPIN.2023.3334693
Danish Gufran;Saideep Tiku;Sudeep Pasricha
Smartphone-based indoor localization has emerged as a cost-effective and accurate solution to localize mobile and IoT devices indoors. However, the challenges of device heterogeneity and temporal variations have hindered its widespread adoption and accuracy. Toward jointly addressing these challenges comprehensively, we propose STELLAR, a novel framework implementing a contrastive learning approach that leverages a Siamese multiheaded attention neural network. STELLAR is the first solution that simultaneously tackles device heterogeneity and temporal variations in indoor localization, without the need for retraining the model (recalibration-free). Our evaluations across diverse indoor environments show 8%–75% improvements in accuracy compared to state-of-the-art techniques, to effectively address the device heterogeneity challenge. Moreover, STELLAR outperforms existing methods by 18%–165% over two years of temporal variations, showcasing its robustness and adaptability.
{"title":"STELLAR: Siamese Multiheaded Attention Neural Networks for Overcoming Temporal Variations and Device Heterogeneity With Indoor Localization","authors":"Danish Gufran;Saideep Tiku;Sudeep Pasricha","doi":"10.1109/JISPIN.2023.3334693","DOIUrl":"https://doi.org/10.1109/JISPIN.2023.3334693","url":null,"abstract":"Smartphone-based indoor localization has emerged as a cost-effective and accurate solution to localize mobile and IoT devices indoors. However, the challenges of device heterogeneity and temporal variations have hindered its widespread adoption and accuracy. Toward jointly addressing these challenges comprehensively, we propose STELLAR, a novel framework implementing a contrastive learning approach that leverages a Siamese multiheaded attention neural network. STELLAR is the first solution that simultaneously tackles device heterogeneity and temporal variations in indoor localization, without the need for retraining the model (recalibration-free). Our evaluations across diverse indoor environments show 8%–75% improvements in accuracy compared to state-of-the-art techniques, to effectively address the device heterogeneity challenge. Moreover, STELLAR outperforms existing methods by 18%–165% over two years of temporal variations, showcasing its robustness and adaptability.","PeriodicalId":100621,"journal":{"name":"IEEE Journal of Indoor and Seamless Positioning and Navigation","volume":"1 ","pages":"115-129"},"PeriodicalIF":0.0,"publicationDate":"2023-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10323477","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138491035","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-11-17DOI: 10.1109/JISPIN.2023.3333852
Min Wang;Aurélie Dommes;Valérie Renaudin;Ni Zhu
Navigating in urban environment is a major challenge for visually impaired people. Spatial landmarks are crucial for them to orient and navigate in their environment. In this paper, the spatial landmarks most important and commonly used by visually impaired people are identified through interviews, and geometric constraints of these landmarks are constructed to facilitate the development of map-matching algorithms. Interviews were conducted with 12 visually impaired people who had a range of visual impairments and used various mobility aids. Data were analyzed by sensory modality, occurrence of use, and number of users. 14 main landmarks for urban navigation were selected and categorized into two groups: Waypoints and Reassurance Points, depending on whether they are directly detected by touch. Geometric constraints were developed for each landmark to prepare their integration into map-matching or path-planning algorithms. The result is a comprehensive dictionary of landmarks and their geometric constraints is created, specifically tailored to help visually impaired people navigate urban environments. Our user-centric approach successfully translates the subjective navigation experiences of visually impaired people into an objective, universally accessible format. This bridges the gap between personal experiences and practical applications and paves the way for more inclusive navigation solutions for visually impaired people in urban environments.
