An Adaptive Informative Path Planning Algorithm for Real-time Air Quality Monitoring Using UAVs

Environmental monitoring is a heavily data driven task where data sample efficiency is paramount due to the shear volumes of gathered data. In particular, air monitoring strongly depends on sensor location. Since the recent past, Unmanned Aerial Vehicles (UAVs) present themselves as a prospective solution for flexible and better air quality data gathering. In this paper, we present a novel adaptive Informative Path Planning (IPP) approach that enables UAVs navigate through a sample utility map based on adaptive Statistical Gas Distribution Models (GDM) for efficient surveying. The presented adaptive IPP approach maximises the amount of gathered information per mission within the system constraints in known and unknown environments with near optimal performance. The effectiveness of the algorithm is tested through extensive simulation. The results showed high quality sample collection, low computational costs and better model prediction metrics against other surveying strategies. Although framed in an air environmental monitoring context, the developed solution can be used for any generic IPP problem by adapting the sample utility map to the particular application.