Mobile Robot Odor Plume Tracking using Three Dimensional Information

—This paper studies the utilization of data gathered from multiple gas sensors placed vertically on a wheeled mobile robot to perform plume tracking strategies. The main research questions to be answered are if is this advantageous and if so, trying to quantify that advantage in terms of plume tracking over traditional two dimensional plume acquisition. This was accomplished by testing a few well known plume tracking algorithms (Surge-Caste, Casting and Gradient following) against a modified version of each one of the algorithms, designed to take into account odor data from the gas sensor vertical array. Extensive real world experiments were conducted in a large indoor environment. It was found that plume tracking with three dimensional acquisition outperformed the two dimensional variants in the tested environment in about 25%.