A Study on the Image Effect of Motorcycle Design Features

1A Study on the Image Effect of Motorcycle Design Features
Chih-Fu Wu*, Chun-Ming Lien**, Jing-Yu Chiang***
*University of Tatung, Graduate School of Industrial Design, Taiwan, [email protected]
**University of Huafan, Department of Industrial Design, Taiwan, [email protected]
***University of Tatung, Graduate School of Industrial Design, Taiwan.
Abstract: Straddling type motorcycles have been popular among young people. However, there are few related researches on motorcycle form. This paper is to collect motorcycle samples and analyze the design elements. From relevant literatures, the model expressions including: streamline, textural, classic, practical, stability, recreative, taste and unique are carried out for measures of image analysis in 7-Level Assessment Scale. By principle component analysis, the result of investigation from young consumers has revealed image of motorcycle form is composed by three factors: “evaluative factor”, “individual factor” and “appearance factor.” Young people’s preferences are affected by two image expressions: “classic-alternative” and “unique-popular.” The researchers applied Quantitative I to other features, exhaust pipe, motorcycle body ratio, fuel tank, trail, engine, motorcycle head, seat cushion, electroplate, wheel rim, back handrail, and motorcycle frame are considered into image evaluation. The finding shows that the main factors affect image are body length ratio, electroplate decoration ratio, motorcycle head, exhaust pipe, trail, fuel tank. From this result, we can draw a conclusion that upper curve of motorcycle is the most important factor in motorcycle form and contrast can be made among motorcycle samples by using slope of contour segment. At last, by applying factor-reducing method on the main part of the upper curve, the result obtains the featured curves of non-geometric products. These discoveries give great helps to motorcycle design in Taiwan. Keywords: Motorcycle form, Principle component analysis, Cluster analysis, Image, Style curve
1. Introduction 1.1 Motivation and Objective In light of the statistic data reported by the Ministry of Transportation And Communications, the ownership for local people has reached 1.89 persons per motorcycle averagely in Taiwan area [1], therefore to develop new products which can meet the needs of consumers has been considered as a key factor by all enterprises for maintaining their competition ability in such a competing motorcycle market. Chan (1992) pointed out that in common features of products, form is the main term to know and recognize style [2]. During the traditional designing procedures, designers always design the products subjectively based on their personal experience, where this kind of design pattern is very difficult to effectively meet the consumer’s needs. Therefore, objectively finding out the viewpoint of consumers will facilitate an excellent interactive relationship between products and consumers [3]. Even though there are various researches relating to the image recognition and investigation for form and
image about product design features. However, some design features could not separate and explain by some design rules. How to separate the design features more detailed and owing to related research did not mention much on how to separate the design features of outline shape more detailed. This research aimed to the design features which influence form image of straddling type motorcycles mainly and deconstruct its characteristic curve of outline shape. Because the designers develop the motorbike design based on the side contour of vehicle, the side contour is an important index to whole design. The design of transport emphasizes on streamline and the contour is continuous curve; as a result, to analyze the side contour objectively is important and it could provide the reference for designers on motorbike outline design. This study took the straddling type motorcycle as an example and execute by means of the practical investigations and analyses, so as to acquire several primary objectives as follows:
and overall preference evaluation on design features of a straddling type motorcycle such as parts of exhaust pipe, fuel tank, engine, etc.
2. Establishing the product’s image configuration for the straddling type motorcycle
3. Studying the individual inference on the product form and design feature of straddling type motorcycle
4. Plotting out the characteristic line segments on the profile curve of the straddling type motorcycle.
1.2 Research Scope and Limitations According to the international standard [4] definition of motorbike, "Motor vehicles have two or three wheels or the net weight is not over 400 kilogram." Basing on the definition above, motorbike widely means motorcycle (Figure 1) and scooter (Figure 2) and mobile bike with two or three wheels and equipped with small exhausting engine and even include beach buggy with three or four wheels. The motorbike manufacturers mainly produce the products of previous two kinds.
