An Evaluation of Gas Source Localization Algorithms for Mobile Robots

This paper presents a comparative evaluation of different approaches to the problem of gas source localization (GSL) with a mobile robot. Concretely, four state-of-the-art algorithms are implemented and evaluated: Surge-Cast, Spiral, Surge-Spiral and a Probabilistic (Particle Filter-based) method. The experiments have been carried out with the gas dispersion simulator GADEN and the robotic tools offered by ROS (Robotic Operating System) under diverse, realistic environments that feature obstacles and turbulent airflows. Our study reveals, among other results, that Surge-Spiral out-performs Surge-Cast under turbulent wind conditions, and the particle filter approach becomes advantageous only when the assumption of a homogeneous wind holds. We believe that these findings can help the research community to decide on the most appropriate GSL method for a given application. Besides, the code that implements these algorithms is made publicly available.