Um ambiente para teste e diagnóstico de drones usando cossimulação

Abstract

The unmanned aerial vehicles (UAVs), also know as drones, are very important to execute flights with no necessary pilot in the vehicle, thus it is programmed to run flight missions. However, they require reliability to execute missions, then with diagnostic it is possible to predict vehicle failure during or before the flight. The objective of this work is to present a testing tool, which analyzes and evaluates drones during the flight in indoor environments. For this purpose, the frameworks Ptolemy II was extended for communication with real drones using the High Level Architecture (HLA). The presented testing environment is extendable for other testing routines, and is ready for integration with other simulation and analysis tools. For testing, a total of 40 flights were performed. From that, 20 were used to train a Decision Tree algorithm, and the other 20 to test the algorithm, where some of them had ananomaly added to one ofthe propellers. The accuraterate of fault detection was 70%.