Longitudinal Aerodynamic Modeling Based on Support Vector Machine

Abstract:

In order to accurately depict the longitudinal aerodynamic characteristics of flight vehicle, a new aerodynamic modeling method based on Support Vector Machine (SVM) is proposed. SVM is a machine learning method that solves the problem by mean of optimization method on the basis of statistics learning theory, and introduces the structural risk minimization to coordinate the relationship between fitting degree and generalization ability, and the training samples are mapped into high dimensional feature space for linear regression, which can solve the practical problems such as nonlinearity, high dimension, over-learning, local minima, curse of dimensionality and so on. The experimental result shows that the method has a good modeling accuracy and speed training, and also is easy to implement. It is fully competent for longitudinal aerodynamic modeling for flight vehicle.

• Periodical: Applied Mechanics and Materials (Volume 327)
• Pages: 290-293
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.327.290
• Online since: June 2013
• Authors: Xu Sheng Gan, Jing Shun Duanmu, Jian Guo Gao
• Keywords: Aerodynamic Modeling, Flight Vehicle, Kernel Function, Structural Risk Minimization, Support Vector Machine (SVM)

References

[1] D. J. Linse, R. F. Strengel. Identification of aerodynamic coefficients using computational neural networks. Journal of Guidance, Control and Dynamics, 16(6), (1993), 1018-1025

DOI: 10.2514/3.21122

[2] K. Rokhsaz, J. E. Steck. Application of artificial neural networks in nonlinear aerodynamics and aircraft design. Journal of Aerospace, 102, (1993), 1790-1798

DOI: 10.4271/932533

[3] G. Peterson, W. Bond, R. Germann, et al. Using neural networks for aerodynamic parameter modeling. Proceedings of the American Control Conference, (1995), 1360-1361

DOI: 10.1109/acc.1995.520972

[4] Y. B. Meng, J. H. Zou, X. S. Gan, et al. Research on WNN aerodynamic modeling from flight data based on improved PSO algorithm, Neurocomputing, 83, (2012), 212-221

DOI: 10.1016/j.neucom.2011.12.015

[5] R. V. Jategaonkar. Flight vehicle system identification: a time domain methodology. Progress in Astronautics and Aeronautics Series, 1st Edition, AIAA, (2006)

[6] R. A. Hess. On the use of back propagation with feed forward neural networks for the aerodynamic estimation problem. Proceedings of AIAA Atmospheric Flight Mechanics Conference, (1993), 233-241

DOI: 10.2514/6.1993-3638

[7] V. N. Vapnik, S. Golowich and A. Smola. Support vector method for function approximation, regression estimation and signal processing. In M. Mozer, M. Jordan, and T. Petsche, editors, Advances in Neural Information Processing Systems 9, Cambridge, MA, MIT Press, (1997), 281-287