무인 비행선 적응 자동비행제어 설계 연구
Adaptive autopilot design for unmanned airships
무인비행선 자동비행제어설계 항공공학;
- 원문 URL
In this paper, the design of a neural network based adaptive autopilot for unmanned airships is presented. The modeling of the airship is implemented using YEZ-2A airship aerodynamic data with the consideration of the difference between an airship and a conventional aircraft. First, using the classical PID control algorithm, the autopilot is designed to meet the performance requirements for the altitude hold and line tracking flight. The autopilot of the classical PID control law has some limitations such as its performance becomes degraded due to parameter variations in the dynamics and high nonlinearity in the aerodynamic characteristics. In this paper, a control scheme to augment an existing classical controller with adaptive neural networks is introduced to overcome the limitation and is applied to the design of airship autopilot. The derivation of the stability of the whole closed-loop system including neural networks is summarized. The benefit of this scheme is that the neural network output is simply added to the nominal control signal, thereby preserving the existing control architecture. By comparing with the simulation results of the classical controller, the performance improvement of the neural network based controller is showed when the model error is introduced. It is concluded that the neural networks augmented to an existing classical controller not only expand the flight region satisfying the performance requirements but also can improve the robustness of the original existing controller.
- 이 논문과 함께 이용한 콘텐츠