{"title":"Analysis of Spatial Landmarks for Seamless Urban Navigation of Visually Impaired People","authors":"Min Wang;Aurélie Dommes;Valérie Renaudin;Ni Zhu","doi":"10.1109/JISPIN.2023.3333852","DOIUrl":"https://doi.org/10.1109/JISPIN.2023.3333852","url":null,"abstract":"Navigating in urban environment is a major challenge for visually impaired people. Spatial landmarks are crucial for them to orient and navigate in their environment. In this paper, the spatial landmarks most important and commonly used by visually impaired people are identified through interviews, and geometric constraints of these landmarks are constructed to facilitate the development of map-matching algorithms. Interviews were conducted with 12 visually impaired people who had a range of visual impairments and used various mobility aids. Data were analyzed by sensory modality, occurrence of use, and number of users. 14 main landmarks for urban navigation were selected and categorized into two groups: Waypoints and Reassurance Points, depending on whether they are directly detected by touch. Geometric constraints were developed for each landmark to prepare their integration into map-matching or path-planning algorithms. The result is a comprehensive dictionary of landmarks and their geometric constraints is created, specifically tailored to help visually impaired people navigate urban environments. Our user-centric approach successfully translates the subjective navigation experiences of visually impaired people into an objective, universally accessible format. This bridges the gap between personal experiences and practical applications and paves the way for more inclusive navigation solutions for visually impaired people in urban environments.","PeriodicalId":100621,"journal":{"name":"IEEE Journal of Indoor and Seamless Positioning and Navigation","volume":"1 ","pages":"93-103"},"PeriodicalIF":0.0,"publicationDate":"2023-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10320446","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138485032","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We consider the robust localization, via Gaussian process regression (GPR), of multiple transmitters/targets based on received signal strength indicator (RSSI) data collected by fixed sensors distributed in the environment. For such a scenario and approach, we contribute both with a novel noise robust procedure to train the parameters of the GPR model, which is achieved via a mini-batch stochastic gradient descent (SGD) scheme with gradients given in closed form, and with a pair of corresponding robust marginalization procedures for the estimation of target locations. Simulation results validate the contributions by showing that the proposed methods significantly outperform the best related state-of-the-art (SotA) alternative and approach the performance of a genie-aided (GA) scheme.
{"title":"Robust Received Signal Strength Indicator (RSSI)-Based Multitarget Localization via Gaussian Process Regression","authors":"Niclas Führling;Hyeon Seok Rou;Giuseppe Thadeu Freitas de Abreu;David González G.;Osvaldo Gonsa","doi":"10.1109/JISPIN.2023.3332033","DOIUrl":"10.1109/JISPIN.2023.3332033","url":null,"abstract":"We consider the robust localization, via Gaussian process regression (GPR), of multiple transmitters/targets based on received signal strength indicator (RSSI) data collected by fixed sensors distributed in the environment. For such a scenario and approach, we contribute both with a novel noise robust procedure to train the parameters of the GPR model, which is achieved via a mini-batch stochastic gradient descent (SGD) scheme with gradients given in closed form, and with a pair of corresponding robust marginalization procedures for the estimation of target locations. Simulation results validate the contributions by showing that the proposed methods significantly outperform the best related state-of-the-art (SotA) alternative and approach the performance of a genie-aided (GA) scheme.","PeriodicalId":100621,"journal":{"name":"IEEE Journal of Indoor and Seamless Positioning and Navigation","volume":"1 ","pages":"104-114"},"PeriodicalIF":0.0,"publicationDate":"2023-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10314734","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135610964","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Traditional pedestrian dead reckoning (PDR) systems have been designed for scenarios where users walk straight ahead. However, user behavior observation at the museum revealed that users often stop or walk sideways to look at the exhibits. If the user's smartphone is moving when the user is stopped, false step detection may occur. In addition, the correct step or change of direction may not be detected in sideways walking. To solve these problems, we propose a novel PDR system. First, we classify the user's walking style to address the problems of false step detection and undetected changes of direction. Next, we use a classifier to detect when the foot touches the ground from smartphone sensor data and perform step detection. Compared with the existing SmartPDR, our proposed method improved positioning accuracy by 20% in straight walking and 70% in sideways walking.
