Car Styling Perceptual Modeling Based on Fuzzy Rules

Abstract:

Abstract. In today’s automotive market with intense competition, car styling preferences directly affect designer’s design outcome. At the meantime, there are various types of car styles, however users’ emotional evaluations of them are messy. Thus establishing a perceptual extraction model of vehicles is necessary. Under common circumstances, we quantify the emotional human perceptual through digitized human perceptual data obtained by questionnaires. However, it is emotional and thus difficult to analyze. In order to analyze perceptual data, the general method is statistical methods, such as factor analysis system. However, the use of statistical methods alone is not sufficient to handle the emotional data, because this method cannot handle non-linear intrinsic emotional data. This paper will discuss the using of fuzzy rules and through inductive data analysis to extract a common perceptual model of consumers’ emotional tendencies towards vehicles. In this paper, methods of variance predicted generalized regression neural network (VP-GRNN) and fuzzy adaptive resonance theory (fuzzy ART) were used to build a common perceptual model.

• Periodical: Applied Mechanics and Materials (Volumes 201-202)
• Pages: 794-797
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.201-202.794
• Online since: October 2012
• Authors: Xin Ru Wan, Jian Ming Che, Lu Han
• Keywords: Car Styling, Fuzzy Adaptive Resonance Theory, Neural Network (NN)

References

[1] D.F. Specht: IEEE Transatcions on Neural Networks, Vol.2 (1991) No.6, pp.567-576.

[2] E. WMLee, C.P. Lim, R. KK Yuken, and SMLo:IEEE Transactions on System, Man, and Cybernetics-PartB: Cybernetics, Vol.34, No. 2, pp.951-960, 2004.

[3] G.A. Carpenter and S.Grossberg: Special Issue on Artificial Neural Systems, Vol. 21 (1988) pp.77-88.