Flight Control Law Design for Launch Vehicles Using Disturbance Observers
- Paper number
IAC-05-C1.8.03
- Author
Dr. Shinji Ishimoto, Japan Aerospace Exploration Agency (ISTA/JAXA), Japan
- Year
2005
- Abstract
The method of dynamic inversion has been widely applied to the design of flight control laws for various aircraft in recent years. In a typical dynamic inversion control law, a classical linear control law generates desired body angular accelerations from guidance or pilot commands, and then a dynamic inversion algorithm converts the desired accelerations into actuator commands. If dynamic inversion is perfect, a combination of the dynamic inversion algorithm and aircraft dynamics becomes equivalent to a bank of decoupled integrators, because the dynamic inversion algorithm cancels unwanted coupling terms or nonlinear terms using an onboard model. Since dynamic inversion controllers inherently depend on on-board models, the designer should take care of robustness against model uncertainties or external disturbances. A common way for good robustness is to include integrators into the linear control law computing the desired accelerations. Although this controller structure has been widely employed, the control law design cannot be done separately for performance and robustness as in the traditional framework of feedback control. In this paper, we propose a new dynamic inversion controller structure using disturbance observers. The disturbance observers estimate disturbances from the desired angular accelerations and measured angular rates based on a nominal plant model as integrators. The disturbance observers will work only when there exist modeling errors or external disturbances. The estimated disturbances are subtracted from the desired angular accelerations to reject external disturbances or to compensate for model uncertainties. Because the disturbance observers are devoted to robustness, the linear control law generating the desired accelerations can be designed only for performance. This point is one of the advantages of the new controller structure. The proposed controller structure is applied to a small experimental reusable vehicle powered by four small rocket engines. This vehicle vertically takes off and vertically lands with the attitude controlled by engine gimbal actuators. The objective of this compact project is to contribute to both expendable and reusable launch vehicles by gaining and demonstrating key technologies that raise system reliability and efficiency in ground operation through repeating a lot of flight tests. At present, we are in the conceptual design phase of the vehicle system. For comparison, two dynamic inversion controllers based on the conventional and new structures are designed in the paper. These controllers are compared and discussed using the results of numerical simulations.
- Abstract document