{"title":"Pedestrian Dead Reckoning for Multiple Walking Styles Using Classifier-Based Step Detection","authors":"Ibuki Yoshida;Takumi Suzaki;Hiroaki Murakami;Hiroki Watanabe;Mananari Nakamura;Hiromichi Hashizume;Masanori Sugimoto","doi":"10.1109/JISPIN.2023.3323937","DOIUrl":"10.1109/JISPIN.2023.3323937","url":null,"abstract":"Traditional pedestrian dead reckoning (PDR) systems have been designed for scenarios where users walk straight ahead. However, user behavior observation at the museum revealed that users often stop or walk sideways to look at the exhibits. If the user's smartphone is moving when the user is stopped, false step detection may occur. In addition, the correct step or change of direction may not be detected in sideways walking. To solve these problems, we propose a novel PDR system. First, we classify the user's walking style to address the problems of false step detection and undetected changes of direction. Next, we use a classifier to detect when the foot touches the ground from smartphone sensor data and perform step detection. Compared with the existing SmartPDR, our proposed method improved positioning accuracy by 20% in straight walking and 70% in sideways walking.","PeriodicalId":100621,"journal":{"name":"IEEE Journal of Indoor and Seamless Positioning and Navigation","volume":"1 ","pages":"69-79"},"PeriodicalIF":0.0,"publicationDate":"2023-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10285345","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136305815","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-10-12DOI: 10.1109/JISPIN.2023.3324279
Chi-Shih Jao;Danmeng Wang;Changwei Chen;Eudald Sangenis;Joe Grasso;Solmaz S. Kia;Andrei M. Shkel
This article proposes a generalized UltraWideBand (UWB)-Zero-velocity-UPdaTe (ZUPT)-simultaneous localization and mapping (SLAM) algorithm, a SLAM approach, utilizing a combination of foot-mounted localization systems integrating inertial measurement units (IMUs), UWB modules, barometers, and dynamically-deployed beacons incorporating UWB, IMUs, and reference barometers. The proposed approach leverages a ZUPT-aided Inertial Navigation System augmented with self-contained sensor fusion techniques to map unknown UWB beacons dynamically deployed in an environment during navigation and then utilizes the localized beacons to bound position error propagation. An experimental testbed was developed, and we conducted two series of experiments to validate the performance of the proposed approach. The first experiment involved high-accuracy motion capture cameras in generating ground truth, and the results showed that the proposed approach estimated positions of UWB beacons with a maximum localization error of 0.36 m, when deployed during the first 15 and 20 s of the navigation. In the second experiment, a pedestrian traveled for around 3.5 km in 1 h in a large multifloor indoor environment and deployed seven beacons, during the first 63, 151, 290, 399, 517, 585, and 786 s of the experiment. The proposed generalized UWB-ZUPT-SLAM had a 3-D mean absolute error of 0.48 m in this experiment, equivalent to 0.013% traveling distance.
