Research on Performance-oriented Perceptual Image Evaluation Method for Cus Tomized Products

FENG Yixiong,LOU Shanhe,WANG Xupeng,ZHENG Hao,GAO Yicong,WANG Yun,TAN Jianrong
DOI: https://doi.org/10.3901/jme.2020.09.181
2020-01-01
Journal of Mechanical Engineering
Abstract:The demand for mechanical products design has increased from the functional level to the emotional level. The function of mechanical products is no longer the only focus. The performance of mechanical products, especially customized products, has attracted more and more attention. The perceptual image is the impression of the user on the performance of the product, which determines whether the product can meet the high-level emotional needs of the user. A performance-oriented method for evaluating the perceptual image of customized products is presented. The main basis for verifying the performance of products is real-time acquired physiological signals of customers. To evaluate the physiological signals of customers, the objective evaluation method and subjective evaluation results are employed, which is able to reflect the performance of customized products accurately. This study takes evaluation of high-speed elevator's ride performance as an example. By wavelet packet transform and random forest algorithm, feature extraction and classification of collected passengers’ EEG signals in the elevator’s operation are respectively implemented. A perceptual image evaluation model of high-speed elevator’s ride performance is established and the reduction of EEG data characteristics is realized. According to the proposed method, passengers can express the pleasure or comfort quantitatively, which provides a new perspective for exploring the evaluation method based on the perceptual image.
What problem does this paper attempt to address?