Analyzing the Relationship between Automobile Head-up Display Presentation Image and Drivers’ Kansei Imagery

本研究係針對汽車用抬頭顯示器顯示內容之設計架構為重點，以感性工學為基礎，建立一感性工學系統，藉由質化與量化分析，進而推論出型態要素與感性意象之關聯性。根據研究中問卷調查分析結果，以市面上汽車用抬頭顯示器樣本中，選出三張，並進行研究驗證，以說明本研究的準確性與可行性。 本研究問卷調查部分分為兩階段，第一階段目的為廣泛蒐集市面上抬頭顯示器樣本，從中分析出抬頭顯示器顯示內容型態要素表，並以田口法L-18直交表繪製出18張新樣本，並從中挑選6張代表性樣本。接著，蒐集形容抬頭顯示器的感性語彙，經過兩次萃取留下32對感性語彙，利用此32對感性語彙配合6張抬頭顯示器樣本，進行第一階段問卷調查，結果以因素分析法與集群分析法將32對感性語彙縮減至最終的5組語彙。第二階段依據第一階段的抬頭顯示器樣本為主軸，稍作修正，重新建構18張新樣本，利用新建構的18張樣本與5對感性語彙進行第二階段問卷調查。結果以數量化I類分析，分析對象分成全數受測者、高年齡層受測者、低年齡層受測者、男性受測者與女性受測者探討研究，依數量化I類分析結果，建立數量化I類評估預測模型。 研究驗證樣本採市面上抬頭顯示器樣本共三張，並以上階段建立的數量化I類評估預測模型計算出三張驗證樣本之預測值。接著，將三張驗證樣本與5對感性語彙進行驗證問卷調查。以驗證結果與評估預測值進行單一樣本T檢定，並分析討論其驗證結果與評估預測模型之合理性。Based on Kansei Engineering, this study explored the relationships between HUD (head-up display) presentation image physical properties and drivers’ Kansei images by quantitative and qualitative analysis. Finally, three existing commercial HUD presentation images on the market were used to verify the accuracy and feasibility of the results. The questionnaire suvery in study was divided into two steps. First, existing HUD presentation images were collected and analyzed. Eighteen new samples, using Taguchi L-18 table, were created, and six out of them were chosen as representative samples. Meanwhile, 32 pairs of Kansei words describing HUD image designs were collected and developed. These Kansei words were then arranged with the six representative samples to create the first-step questionnaire. The results were analyzed using factor analysis and cluster analysis, and the original 32 pairs of Kansei words were classified and reduced into five pairs of final representative Kansei words. The second step was to modify the HUD image samples in the first questionnaire survery and reconstruct them into 18 new samples. Then, the second-step questionnaire was developed by combining the new 18 samples with the five pairs of final Kansei words. The results were analyzed using Quantification Theory Type 1 (QT1) method based on subjects’ age and gender. According to the results, a predict model based on QT1 was generated. To test the validity of the results, three existing commerical HUD presentation images were chosen and validity questionnaire survey was conducted. The predicting Kansei values were calculated using QT1 predicting model built in the second-step. Finally, the validity and feasibility of the results were tested using one-sample t-test