Odor plume source localization with a Pioneer 3 Mobile Robot in an indoor airflow environment

Abstract

Olfactory-based mobile robots use odors as a guide to navigate and track in the unknown environments. The key issue of localizing the odor plume source is how to trace odor plume effectively. This paper presents an effective olfactory-based planning and search algorithms for using on mobile robots. The algorithms are based on Bayesian inference theory and artificial potential field methods. The Bayesian inference theory is used to construct the probability map of plume source based on the flow records and plume detection information collected from sensors. Then the Artificial Potential Field (APF) method is used for planning a plume tracking path. The robot follows the path to trace odor plume until the source is detected. The algorithms were implemented on a Pioneer 3 Mobile Robot in an indoor airflow environment. Experiment results show that the search algorithms are effective and feasible to odor plume source localization problem.