{"title":"Augmented UWB-ZUPT-SLAM Utilizing Multisensor Fusion","authors":"Chi-Shih Jao;Danmeng Wang;Changwei Chen;Eudald Sangenis;Joe Grasso;Solmaz S. Kia;Andrei M. Shkel","doi":"10.1109/JISPIN.2023.3324279","DOIUrl":"10.1109/JISPIN.2023.3324279","url":null,"abstract":"This article proposes a generalized UltraWideBand (UWB)-Zero-velocity-UPdaTe (ZUPT)-simultaneous localization and mapping (SLAM) algorithm, a SLAM approach, utilizing a combination of foot-mounted localization systems integrating inertial measurement units (IMUs), UWB modules, barometers, and dynamically-deployed beacons incorporating UWB, IMUs, and reference barometers. The proposed approach leverages a ZUPT-aided Inertial Navigation System augmented with self-contained sensor fusion techniques to map unknown UWB beacons dynamically deployed in an environment during navigation and then utilizes the localized beacons to bound position error propagation. An experimental testbed was developed, and we conducted two series of experiments to validate the performance of the proposed approach. The first experiment involved high-accuracy motion capture cameras in generating ground truth, and the results showed that the proposed approach estimated positions of UWB beacons with a maximum localization error of 0.36 m, when deployed during the first 15 and 20 s of the navigation. In the second experiment, a pedestrian traveled for around 3.5 km in 1 h in a large multifloor indoor environment and deployed seven beacons, during the first 63, 151, 290, 399, 517, 585, and 786 s of the experiment. The proposed generalized UWB-ZUPT-SLAM had a 3-D mean absolute error of 0.48 m in this experiment, equivalent to 0.013% traveling distance.","PeriodicalId":100621,"journal":{"name":"IEEE Journal of Indoor and Seamless Positioning and Navigation","volume":"1 ","pages":"80-92"},"PeriodicalIF":0.0,"publicationDate":"2023-10-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10283865","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136302860","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-27DOI: 10.1109/JISPIN.2023.3299433
Darwin P. Quezada Gaibor;Lucie Klus;Roman Klus;Elena Simona Lohan;Jari Nurmi;Mikko Valkama;Joaquín Huerta;Joaquín Torres-Sospedra
Indoor positioning based on machine-learning (ML) models has attracted widespread interest in the last few years, given its high performance and usability. Supervised, semisupervised, and unsupervised models have thus been widely used in this field, not only to estimate the user position, but also to compress, clean, and denoise fingerprinting datasets. Some scholars have focused on developing, improving, and optimizing ML models to provide accurate solutions to the end user. This article introduces a novel method to initialize the input weights in autoencoder extreme learning machine (AE-ELM), namely factorized input data (FID), which is based on the normalized form of the orthogonal component of the input data. AE-ELM with FID weight initialization is used to efficiently reduce the radio map. Once the dimensionality of the dataset is reduced, we use $k$ -nearest neighbors to perform the position estimation. This research work includes a comparative analysis with several traditional ways to initialize the input weights in AE-ELM, showing that FID provide a significantly better reconstruction error. Finally, we perform an assessment with 13 indoor positioning datasets collected from different buildings and in different countries. We show that the dimensionality of the datasets can be reduced more than 11 times on average, while the positioning error suffers only a small increment of 15% (on average) in comparison to the baseline.
{"title":"Autoencoder Extreme Learning Machine for Fingerprint-Based Positioning: A Good Weight Initialization is Decisive","authors":"Darwin P. Quezada Gaibor;Lucie Klus;Roman Klus;Elena Simona Lohan;Jari Nurmi;Mikko Valkama;Joaquín Huerta;Joaquín Torres-Sospedra","doi":"10.1109/JISPIN.2023.3299433","DOIUrl":"https://doi.org/10.1109/JISPIN.2023.3299433","url":null,"abstract":"Indoor positioning based on machine-learning (ML) models has attracted widespread interest in the last few years, given its high performance and usability. Supervised, semisupervised, and unsupervised models have thus been widely used in this field, not only to estimate the user position, but also to compress, clean, and denoise fingerprinting datasets. Some scholars have focused on developing, improving, and optimizing ML models to provide accurate solutions to the end user. This article introduces a novel method to initialize the input weights in autoencoder extreme learning machine (AE-ELM), namely factorized input data (FID), which is based on the normalized form of the orthogonal component of the input data. AE-ELM with FID weight