Odor-source searching using a mobile robot in time-variant airflow environments with obstacles

Abstract

A new path planning method for a mobile robot to search for an odor source in time-variant airflow environments with obstacles has been developed and implemented. The proposed method subsequently employs a three-stage process in order to plan a desired path for the robot to find the odor source and avoid collisions with obstacles in search area. In the first stage, when an odor clue is found by the mobile robot, an odor-patch path, indicating where the detected odor patch came from, is estimated based on the airflow velocities recorded by the robot. And then, a search route is planned for the robot to find the odor source based on the estimated odor-patch path and the recent airflow directions. In the third stage, when obstacles are detected by the robot, the planned search route will be modified with range data sampled by on-board range sensors. The proposed method has a low computational cost, thus it can be used in online applications. The simulations in time-variant airflow environments show that the robot has a purposeful odor-source search behavior in clustered obstacles.