Fig. 1 Straddling type motorcycle Fig. 2 Scooter
This study discussed style from the angle of form and to simplify the studying objects, the researchers chose the side view form of motorcycle as the studying object. Besides, to confer the influence of the engine form toward image, the selection of sample mainly is the motorcycles without side cowling and the interviewees will be able to observe engine form clearly. Owing to the present study is only focusing on the side view of motorcycle body, as to the side form of tire and rim will be considered as a simple geometrical pattern in round shape only, without any further discussion as well as analysis scope at all. The primary investigation objects are the young people group and the age is between 18 to 25 years old. 2. Method Kansei Engineering is defined as “translating technology of a consumer's feeling and image for a product into design elements” [5]. Tanoue et al.(1997)
have had the experience of applied to evaluate vehicle interior image, especially roominess and oppressiveness
[6]. And now, this study used the Kansei Engineering to research the image effect of motorcycle design features.. 2.1 Experiment Desig 1. Collecting Samples and Classifying Its Form
Features: This study collected sight draft of existing straddling type motorcycle from the market in accordance with the scope of research and acquired the recognition of interviewees on destructed product form by the questionnaire, and then eliminated the samples with similar form features based on their form feature.
2. Collecting and Selecting Related Sensational
Glossary: The researchers collected adjectives about vehicles and picked out the appropriate one from the questionnaire. After selecting, the researchers chose the representative samples with adjectives and provide for the purpose of image investigation.
3. The Experiment of Image Characteristic Extraction:
Setting up a primary image space after executing “Method of Semantic Differential (SD)” toward interviewees, and carrying out the Cluster Analysis with coordinates data of each sample locating in image space, and then studying the applicable sensational glossary which affects the teenager's preference by means of “Multiple Regression Analysis” as well as studying various form features which are capable to affect different images by quantifying the Category I.
4. The Experiment of Characteristic Curve Extraction:
Upon this stage, further study for the upper contour (fuel tank, seat cushion and tail curve) of straddling type motorcycle is carried out by quantifying the Category I. and then transformed the curve slope into dimensionless vector which represents the tendency of each profile respectively. Besides, the researchers extracted the characteristic curve which represents the upper profile, by the method of Principle Component Analysis.
2.2 Experimental Procedures 2.2.1 Collecting and Arranging Samples with Related
Image Glossary
The scope of this study focused on the straddling type motorcycles as well as collecting 2D side view samples and after initial arrangement, there were 48 samples from 10 different brands. Wherein the 10 samples were picked up randomly, and freely-selecting by 50 teenagers in the age of 18-25, in order to pick out the most attractive profile among them, and in this way the interviewees’ recognition on the features of motorcycle would then be acquired from the results. In light of such results, the 48 tested samples will be classified and eliminated by crossed classification. As to the sensational glossary, the first, 68 adjectives have been collected from the related reference, and 40 adjective modifiers will be obtained after deleting the inapplicable ones, and then an investigation on 188 motorcycle riders will be executed afterwards so as to pick out the 8 sets of adjectives which are more applicable than others. 2.2.2 Analyzing and Verifying the Influence of Form
Features against Overall Image During this experiment, the researchers transformed the tested samples into diagrams with resolution of 800*600 and displayed them on computer screen and used grayscale treatment to avoid the problem of color-shift. The experiment was proceeded by means of a questionnaire prepared according to 7-Level Assessment Scale provided by “Method of Semantic Differential (SD)”, and investigating 120 persons with questionnaire containing 8 image glossaries and 1 comprehensive evaluation. A Principle Component Analysis is carried out according to data obtained from the image investigation procedure, so as to set up the primary image space, and then carrying on a Cluster Analysis, wherein a Regression Analysis will be proceeded based on both the quantifiable form features and image scores. From this, the researchers will be able to study the applicable sensational glossary which affects the teenager's preference as well as comparing both the quantifiable category I and qualitative results of form features. 2.2.3 The Extraction of Characteristic Curve for the Form Profiles The further study on upper profiles (such as fuel tank, seat cushion, tail curve) of a straddling type motorcycle will be proceeded and in order to acquire a trustable result, firstly, the researchers will plot upper profiles of
motorcycle by means of AUTOCAD software, and then classify the fuel tank and tail portion into 25 segments respectively, so as to calculate its slope in-between two line segments, and acquire 50 data sets, which represent the upper curve of the straddling type motorcycle. Thus, the line segments would be reduced with the Principle Component Analysis (Figure 3).
