{"paper":{"title":"Magnetic Indoor Localization through CNN Regression and Rotation Invariance","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Rotation-invariant 2D magnetic features let a CNN regress accurate indoor positions without device orientation alignment.","cross_cats":["cs.LG"],"primary_cat":"cs.RO","authors_text":"Bernd Sch\\\"aufele, Helge Ros\\'e, Ilja Radusch, Konstantin Klipp, Tom Koubek","submitted_at":"2026-04-24T12:06:08Z","abstract_excerpt":"Indoor positioning is an essential technology for a wide range of applications in GNSS-denied environments, including indoor navigation and IoT systems. Combining convolutional neural networks (CNNs) and magnetic field-based features offers a low-cost, infrastructure-free solution for precise positioning. While magnetic fingerprints are a promising approach for indoor positioning, models trained on raw 3D magnetometer data are highly sensitive to device orientation. We address this by using two rotation invariant features derived from the 3D magnetic field: the norm (Mn) and the projection ont"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"2D (Mn, Mg) inputs maintain rotation invariant accuracy and surpass the 3D inputs once rotation exceeds building-specific thresholds for three reference buildings: 0° for Loomis (large), 5° for Talbot (medium), and 6° for CSL (small).","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the chosen rotation-invariant features preserve enough position-specific information from the magnetic field and that the MagPie dataset trajectories represent realistic usage conditions without significant temporal or device variation.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Rotation-invariant magnetic features (norm and gravity projection) with a dilated CNN enable accurate, orientation-robust indoor position regression on the MagPie dataset.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Rotation-invariant 2D magnetic features let a CNN regress accurate indoor positions without device orientation alignment.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"c591e01aba016a2aa17695e7e5a7af9254c47827058fc8c5d242dc802804937e"},"source":{"id":"2604.22896","kind":"arxiv","version":2},"verdict":{"id":"43ae2a84-3346-401d-b7d7-a45fd8cd0088","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-08T11:28:19.153103Z","strongest_claim":"2D (Mn, Mg) inputs maintain rotation invariant accuracy and surpass the 3D inputs once rotation exceeds building-specific thresholds for three reference buildings: 0° for Loomis (large), 5° for Talbot (medium), and 6° for CSL (small).","one_line_summary":"Rotation-invariant magnetic features (norm and gravity projection) with a dilated CNN enable accurate, orientation-robust indoor position regression on the MagPie dataset.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the chosen rotation-invariant features preserve enough position-specific information from the magnetic field and that the MagPie dataset trajectories represent realistic usage conditions without significant temporal or device variation.","pith_extraction_headline":"Rotation-invariant 2D magnetic features let a CNN regress accurate indoor positions without device orientation alignment."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.22896/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-21T10:39:48.230826Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T23:56:36.537194Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"72662bc802de3e52e827992b88f1ff95534016709489344ffe223eb26bb08052"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}