UAV-based air pollutant source localization using gradient and probabilistic methods

Abstract

This work proposes an algorithm for air pollutant source localization using an Unmanned Aerial Vehicle (UAV). The algorithm combines a gradient-based search with a probabilistic method to localize the pollutant source. The design of the gradient-based search component is based on the simulated annealing metaheuristic and allows to trace the plume of pollutant. The probabilistic component contributes to generate a heuristic position of the source location, which is used by the gradient-based metaheuristic to navigate towards the source position, reducing the searching region at each sampling time. The proposed algorithm was tested in a simulated polluted environment. The results showed high effectiveness and robustness of the proposed strategy.