Autonomous Visual Detection of Bio-hazards Locations with Aerial Drones

Time period: 2019-01-01 to 2020-12-31

Project leader: Thiemo Voigt

Other personnel: Christian Rohner, Luca Mottola

Funder: Swedish Research Council

Type of award: Research environment

Total fundning: 787 500 SEK

Dengue and Zika are arboviral viruses that affect a significant portion of the world population causing many deaths. While there is no direct correlation between the income level and the possibility of being infected by the dengue virus, the economic impact on the poor is much larger since the funds spent by households below the poverty-line for the treatment of dengue amounts to 93.7% of their monthly per capita income even though free health care is available in Sri Lanka.The main vectors species of the Dengue and Zika viruses are mosquitoes that breed in very slow-flowing or standing water pools. It is important to reduce and control such breeding grounds to contain the spread of the diseases. We propose to use drones, unmanned micro aerial vehicles (UAV), to locate mosquito breeding grounds. UAVs equipped with cameras, accelerometers and other sensors can provide close-range aerial-inspection-capabilities. Using drones is a low-cost and hence sustainable solution. The main scientific challenges are to conceive visual-based, active sensing functionality to allow drones autonomously locate the relevant sites with little to no human intervention. Further, drones must be enabled to maneuver in confined areas to carry out the required tasks in built-up urban environments. Lastly, we must adaptively coordinate the operation of multiple drones to balance the application load among available resources, particularly with respect to the limited battery lifetime.