Grid2Guide: A* Enabled Small Language Model for Indoor Navigation

First Author, Md. Wasiul Haque, University of Alabama Sagar Dasgupta, University of Alabama Mizanur Rahman, University of Alabama · 2025

Read Grid2Guide: A* Enabled Small Language Model for Indoor Navigation by First Author, Md. Wasiul Haque, University of Alabama Sagar Dasgupta, University of Alabama Mizanur Rahman, University of Alabama (2025) 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/arxiv-2508-08100.txt.

Description

Wasiul Haque, Sagar Dasgupta, and Mizanur Rahman pair a classic pathfinding algorithm with a small language model for indoor directions, in a 2025 paper. Indoor spaces defeat GPS, and many buildings lack dedicated navigation infrastructure. Their system, Grid2Guide, first converts a floor map into a grid of free and blocked cells. The A-star algorithm then computes the best route between start and finish as a list of terse steps.

A compact language model rewrites those steps into natural, human-readable instructions. Tests across indoor scenarios produced accurate, timely guidance. The pairing is instructive: let a proven algorithm handle geometry, and let the language model handle only the talking. The result is a lightweight system needing no beacons or signals.

Catalog record

CREATORFirst Author, Md. Wasiul Haque, University of Alabama Sagar Dasgupta, University of Alabama Mizanur Rahman, University of Alabama
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE46 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-08100

Browse the full library · About the Tower of Babel