Physics-Informed Transformer-LSTM Ensemble for Event-Based Acoustic Indoor Localization with Tokenized TDoA and Relative-Speed Cues

Physics-Informed Transformer-LSTM Ensemble for Event-Based Acoustic Indoor Localization with Tokenized TDoA and Relative-Speed Cues | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Physics-Informed Transformer-LSTM Ensemble for Event-Based Acoustic Indoor Localization with Tokenized TDoA and Relative-Speed Cues Deniz Berke Özsoy, Atakan Özcan, Mohammed Al-Hubaishi This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8992737/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Acoustic indoor localization can be deployed with minimal infrastructure, but indoor reverberation and intermittent non-line-of-sight propagation often bias delay estimates and destabilize purely geometry-based solvers. We study the two-base-station setting with band-separated ultrasonic chirps and propose an event-based learning pipeline that (i) pairs chirp detections across the two recording streams and (ii) converts them into fixed-length windows of discrete tokens summarizing residual (offset-corrected) TDoA, a Doppler-inspired relative-speed difference, confidence, and inter-event timing. A lightweight scenario- and field-aware Transformer maps each 6-event window (38 tokens including boundary tokens) to a 2D position; during training, we additionally examine differentiable regularizers that encourage TDoA-consistent geometry and locally smooth trajectories.On two indoor scenarios and a leakage-safe held-out split, the tuned Transformer outperforms classical regressors, achieving Mean/Median/P90 errors of 0.529/0.517/0.965\,m in Scenario~A and 0.961/0.776/1.639\,m in Scenario~B. On the combined held-out set ($n{=}24$ windows), a late-fusion Transformer+LSTM ensemble further improves reliability, reaching 0.427/0.325/0.903\,m. Acoustic indoor localization indoor positioning time-difference-of-arrival Doppler/relative-speed features physics-informed learning Transformer ensemble learning Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 26 Mar, 2026 Reviews received at journal 25 Mar, 2026 Reviewers agreed at journal 13 Mar, 2026 Reviewers invited by journal 08 Mar, 2026 Editor assigned by journal 02 Mar, 2026 Submission checks completed at journal 02 Mar, 2026 First submitted to journal 28 Feb, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8992737","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":603466091,"identity":"965fc699-428c-4b89-83c1-215fd56c8b2d","order_by":0,"name":"Deniz Berke Özsoy","email":"data:image/png;base64,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","orcid":"","institution":"","correspondingAuthor":true,"prefix":"","firstName":"Deniz","middleName":"Berke","lastName":"Özsoy","suffix":""},{"id":603466092,"identity":"df51ad10-fde5-421b-8976-1a08e012d3be","order_by":1,"name":"Atakan Özcan","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Atakan","middleName":"","lastName":"Özcan","suffix":""},{"id":603466093,"identity":"51724180-45e7-458c-bfc2-34a8266e2ecc","order_by":2,"name":"Mohammed Al-Hubaishi","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Mohammed","middleName":"","lastName":"Al-Hubaishi","suffix":""}],"badges":[],"createdAt":"2026-02-28 07:40:13","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8992737/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8992737/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104780744,"identity":"8f6feba7-4d3c-48fa-86a6-742646d71cce","added_by":"auto","created_at":"2026-03-17 07:53:45","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":11535714,"visible":true,"origin":"","legend":"","description":"","filename":"ManuscriptFiles.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8992737/v1_covered_4ac61560-8ea5-4852-909d-556c4023cee0.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Physics-Informed Transformer-LSTM Ensemble for Event-Based Acoustic Indoor Localization with Tokenized TDoA and Relative-Speed Cues","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"signal-image-and-video-processing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"sivp","sideBox":"Learn more about [Signal, Image and Video Processing](http://link.springer.com/journal/11760)","snPcode":"11760","submissionUrl":"https://submission.nature.com/new-submission/11760/3","title":"Signal, Image and Video Processing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Acoustic indoor localization, indoor positioning, time-difference-of-arrival, Doppler/relative-speed features, physics-informed learning, Transformer, ensemble learning","lastPublishedDoi":"10.21203/rs.3.rs-8992737/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8992737/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAcoustic indoor localization can be deployed with minimal infrastructure, but indoor reverberation and intermittent non-line-of-sight propagation often bias delay estimates and destabilize purely geometry-based solvers. We study the two-base-station setting with band-separated ultrasonic chirps and propose an event-based learning pipeline that (i) pairs chirp detections across the two recording streams and (ii) converts them into fixed-length windows of discrete tokens summarizing residual (offset-corrected) TDoA, a Doppler-inspired relative-speed difference, confidence, and inter-event timing. A lightweight scenario- and field-aware Transformer maps each 6-event window (38 tokens including boundary tokens) to a 2D position; during training, we additionally examine differentiable regularizers that encourage TDoA-consistent geometry and locally smooth trajectories.On two indoor scenarios and a leakage-safe held-out split, the tuned Transformer outperforms classical regressors, achieving Mean/Median/P90 errors of 0.529/0.517/0.965\\,m in Scenario~A and 0.961/0.776/1.639\\,m in Scenario~B. On the combined held-out set ($n{=}24$ windows), a late-fusion Transformer+LSTM ensemble further improves reliability, reaching 0.427/0.325/0.903\\,m.\u003c/p\u003e","manuscriptTitle":"Physics-Informed Transformer-LSTM Ensemble for Event-Based Acoustic Indoor Localization with Tokenized TDoA and Relative-Speed Cues","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-12 06:21:27","doi":"10.21203/rs.3.rs-8992737/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-26T23:44:33+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-25T23:28:45+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"133115944900782492407461123115601762563","date":"2026-03-14T03:25:15+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-08T21:20:05+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-02T17:04:17+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-02T16:58:26+00:00","index":"","fulltext":""},{"type":"submitted","content":"Signal, Image and Video Processing","date":"2026-02-28T07:31:56+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"signal-image-and-video-processing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"sivp","sideBox":"Learn more about [Signal, Image and Video Processing](http://link.springer.com/journal/11760)","snPcode":"11760","submissionUrl":"https://submission.nature.com/new-submission/11760/3","title":"Signal, Image and Video Processing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"84c2d855-18c0-41fe-91fc-6a0abad1db4c","owner":[],"postedDate":"March 12th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-16T17:08:12+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-12 06:21:27","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8992737","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8992737","identity":"rs-8992737","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}