Research Analyzer
← Back ICRA 2026

SignLoc: Robust Localization Using Navigation Signs and Public Maps

Nicky Zimmerman, Joel Loo, Ayush Agrawal, David Hsu

PDF

AI summary

Key figure (auto-extracted from paper)
SignLoc enables robust global robot localization in large indoor-outdoor environments by matching detected navigation signs to public maps, requiring only one to two sign observations.
Global localization navigation signs public maps semantic navigation particle filter indoor-outdoor navigation

Problem

Robots struggle to localize in large, heterogeneous human environments without extensive pre-deployment sensor mapping, despite humans routinely using ubiquitous navigation signs and public maps for wayfinding.

Approach

SignLoc extracts a navigational graph from public floor plans and OpenStreetMap data, then uses a probabilistic observation model within a particle filter to match directional and locational cues from detected signs to the graph for global localization.

Key results

  • Extracts unified navigational graphs from floor plans and OpenStreetMap
  • Matches sign-derived directional and locational cues via a probabilistic observation model
  • Achieves reliable global localization after observing only one to two signs
  • Demonstrates robust performance across diverse large-scale indoor-outdoor environments

Why it matters

Enables robots to navigate complex public spaces immediately upon deployment without costly pre-mapping, bridging the gap between human wayfinding aids and autonomous systems.

Abstract

Navigation signs and maps, such as floor plans and street maps, are widely available and serve as ubiquitous aids for way-finding in human environments. Yet, they are rarely used by robot systems. This paper presents SignLoc, a global localization method that leverages navigation signs to localize the robot on publicly available maps—specifically floor plans and OpenStreetMap (OSM) graphs–without prior sensor-based mapping. SignLoc first extracts a navigation graph from the input map. It then employs a probabilistic observation model to match directional and locational cues from the detected signs to the graph, enabling robust topo-semantic localization within a Monte Carlo framework. We evaluated SignLoc in diverse large- scale environments: part of a university campus, a shopping mall, and a hospital complex. Experimental results show that SignLoc reliably localizes the robot after observing only one to two signs.

Index terms

Localization Semantic Scene Understanding

Related papers