Fig. 3 Quantification of Curve
0
5
10
15
20
1 3 5 7 9 11 13 15 17 19 21 23 25
3. Results and Analysis 3.1 The Results of Selecting Experimental Sensational
Glossary The researchers collected 40 adjectives of sensational glossary from related reference and proceeded investigation toward 188 motorcycle riders to gain more adaptable adjectives. The total amount is 721 items and 8 categories of adjectives have been selected much more than others, which include streamline, textural, classic, practical, stability, recreative, taste and unique. (Figure 4)
0
10
20
30
40
50
60
70
80
90
Streamline Textural Durable Practical Stable Recreative Taste Unique Techno Safe ect.
Fig. 4 Bar Chart of Sensational Glossary 3.2 The Results of Destructing the Form Features of Straddling Type Motorcycle In light of interviewees’ recognition on destruction of motorcycle form, the top 5 product forms are exhaust pipe, proportion of body, fuel tank, tail portion, and engine (Figure 5). The other parts selected including exhaust pipe, proportion of body, fuel tank, tail portion, engine, head, seat cushion, metal-plated parts, rim, back handrail and frame etc (Figure 6).
Exhaust pipe
Fig. 5 Bar Table of Attractive Ability of Design Features
Fig. 6 Deconstruct Draft of Motorcycle Design Features Tested samples are reduced from 48 separated by form design factors to 25 through eliminating the samples with similar form features according to results of destruction for the design features of motorcycle appearance to speed the procedure of experiment. (Figure 7)
Fig. 7 Tested Samples 3-3 The cluster of form features in image space for motorcycle In this experiment, the researchers transformed the tested samples into diagrams with resolution of 800*600, displayed samples on computer screen and used grayscale treatment to avoid the problem of color-shift. Total testees were 120 people, including 111 males and 9 females. The experiment proceeded 7-Level Assessment Scale provided by “Method of Semantic Differential (SD)” to 25 samples with 9 sets of image judge glossary (Table 1). From the result of straddling type motorcycle image,
there are significant differences between each sample. As a result, the researchers proceeded further cluster analysis and quantifiable category I to discuss the influence of features to image.
0
50
100
150
200
250
300
1










exhaust pipe proportion fuel tank engine tail portion head seat cushion frame metal-plated
Table 1. Image Judge Glossary
1 Streamline-Square 5 Steady-Handy 2 Textural-Rough 6 Recreative-Formal 3 Classical-Alternative 7 Taste-Vulgar 4 Practical-Decorated 8 Unique-Popular
9 Favor-Disfavor 3.3.1 The Causal Relationship between Preference and Each Individual Image Preference is an important considerable factor for general consumers to make decision on purchasing product, and this factor is mainly subjected to the image of consumer according to the product form. Therefore, if there is a possibility to find out further what kind of image would affect the preference of motorcycle against a consumer, then it will be a higher potentiality for the manufacturers to design a product which can meet the expectation of consumers. Therefore, in this chapter, the researchers will adopt the overall evaluation of favor-disfavor as dependent variable and the remaining 8 images as independent variable in order to study the co-linear for each individual adjective factor by analyzing the partial correlation value (Table 2), also defining the partial correlation value larger than 0.7 is a high correlation, and finding out the "textural-roughness" and "classic-alternative", "stability-handy", and "taste-vulgar" etc., all these three glossary sets have a high linear relationship, and "classical-alternative" and “stability-handy”, “recreative-formal” have a high linear relationship as well, where "stability-handy" and “recreative-formal”, "taste-vulgar" have a high linear relationship too, wherein these 5 adjective glossaries have interactive relationship. As a result, the researchers execute a “Multiple Regression Analysis” toward 8 glossaries. Where only the image factor of "classic-alternative” with a reliability value of 0.5692 (adjusted R square) has a obvious relationship. As there are 5 adjective glossaries all have a high linear relationship and it is potential to build-up a co-linear relationship; hence, the searchers have to give up partial variables, and "classical-alternative”, "taste-vulgar" these 2 items are presenting preferably the overall style of the product form. Thus, the researchers utilize the following 5 adjective glossaries, “streamline-square”, "practical-decorative",
