Long-Term Robot Navigation in Indoor Environments Estimating Patterns in Traversability Changes

TL;DR

This paper presents a method for long-term indoor robot navigation that learns and exploits patterns in traversability changes to improve path planning over time.

Contribution

It introduces an incremental model of traversability patterns combined with probabilistic predictions to enhance navigation decisions in dynamic indoor environments.

Findings

Long-term navigation paths are shorter than greedy shortest paths.

The approach effectively captures and utilizes traversability change patterns.

Experiments demonstrate improved navigation efficiency over time.

Abstract

Nowadays, mobile robots are deployed in many indoor environments, such as offices or hospitals. These environments are subject to changes in the traversability that often happen by following repeating patterns. In this paper, we investigate the problem of navigating in such environments over extended periods of time by capturing these patterns and exploiting this knowledge to make informed decisions. Our approach incrementally estimates a model of the traversability changes from robot's observations and uses a probabilistic graphical model to make predictions at currently unobserved locations. In the belief space defined by the predictions, we plan paths that trade off the risk to encounter obstacles and the information gain of visiting unknown locations. We implemented our approach and tested it in different indoor environments. The experiments suggest that in the long run, our approach leads to navigation along shorter paths compared to following a greedy shortest path policy.