NeurIT: Pushing the Limit of Neural Inertial Tracking for Indoor Robotic IoT

Inertial tracking is vital for robotic IoT and has gained popularity thanks to the ubiquity of low-cost Inertial Measurement Units (IMUs) and deep learning-powered tracking algorithms. Existing works, however, have not fully utilized IMU measurements, particularly magnetometers, nor maximized the po...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zheng, Xinzhe, Ji, Sijie, Pan, Yipeng, Zhang, Kaiwen, Wu, Chenshu
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Zheng, Xinzhe
Ji, Sijie
Pan, Yipeng
Zhang, Kaiwen
Wu, Chenshu
description Inertial tracking is vital for robotic IoT and has gained popularity thanks to the ubiquity of low-cost Inertial Measurement Units (IMUs) and deep learning-powered tracking algorithms. Existing works, however, have not fully utilized IMU measurements, particularly magnetometers, nor maximized the potential of deep learning to achieve the desired accuracy. To enhance the tracking accuracy for indoor robotic applications, we introduce NeurIT, a sequence-to-sequence framework that elevates tracking accuracy to a new level. NeurIT employs a Time-Frequency Block-recurrent Transformer (TF-BRT) at its core, combining the power of recurrent neural network (RNN) and Transformer to learn representative features in both time and frequency domains. To fully utilize IMU information, we strategically employ body-frame differentiation of the magnetometer, which considerably reduces the tracking error. NeurIT is implemented on a customized robotic platform and evaluated in various indoor environments. Experimental results demonstrate that NeurIT achieves a mere 1-meter tracking error over a 300-meter distance. Notably, it significantly outperforms state-of-the-art baselines by 48.21% on unseen data. NeurIT also performs comparably to the visual-inertial approach (Tango Phone) in vision-favored conditions and surpasses it in plain environments. We believe NeurIT takes an important step forward toward practical neural inertial tracking for ubiquitous and scalable tracking of robotic things. NeurIT, including the source code and the dataset, is open-sourced here: https://github.com/NeurIT-Project/NeurIT.
doi_str_mv 10.48550/arxiv.2404.08939
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2404_08939</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2404_08939</sourcerecordid><originalsourceid>FETCH-LOGICAL-a679-50e294f5765ab06f83bb4a53c600127cae09f0f97d80c34f4e746e031d8b793</originalsourceid><addsrcrecordid>eNotj8tOwzAURL1hgQofwAr_QMJN_O4OVTwiRQVV2UfXjk2tPozcFMHfkxRWM5oZjXQIuaug5FoIeMD8Hb_KmgMvQRtmrkm79ufcdEv6fj5t4_GDjltP23iII02BziXuaXP0eYyT6TK63bwKKU_pkCbZJJvG6GiTuhtyFXB_8rf_uiCb56du9Vq0by_N6rEtUCpTCPC14UEoKdCCDJpZy1EwJwGqWjn0YAIEowYNjvHAveLSA6sGbZVhC3L_d3qB6T9zPGD-6Weo_gLFfgGEUUYX</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>NeurIT: Pushing the Limit of Neural Inertial Tracking for Indoor Robotic IoT</title><source>arXiv.org</source><creator>Zheng, Xinzhe ; Ji, Sijie ; Pan, Yipeng ; Zhang, Kaiwen ; Wu, Chenshu</creator><creatorcontrib>Zheng, Xinzhe ; Ji, Sijie ; Pan, Yipeng ; Zhang, Kaiwen ; Wu, Chenshu</creatorcontrib><description>Inertial tracking is vital for robotic IoT and has gained popularity thanks to the ubiquity of low-cost Inertial Measurement Units (IMUs) and deep learning-powered tracking algorithms. Existing works, however, have not fully utilized IMU measurements, particularly magnetometers, nor maximized the potential of deep learning to achieve the desired accuracy. To enhance the tracking accuracy for indoor robotic applications, we introduce NeurIT, a sequence-to-sequence framework that elevates tracking accuracy to a new level. NeurIT employs a Time-Frequency Block-recurrent Transformer (TF-BRT) at its core, combining the power of recurrent neural network (RNN) and Transformer to learn representative features in both time and frequency domains. To fully utilize IMU information, we strategically employ body-frame differentiation of the magnetometer, which considerably reduces the tracking error. NeurIT is implemented on a customized robotic platform and evaluated in various indoor environments. Experimental results demonstrate that NeurIT achieves a mere 1-meter tracking error over a 300-meter distance. Notably, it significantly outperforms state-of-the-art baselines by 48.21% on unseen data. NeurIT also performs comparably to the visual-inertial approach (Tango Phone) in vision-favored conditions and surpasses it in plain environments. We believe NeurIT takes an important step forward toward practical neural inertial tracking for ubiquitous and scalable tracking of robotic things. NeurIT, including the source code and the dataset, is open-sourced here: https://github.com/NeurIT-Project/NeurIT.</description><identifier>DOI: 10.48550/arxiv.2404.08939</identifier><language>eng</language><subject>Computer Science - Artificial Intelligence ; Computer Science - Human-Computer Interaction ; Computer Science - Robotics</subject><creationdate>2024-04</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2404.08939$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2404.08939$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Zheng, Xinzhe</creatorcontrib><creatorcontrib>Ji, Sijie</creatorcontrib><creatorcontrib>Pan, Yipeng</creatorcontrib><creatorcontrib>Zhang, Kaiwen</creatorcontrib><creatorcontrib>Wu, Chenshu</creatorcontrib><title>NeurIT: Pushing the Limit of Neural Inertial Tracking for Indoor Robotic IoT</title><description>Inertial tracking is vital for robotic IoT and has gained popularity thanks to the ubiquity of low-cost Inertial Measurement Units (IMUs) and deep learning-powered