Extendable Matrix Camera Using Aerial Networks

Abstract

Recent advances in Unmanned Aerial Vehicles (UAVs) have enabled a myriad of new applications many of which provide aerial vision-based sensing. In intrusion detection or target tracking applications, it is important to reach a given area of interest in the shortest time, and create an online data streaming connection to a monitoring station for immediate delivery of content to the operator. However, if the area of interest (AOI) is not contained in the field of view of a single UAV, it is necessary to move the sensor-UAV to sweep the region in order to provide the most fresh information as possible. In order to improve the collection time of the AOI, we can cooperatively use multiple UAVs creating an array of moving cameras, that always remain connected without breaks in communication. In this work, we validate our approach with a simulation that captures physical models and the application layer of each UAV, as well as the wireless network. We propose an optimal solution to sweep the AOI as well as a decentralised formation control algorithm that maintains UAVs equally separated along the optimal path that covers the whole AOI, even with external disturbances such as wind gusts. We show a seven-fold increase in the refresh rate of the AOI coverage when comparing to a solution without the optimal sweeping and decentralised formation control.