Title :
A modeling method based on Wavelet Support Vector Machine
Author :
Wang, Shuzhou ; Meng, Bo ; Tian, Huixin
Author_Institution :
Sch. of Electr. Eng. & Autom., Tianjin Polytech. Univ., Tianjin, China
Abstract :
Neural networks with good nonlinear mapping abilities can be applied to build simulation model of helicopter. But they have some difficulties such as hardness of selecting network structure, slow convergence speed, local minimum, and over-fitting. To avoid above problems, a modeling method based on Wavelet Support Vector Machine (WSVM) is proposed. Marr wavelet is used to construct wavelet kernel. And the rationality of the multidimensional wavelet kernel is proved. Based on pretreatment of practical flight data, rotational speed model for landing process of helicopter with rotor self-rotating is built with WSVM. Compared with neural network model, WSVM model possess some advantages such as simple structure, fast convergence speed and high generalization ability.
Keywords :
aerospace computing; aircraft control; helicopters; neural nets; support vector machines; wavelet transforms; Marr wavelet; convergence speed; helicopter landing process; helicopter simulation model; modeling method; multidimensional wavelet kernel; network structure; neural networks; nonlinear mapping; rotational speed model; wavelet support vector machine; Aerodynamics; Aerospace simulation; Convergence; Helicopters; Kernel; Neural networks; Pattern classification; Rotors; Support vector machine classification; Support vector machines; Helicopter; Simulation Model; Support Vector Machine; Wavelet Kernel;
Conference_Titel :
Control and Decision Conference (CCDC), 2010 Chinese
Conference_Location :
Xuzhou
Print_ISBN :
978-1-4244-5181-4
Electronic_ISBN :
978-1-4244-5182-1
DOI :
10.1109/CCDC.2010.5498645
