- 著者
-
松浦 晃子
鈴木 真二
- 出版者
- 一般社団法人日本機械学会
- 雑誌
- 交通・物流部門大会講演論文集
- 巻号頁・発行日
- vol.2006, no.15, pp.245-248, 2006-12-13
This paper presents an on-line navigation and adaptive flight control for waypoint tracking and course tracking of an Unmanned Aerial Vehicle (UAV) using neural network (NN). The UAV learns to fly above the designated points and course. This system consists of a navigation guidance part for making rolling angle command from the relative position of the aircraft and the target or the course, and the Feedback Error Learning (FEL) part for rolling angle control. For waypoint tracking, the rotation rate of the line of sight is needed, and for course tracking, the relative position of the aircraft is needed. FEL is a learning scheme of NN proposed by Kawato et al. The NN controller is included in a conventional feedback (CFB) controller. By minimizing CFB signals through a learning process in FEL, the NN obtains inverse feedforward dynamics of a nonlinear plant, in this case the aileron-rolling angle system. A numerical simulation using this method is performed to show the effectiveness of the method. It is considered that this system is appropriate for small UAVs whose computer resources and measurement hardware are restricted.