"unique-popular", "classical-alternative" and "taste-vulgar" to proceed “Multiple Regression Analysis” against the “favor-disfavor” comprehensive evaluation, and the R square after adjustment was 0.5976. It means that these 5 variables have achieved a 59.76% explanation capability upon comprehensive evaluation, where it has an improvement of 0.0284(5%) on the reliability while comparing to the result without co-linear variables. For this reason, after considering the improved explanation capability upon comprehensive evaluation, the researchers decide to utilize a reduced model (Table 3, 4) instead in the analysis process in this study.
Table 2. Correlation Matrix of Image Streamline Textural Classical Practical Stability Recreative Taste Unique
Streamline 1
Stability 0.082307 0.740382 0.701031 0.120351 1
Recreative 0.338196 0.689592 0.736468 0.351644 0.746137 1
Taste 0.375986 0.826857 0.538096 -0.14162 0.74444 0.682747 1
Unique 0.228847 0.244241 -0.31463 -0.67144 -0.10853 -0.09527 0.444366 1
* Partial correlation value larger than 0.7 is a high correlation
Model Establishment Y= 5544332211 χβχβχβχβχβα +++++
Table 3. Result 1 of ANOVA
D.f. SS MS F Sig. Regression 5 7.111697 1.422339 6.799775 0.000862 Residual 19 3.974315 0.209174 Sum 24 11.08601 Table 4. Result 2 of ANOVA Y=Comprehensive evaluation
Coefficient Error st. t P Intercept 1.805942 1.74391 1.03557 0.313404 X streamline-square 0.188509 0.327032 0.576423 0.571096 X classical-alternative 0.920887 0.467671 1.96909 0.063703 X practical-decorative 0.161968 0.348651 0.464554 0.64753 X taste-vulgar -0.56956 0.461379 -1.23448 0.232075 X unique-popular -0.58955 0.347327 -1.69738 0.095944 R square 0.641502 R square adj. 0.547160
It can be found in the model built up by the present study that the P value is 0.000862, which is smaller than the significance level 0.1, indicating this model is a linear model. Where the model established is as follows: Favor score = 0.188509(streamline)+ 0.920887(classical)+ 0.161968(practical)- 0.56956(taste)- 0.58955(unique)+ 1.805942 (sort of Comprehensive value EX:like:1 dislike:7)
In light of the above model, we realize the P values of these 2 variables of "classic-alternative" and “unique-popular” are all less than the significance level 0.1, meaning that it has a significant influence on the comprehensive evaluation. While looking into the P value, we find out that the comprehensive evaluation is primarily affected by the "classic-alternative" image, secondarily the "unique-popular", therefore it is inferable that the gear-shiftable type motorcycle combined with classical and popularized style would be much more preferable to young consumers. 3.3.2 The Causal Relationship between the Image
Space and Form Features of Straddling Type Motorcycle
Averaging all scores of individual sensational glossary for every sample given by all interviewees and proceeding the statistic analysis by the method of Principle Component Analysis (Table 5), to generate a new principle component value and so as to acquire the coordinate locations of samples in image space. The Hierarchical Cluster Analysis will be carried on against all samples so that the researchers can study the relationship in-between each sample group and image space as well as its inclusive image tendency. The analysis results are shown in the following table. Table 5. Result 1 of Principle Component Analysis
Component 1 Component 2 Component 3 streamline 0.961 -0.125 qualitative 0.859 -0.363 0.234 durable 0.846 0.235 0.188 practical 0.829 -0.148 0.45 stable 0.727 0.461 0.419 recreative -0.923 0.27 taste 0.895 0.266 unique 0.192 0.968 Charact value 4.164 2.056 1.066 Rate of description 52.053% 25.702% 13.324% Accumulation of rate 52.053% 77.755% 91.79%
It has reached an accumulation explanation rate of 91.79% till the third principle component, and the characteristic value of three principle components are all more than 1, so we extract samples from the third principle component for analysis. These three principle components are named as the evaluative factor, individual factor, and formal factor. In evaluative factor, its characteristic value occupies 52.1% of total variation, the individual factor occupies 25.7% of total variation, and the formal fact is 13.3% instead, and all three factors occupy 91.8% of total variations. The first principle component consists of images, such
