Motivation and Background

Abstract

The successful development of an autonomous aerial vehicle requires solving complex engineering problems. Toward the end of the 1980’s, the shrinking size of computers and the arrival of small commercial global positioning systems (GPS) units and inertial sensor packages made it possible to instrument a small-scale rotorcraft for free and automatic flight. In these early years of aerial robotics a small number of these systems were built, mainly in academic institutions, and a few were able to demonstrate basic flight capabilities such as hover at a position and slow flight following waypoints. In the period following these important breakthroughs, little progress were made in improving the automatic flying capability; the results were far from the theoretical potential of the vehicle. The main reason for this limitation was the absence of an accurate model of the vehicle dynamics that could be used for the analysis and design of the flight control system. These early results were often obtained using classical control architectures with empirically tuned feedback gains.