Multi-Objective Optimization of Unmanned Aerial Vehicle Three-Dimensional Paths for Air Quality Monitoring and Environmental Sensing

The negative impact of air pollution on human health and the environment requires close monitoring of air quality. Unmanned aerial vehicles (UAVs) are gaining popularity in environmental sensing because they provide higher spatial resolution than fixed ground sensors and can be deployed on demand. However, the short flight time of UAVs makes them unsuitable for complete surveys in extensive areas. Instead, UAVs may be more effectively employed to refine and verify coarse estimates provided by other, usually stationary, systems. This work focuses on optimizing UAV flying trajectories to efficiently acquire samples in specific regions of interest, for which prior information about levels of pollutants is available, while maximizing the time spent in each region. Our approach is based on a Multi-Objective Genetic Algorithm that deals with the conflicting requirements emerging from the characteristics of the analyzed problem. Maximization of time spent in regions of interest potentially broadens the scope of the approach to the integration of data from multiple sensors and other non-sampling tasks. The effectiveness of the proposed method over traditional single-objective optimization techniques is shown through numerical simulation with synthetic data and real-world Air Quality Index data taking into account multiple pollutants. Results on two-dimensional and three-dimensional scenarios confirm the suitability of the proposed approach for UAV path planning in air quality monitoring tasks.