initialization is used to efficiently reduce the radio map. Once the dimensionality of the dataset is reduced, we use \u0000<inline-formula><tex-math>$k$</tex-math></inline-formula>\u0000-nearest neighbors to perform the position estimation. This research work includes a comparative analysis with several traditional ways to initialize the input weights in AE-ELM, showing that FID provide a significantly better reconstruction error. Finally, we perform an assessment with 13 indoor positioning datasets collected from different buildings and in different countries. We show that the dimensionality of the datasets can be reduced more than 11 times on average, while the positioning error suffers only a small increment of 15% (on average) in comparison to the baseline.","PeriodicalId":100621,"journal":{"name":"IEEE Journal of Indoor and Seamless Positioning and Navigation","volume":"1 ","pages":"53-68"},"PeriodicalIF":0.0,"publicationDate":"2023-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/iel7/9955032/9962767/10195972.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"50323047","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-20DOI: 10.1109/JISPIN.2023.3287508
Wei Guo;Lei Jing
With the popularity of Wi-Fi devices and the development of the Internet of Things (IoT), Wi-Fi-based passive motion tracking has attracted significant attention. Most existing works utilize the Angle of Arrival (AoA), Time of Flight (ToF), and Doppler Frequency Shift (DFS) of the Channel State Information (CSI) to track human motions. However, they usually require multiple pairs of Wi-Fi devices and extensive data training to achieve accurate results, which is unrealistic in practical applications. In this article, we propose Wi -Fi M otion T racking ( WiMT ), a low-cost passive motion tracking system based on a single pair of commodity Wi-Fi devices. WiMT calculates the Doppler velocity and phase difference using the CSI obtained from the transmitter with one antenna and the receiver with three antennas. The Z ero V elocity I dentification and C alibration ( ZVIC ) algorithm is proposed to remove the random noise of Doppler velocity when the target is stationary. We take the Doppler velocity as the measurement and employ a particle filter to estimate the motion trajectory. A particle weight update method based on phase difference information is developed to eliminate particles with low confidence. Experimental results in real indoor environment show that WiMT achieves great performance with a motion tracking median error of 7.28 cm and a nonmoving recognition accuracy of 92.6%.
{"title":"Toward Low-Cost Passive Motion Tracking With One Pair of Commodity Wi-Fi Devices","authors":"Wei Guo;Lei Jing","doi":"10.1109/JISPIN.2023.3287508","DOIUrl":"https://doi.org/10.1109/JISPIN.2023.3287508","url":null,"abstract":"With the popularity of Wi-Fi devices and the development of the Internet of Things (IoT), Wi-Fi-based passive motion tracking has attracted significant attention. Most existing works utilize the Angle of Arrival (AoA), Time of Flight (ToF), and Doppler Frequency Shift (DFS) of the Channel State Information (CSI) to track human motions. However, they usually require multiple pairs of Wi-Fi devices and extensive data training to achieve accurate results, which is unrealistic in practical applications. In this article, we propose \u0000<bold>Wi</b>\u0000-Fi \u0000<bold>M</b>\u0000otion \u0000<bold>T</b>\u0000racking (\u0000<bold>WiMT</b>\u0000), a low-cost passive motion tracking system based on a single pair of commodity Wi-Fi devices. WiMT calculates the Doppler velocity and phase difference using the CSI obtained from the transmitter with one antenna and the receiver with three antennas. The \u0000<bold>Z</b>\u0000ero \u0000<bold>V</b>\u0000elocity \u0000<bold>I</b>\u0000dentification and \u0000<bold>C</b>\u0000alibration (\u0000<bold>ZVIC</b>\u0000) algorithm is proposed to remove the random noise of Doppler velocity when the target is stationary. We take the Doppler velocity as the measurement and employ a particle filter to estimate the motion trajectory. A particle weight update method based on phase difference information is developed to eliminate particles with low confidence. Experimental results in real indoor environment show that WiMT achieves great performance with a motion tracking median error of 7.28 cm and a nonmoving recognition accuracy of 92.6%.","PeriodicalId":100621,"journal":{"name":"IEEE Journal of Indoor and Seamless Positioning and Navigation","volume":"1 ","pages":"39-52"},"PeriodicalIF":0.0,"publicationDate":"2023-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/iel7/9955032/9962767/10158358.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"50323046","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：