Modelling Data Aided Sensing With UAVs for Efficient Data Collection

With the increasing adoption of fifth-generation (5G) networks and wide bandwidth availability, unmanned aerial vehicles (UAVs) should now be practicable to stream and operate at different altitudes and upload incredibly large quantities of data from sensors. We develop a novel intelligent sensing framework (Choi, 2020) so that sensing and communication occur on demand, according to a given query. Our query-based Data Aided Sensing (DAS) approach differs from existing methods as we start with selecting relevant nodes based on the query and then allocating UAVs for collecting measurements. We develop an economical and efficient approach to collect measurements from massive passive sensors, which adaptively minimize the entropy gap while scheduling UAVs to sensors. At the heart of our solution is a mathematical approach for estimating entropy, cost and energy consumption, thereby deploying dynamically optimal UAV-sensor association (e.g., optimizing entropy and employing bipartite graph). Our system has been evaluated with extensive simulations, providing insightful observations and findings.