Aircraft parameter estimation — A tool for development of aerodynamic databases

With the evolution of high performance modern aircraft and spiraling developmental and experimental costs, the importance of flight validated databases for flight control design applications and for flight simulators has increased significantly in the recent past. Ground-based and in-flight simulators are increasingly used not only for pilot training but also for other applications such as flight planning, envelope expansion, design and analysis of control laws, and handling qualities investigations. Most of these demand a high-fidelity aerodynamic database representing the flight vehicle. System identification methodology, evolved over the past three decades, provides a powerful and sophisticated tool to identify from flight data aerodynamic characteristics valid over the entire operational flight envelope. This paper briefly presents aircraft parameter estimation methods for both stable and unstable aircraft, highlighting the developmental work at the DLR Institute of Flight Mechanics. Various aspects of database identification and its validation are presented. Practical aspects like the proper choice of integration and optimization methods as well as limitations of gradient approximation through finite-differences are brought out. Though the paper focuses on application of system identification methods to flight vehicles, its use in other applications, like the modelling of inelastic deformations of metallic materials, is also presented. It is shown that there are many similar problems and several challenges requiring additional concepts and algorithms.