as “textural-rough”, “classic-alternative”, “stability-handy”, “recreative-formal”, “tasty-vulgar” etc., and all having meaning of product evaluation inclusively. Therefore, it is defined as the “evaluative” factor. The second principle component consists of images, such as “practical-decorative”, “unique-popular”, and these two adjective glossaries are showing a negative relationship, with a meaning similar to be practical and popularizes, and further, the high decorative is very similar to the unique image. Hence, this fact is defined as "characterized" factor. The third principle component consists of “streamline-square” image only, which is defined as the "formal" factor. In order to acquire the corresponding relationship between the sample and image factor, the researchers spread out scores of principle component of each sample in the image space created in-between two main axes (Figure 8, 9), and combining it with the results of cluster analysis (Table 6).
Fig. 8 Practical-Luxury Image Space Histogram
Fig. 9 Streamline-Practical Image Space Histogram
Table 6. Relation of Sample Group between Image Spread and Design Features
Image Space Spread Common Design Features
G1 Valuable Special
Lots of electroplate decoration, curve of tail and fuel tank are more round and the ratio of body ratio is obviously longer than other group
G2 Slightly Special Slightly Square
The height between fuel tank and seat cushion is large (the curve of fuel tank is high) and tail is more sharp
G3 Slightly Square Practical and not attractive
The height between fuel tank and seat cushion is large (the curve of fuel tank is high) and round, slanting tail and side exhaust pipe
G4 Streamline With cowling and the height between fuel tank and seat cushion is large (the curve of fuel tank is high)
G5 Practical and valuable
Unitive seat cushion with back handrail and the height between fuel tank and seat cushion is small (fuel tank curve is gradual)
G6 Cheap Square and unique
Extreme form, body ratio is foursquare and tail is short and thick
3-3.3 Summary It is found according to the results of experiments that the motorcycle equipped with a cowling has a streamline image obviously, and in addition, the motorcycle with a longer body ratio as well as sharper tail portion has also related to the streamline image. Originally, the experiment would like to predict the difference in-between the high/low positioning of the fuel tank and seat cushion, as well as the influence of curve of fuel tank against the overall image of the motorcycle, but the result turned out has no any difference at all, on the contrary it can affect the length of overall body ratio. Therefore it is predicted that the style of curve is impossible to be simply represented by the altitude difference of partial lines accordingly. The plated parts, body’s reflecting area and the length of body ratio are all capable to affect the overall “luxury-normal” image. The increasing numbers of plated parts and the longer body ratio has a higher score in luxury evaluation. The style of seat cushion and the rear seat’s arm-rest would affect the practical and popularized evaluation by consumers, wherein an integral type seat cushion has a tendency of common or normal image than the motorcycle with a separated seat cushion. All samples equipped with rear seat’s arm-rest having a tendency of practical and popularized image. The thickness of tail portion and the volume of exhaust pipe will affect the image of "stability-handy", where the larger volume of tail and exhaust pipe representing a more stable image. While performing a Regression Analysis against the major body ratio, it is found that both images of "stability-handy" and "recreative-formal" have become more significant. According to the analysis on 8 image factors related to
