A wavelet neural network based on genetic algorithm and it's apply to gain scheduling flight control

Xun Sun, Weiguo Zhang, Wei Yin, Aijun Li

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

As enlarging of the flight envelop, the aerodynamic derivative of the airplane varies enormous. The gain scheduling method is usually used to deal with it. But the workload is enormously and the stability is difficulty to be assured. To solve the above problem, a large envelope wavelet neural network gain scheduling flight control law design method based on genetic algorithm is presented in this paper. Wavelet has good time accuracy in high frequency-domain and the good frequency accuracy in low frequency-domain. Neural network has the self-learning character. In this method, wavelet function instead of Sigmoid function as the excitation function. So the two merits are merged and the high nonlinear function approximation capability could be achieved. In order to obtain higher accuracy and faster speed, genetic algorithm is used to optimize the parameters of the wavelet neural network. This method is used in design the large envelope gain scheduling flight control law. This simulation results show that good control capability could be achieved in large envelope and the system is still stable when modeling error is 20%. In the situation of 20% modeling error, the maximum overshoot is only 12m and it is 35% of the maximum overshoot using normal method.

Original languageEnglish
Title of host publicationSixth International Symposium on Instrumentation and Control Technology
Subtitle of host publicationSensors, Automatics Measurement, Control, and Computer Simulation
DOIs
StatePublished - 2006
EventSixth International Symposium on Instrumentation and Control Technology: Sensors, Automatic Measurement, Control and Computer Simulation - Beijing, China
Duration: 13 Oct 200615 Oct 2006

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
Volume6358 I
ISSN (Print)0277-786X

Conference

ConferenceSixth International Symposium on Instrumentation and Control Technology: Sensors, Automatic Measurement, Control and Computer Simulation
Country/TerritoryChina
CityBeijing
Period13/10/0615/10/06

Keywords

  • Gain scheduling
  • Genetic algorithm
  • Large envelope flight control
  • Wavelet neural network

Fingerprint

Dive into the research topics of 'A wavelet neural network based on genetic algorithm and it's apply to gain scheduling flight control'. Together they form a unique fingerprint.

Cite this