A Large Language Model-Guided IMU-PDR Method for Detection and Error Correction
Guowei Liang, Hongyu Wu, Siqi Bai, Hong Tang · 2026
Read A Large Language Model-Guided IMU-PDR Method for Detection and Error Correction by Guowei Liang, Hongyu Wu, Siqi Bai, Hong Tang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13611523.txt.
Description
Guowei Liang and colleagues proposed a new indoor positioning method in 2026 for places where satellite signals fail. Phones and wearables can track walking steps with motion sensors, but errors build up fast as conditions change. The team fed motion data, reshaped into a structured form, to a large language model loaded with expert knowledge about walking. The model handled step counting, step length, heading fixes, and rebuilding the walked path.
In tests, the method counted steps more accurately and held its own on step length. It also curbed the slow drift in walking direction that ruins indoor tracks, and the rebuilt paths stayed more consistent overall. The authors argue that, with the right background knowledge, language models can reason usefully over raw motion signals. They present the work as a fresh route to smarter indoor navigation without extra building equipment.
Catalog record
| CREATOR | Guowei Liang, Hongyu Wu, Siqi Bai, Hong Tang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 80 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |