Applications of Fuzzy Neural Networks and Genetic Algorithms to Aircraft Landing Control
碩士 === 國立海洋大學 === 導航與通訊系碩士班 === 91 === Abstract The atmospheric disturbances affect not only flying qualities of an airplane but also flight safety. According to a survey of the National Transportation Safety Board, 22.6 percent of aircraft accidents in the years of 1989 to 1999 were weat...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2003
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/07008883091336146700 |
| Summary: | 碩士 === 國立海洋大學 === 導航與通訊系碩士班 === 91 === Abstract
The atmospheric disturbances affect not only flying qualities of an airplane but also flight safety. According to a survey of the National Transportation Safety Board, 22.6 percent of aircraft accidents in the years of 1989 to 1999 were weather related. When aircraft approaches landing phase the altitude is low and the speed is slow. If the aircraft encountered wind shear or turbulence while landing it could cause altitude loss, heading variation and even crash. Take off and landing are the most difficult operations of a flight. Most aircrafts have installed the Automatic Landing System (ALS) which helps aircraft landing stably and reduces pilot’s work loading greatly. Controller of the conventional ALS usually uses gain-scheduling techniques. If the landing environment is beyond predefined conditions, the ALS must disable and the pilot has to operate the aircraft manually. Most pilots have no experience on wind shear or turbulence environment. Usually it ends up with airplane crash. Neural network and fuzzy logic system are complementary techniques to each other. The combination of these two techniques is called Fuzzy Neural Network which has both policy decision of fuzzy system and adaptive learning of neural network. This paper presents an aircraft automatic landing control scheme that uses a fuzzy neural network controller. The controller is constructed by linguistic fuzzy rule or functional fuzzy rule. A real number type genetic algorithm is used to adjust control gains of the pitch autopilot. A neural network device emulates wind disturbances, which provides adaptive control to the system. A backpropagation through time algorithm with a linearized inverse aircraft model refines the connection weights of the fuzzy neural network controller. The proposed controller can overcome violent variation of environment and enhance capability of control for wind disturbances during landing.
|
|---|