tracking algorithms. Existing works, however, have not fully utilized IMU measurements, particularly magnetometers, nor maximized the potential of deep learning to achieve the desired accuracy. To enhance the tracking accuracy for indoor robotic applications, we introduce NeurIT, a sequence-to-sequence framework that elevates tracking accuracy to a new level. NeurIT employs a Time-Frequency Block-recurrent Transformer (TF-BRT) at its core, combining the power of recurrent neural network (RNN) and Transformer to learn representative features in both time and frequency domains. To fully utilize IMU information, we strategically employ body-frame differentiation of the magnetometer, which considerably reduces the tracking error. NeurIT is implemented on a customized robotic platform and evaluated in various indoor environments. Experimental results demonstrate that NeurIT achieves a mere 1-meter tracking error over a 300-meter distance. Notably, it significantly outperforms state-of-the-art baselines by 48.21% on unseen data. NeurIT also performs comparably to the visual-inertial approach (Tango Phone) in vision-favored conditions and surpasses it in plain environments. We believe NeurIT takes an important step forward toward practical neural inertial tracking for ubiquitous and scalable tracking of robotic things. NeurIT, including the source code and the dataset, is open-sourced here: https://github.com/NeurIT-Project/NeurIT.</description><subject>Computer Science - Artificial Intelligence</subject><subject>Computer Science - Human-Computer Interaction</subject><subject>Computer Science - Robotics</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj8tOwzAURL1hgQofwAr_QMJN_O4OVTwiRQVV2UfXjk2tPozcFMHfkxRWM5oZjXQIuaug5FoIeMD8Hb_KmgMvQRtmrkm79ufcdEv6fj5t4_GDjltP23iII02BziXuaXP0eYyT6TK63bwKKU_pkCbZJJvG6GiTuhtyFXB_8rf_uiCb56du9Vq0by_N6rEtUCpTCPC14UEoKdCCDJpZy1EwJwGqWjn0YAIEowYNjvHAveLSA6sGbZVhC3L_d3qB6T9zPGD-6Weo_gLFfgGEUUYX</recordid><startdate>20240413</startdate><enddate>20240413</enddate><creator>Zheng, Xinzhe</creator><creator>Ji, Sijie</creator><creator>Pan, Yipeng</creator><creator>Zhang, Kaiwen</creator><creator>Wu, Chenshu</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20240413</creationdate><title>NeurIT: Pushing the Limit of Neural Inertial Tracking for Indoor Robotic IoT</title><author>Zheng, Xinzhe ; Ji, Sijie ; Pan, Yipeng ; Zhang, Kaiwen ; Wu, Chenshu</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a679-50e294f5765ab06f83bb4a53c600127cae09f0f97d80c34f4e746e031d8b793</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Computer Science - Artificial Intelligence</topic><topic>Computer Science - Human-Computer Interaction</topic><topic>Computer Science - Robotics</topic><toplevel>online_resources</toplevel><creatorcontrib>Zheng, Xinzhe</creatorcontrib><creatorcontrib>Ji, Sijie</creatorcontrib><creatorcontrib>Pan, Yipeng</creatorcontrib><creatorcontrib>Zhang, Kaiwen</creatorcontrib><creatorcontrib>Wu, Chenshu</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Zheng, Xinzhe</au><au>Ji, Sijie</au><au>Pan, Yipeng</au><au>Zhang, Kaiwen</au><au>Wu, Chenshu</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>NeurIT: Pushing the Limit of Neural Inertial Tracking for Indoor Robotic IoT</atitle><date>2024-04-13</date><risdate>2024</risdate><abstract>Inertial tracking is vital for robotic IoT and has gained popularity thanks to the ubiquity of low-cost Inertial Measurement Units (IMUs) and deep learning-powered tracking algorithms. Existing works, however, have not fully utilized IMU measurements, particularly magnetometers, nor maximized the potential of deep learning to achieve the desired accuracy. To enhance the tracking accuracy for indoor robotic applications, we introduce NeurIT, a sequence-to-sequence framework that elevates tracking accuracy to a new level. NeurIT employs a Time-Frequency Block-recurrent Transformer (TF-BRT) at its core, combining the power of recurrent neural network (RNN) and Transformer to learn representative features in both time and frequency domains. To fully utilize IMU information, we strategically employ body-frame differentiation of the magnetometer, which considerably reduces the tracking error. NeurIT is implemented on a customized robotic platform and evaluated in various indoor environments. Experimental results demonstrate that NeurIT achieves a mere 1-meter tracking error over a 300-meter distance. Notably, it significantly outperforms state-of-the-art baselines by 48.21% on unseen data. NeurIT also performs comparably to the visual-inertial approach (Tango Phone) in vision-favored conditions and surpasses it in plain environments. We believe NeurIT takes an important step forward toward practical neural inertial tracking for ubiquitous and scalable tracking of robotic things. NeurIT, including the source code and the dataset, is open-sourced here: https://github.com/NeurIT-Project/NeurIT.</abstract><doi>10.48550/arxiv.2404.08939</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2404.08939
ispartof
issn
language eng
recordid cdi_arxiv_primary_2404_08939
source arXiv.org
subjects Computer Science - Artificial Intelligence
Computer Science - Human-Computer Interaction
Computer Science - Robotics
title NeurIT: Pushing the Limit of Neural Inertial Tracking for Indoor Robotic IoT
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-21T11%3A27%3A59IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=NeurIT:%20Pushing%20the%20Limit%20of%20Neural%20Inertial%20Tracking%20for%20Indoor%20Robotic%20IoT&rft.au=Zheng,%20Xinzhe&rft.date=2024-04-13&rft_id=info:doi/10.48550/arxiv.2404.08939&rft_dat=%3Carxiv_GOX%3E2404_08939%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true