• Design and Implementation of a UAV-based Platform for Air Pollution Monitoring and Source Identification

      Garza Castañón, Luis Eduardo; Yungaicela Naula, Noé Marcelo; Ponce Cuspinera, Luis; Vargas Martínez, Adriana (Instituto Tecnológico y de Estudios Superiores de Monterrey, 2018-05-15)
      This document presents the thesis proposal for obtaining the Master of Science in Intelligent Systems. Technology, industry and government forecasts coincide that the planet will withstand a maximum of 50 years at the rate of current air pollution. Air pollution has reached critical levels causing major impacts on health and economy across the globe. Environmental monitoring and control agencies, as well as industries, require a reliable and cost-effective tool that is easy to deploy where required to assess contamination levels, and on that basis, take the necessary actions. Current measurement methods using pressurized balloons, satellite imagery, or earth stations result in considerable investment, as well as providing low spatial and temporal resolution. There are also systems for measuring air pollution using Unmanned Air Vehicles (UAV), which are financed by large government institutions or international organizations whose budget and resources allow costly implementations. Other related works are limited to the capture of atmospheric data using the UAVs and offline analysis. This work presents the design and implementation of an open-source UAV-based platform for measuring atmospheric pollutants and an algorithm for the localization of the air pollutant sources with the use of a UAV and in-line processing of the pollutants data. The development of the UAV-based platform includes: the UAV mounting and characterization and the control system to guide the navigation of the vehicle, the appropriate sensors selection and integration to the UAV, the data transmission from the sensors onboard the UAV to the ground station, and the implementation of the user interface which is based on a web design. The algorithm for the air pollutant source localization is based on a metaheuristic component, to follow the increasing gradient of the pollutant concentration, and complemented with a probabilistic component to concentrate the searching to the most promising areas in the targeted environment. The results of this work are: Outdoors experiments of the UAV-based platform for the air pollutant monitoring and indoor experiments to validate the algorithm for the source localization. The results show effectiveness and robustness of the UAV-based platform and of the algorithm for the source identification.