An integrated decision-making method for product design scheme evaluation based on cloud model and EEG data. (January 2020)

• Record Type: Journal Article
• Authors: Lou, Shanhe; Feng, Yixiong; Li, Zhiwu; Zheng, Hao; Tan, Jianrong
• Is Part Of: Advanced engineering informatics. Volume 43(2020)
• Publication Date: 2020-01
• Subjects: Design scheme evaluation -- Multi-criteria decision making -- Cloud model -- EEG; Computer-aided engineering -- Periodicals; Engineering -- Data processing -- Periodicals; 620.00285
• DOI: 10.1016/j.aei.2019.101028
• Languages: English
• ISSNs: 1474-0346

Abstract

Highlights: An integrated decision-making method is proposed to evaluate product design schemes. The cloud model is applied to facilitate the evaluation process of experts. The EEG data of customers when they are taking the elevators are analyzed. Abstract: Selecting the optimal design scheme is a vital task in the product design area. It not only improves the performance of the product, but also leads to the greatest satisfaction of customers. However, existing methods express qualitative evaluation information roughly, and none of them has taken the implicit psychological states of customers into consideration. Therefore, an integrated decision-making method for product design scheme evaluation is proposed. This method applies the cloud model to facilitate the evaluation process of experts and uses the EEG data to reveal the psychological states of customers. Benefit from the probability theory and fuzzy set theory, the cloud model deals with the fuzziness and randomness simultaneously. It can decrease the cognitive discrepancy of experts and allow the information distortion to be neutralized to a great extent. Since the experts are not the final users of products, the evaluation results from experts cannot truly reflect the psychological states of customers when they use the product. An experiment is designed to collect the EEG data which can reveal the implicit psychological states of customers. The recorded data are segmented based on the operation process and tagged with the self-reported psychological states. Subsequently, the wavelet packet decomposition is applied and the sample entropy of each EEG frequency band is extracted as the feature. Taking advantage of the random forest classifier, the psychological states of customers can be classified with the average accuracy of 90.76%. This study can lead to a practical system for automatic assessment of psychological states in future applications. The evaluation process of elevator design schemes is conducted as a case study to illustrate the feasibility of the proposed method.