Multi-Modal Robotic Platform Development for Odor Source Localization

Abstract

This paper discusses customizing a popular robot development platform ”Turtlebot3” for Odor Source Localization (OSL) task. OSL technology allows a robot to locate and navigate to odor sources in an unknown environment. Turtlebot3 is an agile robot platform that includes Raspberry Pi for on-device computation and an Open-source Control module for ROS (OpenCR) board for additional sensor connection. It runs on Ubuntu and “Robot Operating System (ROS)” that allows it to send sensor data to and receive heading commands from a remote computer that can run complex algorithms. In combination, this robotics platform can be customized to perform a wide variety of robot tasks. This paper focuses on the additional olfactory sensor installation for the OSL experiment. It also discusses an olfactorybased navigation algorithm named moth-inspired algorithm for OSL task. The algorithm was applied to real-world experiments with varying conditions. The experiments show that the mothinspired algorithm successfully navigates to the odor source in laminar airflow environments. The paper also discusses the future scope of adding vision sensors and machine learning methods.