Flight Test Oriented Autopilot Design for Improved Aerodynamic Parameter Identification

Abstract

In order to reduce development costs and time, model-based design is widely introduced in the industry leading to a strong need for verified high-fidelity simulation models. An inevitable, but challenging process step to obtain such simulation models for GNC-applications is the aerodynamic parameter identification on the basis of real flight test data. The identification process requires distinct excitation maneuvers in order to constrain the design space to a subset of model parameters reducing the complexity of the identification problem and the correlation within the overall parameter set. Typically, manually flown excitation maneuvers are not exact and fully reproducible concerning the requirements and therefore the amount of rejected data points is significant. In case of remotely piloted aircraft systems, the decoupling of the aircraft and the ground pilot in charge leads to an even less sensitive maneuver control, a further reduced disturbance suppression and even greater difficulties in meeting the initialization requirements. This scenario calls for an automation of aerodynamic parameter identification related flight tests. A practical approach to a flight test oriented autopilot for improved aerodynamic parameter identification is proposed within this paper. The requirements for identification excitation maneuvers and the corresponding design of the autopilot are emphasized and flight test results are presented.