A Biologically-Inspired Algorithm for Gas/Odor Source Localization in an Indoor Environment with no Strong Airflow: First Experimental Results

.Abstract – This paper describes the design of a biologicallyinspired SPIRAL (Searching Pollutant Iterative Rounding ALgorithm) algorithm, for the localization of a gas source in an indoor environment with no strong airflow. Such environment shows a few aspects that make the issue of finding an odor source much harder than in the presence of a strong wind. In fact, in a windless room, the gas diffusion mechanism is strongly perturbed by convection flows and turbulence, which make gas distribution very complex. Moreover, the algorithms used for gas source localization with no predominant airflow cannot exploit the upwind surge movement, usually exploited in traditional algorithms. We discuss the results of SPIRAL implemented on a robotic platform, called MOMO (Multirobots for Odor MOnitoring). Tests show that SPIRAL algorithm localizes a gas source with good results, by using only one gas sensor, and without either referring to anemometers or to any information about wind distribution.