Non-linear Dynamic Inversion versus Gain Scheduling Techniques for Re-entry Flight Control System

Paper number

IAC-05-C1.3.09

Author

Mrs. Sinar Juliana, Technical University of Delft (TUDelft), The Netherlands

Coauthor

Dr. Q.P. Chu, Technical University of Delft (TUDelft), The Netherlands

Coauthor

Prof. J.A. Mulder, Technical University of Delft (TUDelft), The Netherlands

Coauthor

Dr. T.J. van Baten, Technical University of Delft (TUDelft), The Netherlands

Year

2005

Abstract

The DART (Delft Aerospace Re-entry Test Demonstrator) project conducted in Delft University of Technology entails the development, building and flights of a small, reusable, low cost re-entry test bed, for the purpose of fundamental studies concerning re-entry. The on-ground research facilities such as wind tunnels are used to its maximum capacity to cover the re-entry flight domain. Nevertheless, there still exist some gaps of information, which lead to an increased margin in re-entry spacecraft design. Moreover, flight data are needed to verify the on-ground test, i.e. to have confident in its validity. Therefore, repeated re-entry flight tests have to be conducted to fill the gaps and to verify the ground tests.

On its re-entry flight mission, the DART vehicle has to follow a predefined path. It will re-enter the atmosphere with a hypersonic speed and an altitude more than 100 km. The vehicle will lose its altitude and velocity rapidly, and has to land in subsonic speed. To fulfill these objectives, the vehicle has to be equipped with flight control to guide it along the re-entry path and to ensure that it has the correct attitude and can be recovered safely on the predetermined landing site.

In this paper, a study is conducted to compare two techniques in design of attitude control system for DART re-entry vehicle. The first technique is Non-linear Dynamic Inversion (NDI) plus Proportional Integral Derivative (PID), whereas the second one is a linear state-feedback control with gain scheduling. The vehicle is an axis-symmetrical spinning body equipped with four pairs of Reaction Control System (RCS) thrusters. The objectives of the control system are to maintain the spinning at constant rate and to track the reference attitude angles.

In NDI plus PID technique, the control is designed in two steps. In the first step, non-linear model of the system is simplified by using feedback linearization. The inversion of the system is used as a feedback to the original model. As a result, the model becomes linear time invariant (LTI). The second step is to apply the linear PID technique for this LTI system to obtain the desired spinning rate and attitude angle. In this technique, no gain scheduling is necessary, because the system is already linear because of NDI application.
In the gain scheduling technique, the non-linear system is linearized around its stationary points, and linear state-feedback control is built on these specific points. The trajectory has to be discretized into several operating points for which the linear control system is designed. 

Comparison between the two designs is made in terms of stability and performance of the resulting control system. The complete results are presented in the full paper.

Abstract document

IAC-05-C1.3.09.pdf

Manuscript document

IAC-05-C1.3.09.pdf (🔒 authorized access only).

To get the manuscript, please contact IAF Secretariat.