the comprehensive evaluation of "favor-disfavor", it is found that "classic-alternative” and “unique-popular" showing significance against the comprehensive evaluation, and the favor and classic, popularization showing a positive linear relationship. The general image presumes that the young men prefer unique, distinguish, hence having a preference for the motorcycle with a more courageous and alternative design. But the research results revealing that the most young men still prefer a motorcycle with a more medium level appearance, additionally, the researchers also realize although the image score for each type of motorcycle is different among each other, but on overall evaluation, the final summation of each sample has a tendency of favor, where the researchers can explain that young man has a high acceptability against various style products, or it could even be explained that the young person has a aspiration to the motorcycle. 3.4 The Causal Relationship between Form Features and Image of Straddling Type Motorcycle To further understand the causal relationship between image and form design features of straddling type motorcycle, in this section, the researchers will adopt each image score in the “Method of Semantic Differential (SD)” as the dependent variable, and set up factor of form design as the independent variable, so as to perform the Lin’s “Quantification I” which analyzes the influence of each partial feature against image. Where the 9 sets of features are defined as the dependent variable and image score in defined as the independent variable, moreover the partial non-quantitative features categories (such as, fuel tank, tail seat etc.) would be classified by 5 experienced designers after an interactive discussion. The specification related to 9 sets of qualitative characteristics is as follows: 1. Head: There are three categories, such as: cowling,
wind goggle, external gauge.
Fig. 10 Types of Head 2. Body ratio: There are three levels, which is classified
into 3 groups by 1.87, 2.07 according to the data distribution of samples.
Fig. 11 Types of Muffle 3. Exhaust pipe: Positioned at right underneath the tail,
positioned at one side and extending upward inclined, positioned at one side and extending horizontally, which is classified by its installation method.
4. Tail: There are four categories, which is classified by
the style.
Fig. 12 Types of Tail 5. Rear handle: There are two categories, included or
excluded.
Fig. 13 Types of rear handle 6. Fuel tank: There are five categories according to their
shape.
Fig. 14 Types of Fuel Tank 7. Plated decoration: It is classified by the proportion of
plated portions at engine, and exhaust pipe etc. 8. Frame color: Due to it is too complicated to be
classified, so the researchers just adopt two categories, such as black color, and silver color.
9. Engine style: It is classified according to its side view,
such as: In-line type engine, V-type engine, Double side cylinders.
Fig.15 Types of Engine 3.4.1 The Result of “Quantification I” Category The factors consist of image “evaluation” for existing straddling type motorcycle are mainly subjected to the following characteristics, such as: body length, proportion of plated parts, head, exhaust pipe and fuel tank. The higher proportion for the body length and plated decorative parts, the higher tendency toward the noble image will be. Additionally, the little wind goggle at head portion, parallel-positioned side exhaust pipe, and water-drop shape anti-looking fuel tank is also a major characteristic which is capable to affect the noble image. The factors consisting of image “individuality” is mainly subjected to the following characteristics, such as: tail, head, fuel tank, exhaust pipe, body ratio, plated decorative parts, and engine type. The shorter and sharper tail similar to sport bike has a tendency toward unique image, instead, hippie type motorcycle, obtuse and flat tail is more acceptable to the public. The head with cowling, anti-looking fuel tank, tail-positioned type exhaust pipe are affecting the unique image, the longer body and larger proportion of plated decorative parts, the higher tendency toward the unique image will be, in addition, the motorcycle equipped with double side cylinders type engine seldom found in domestic market is also one of the affecting factor to unique image as well. The factors consisting of image “formation” is mainly subjected to the following characteristics, such as: head, tail, body ratio etc. The characteristic of cowling is obvious affecting to the streamline image of motorcycle, and the second is the characteristic with a little wind goggle, the open type gauge without any windshield having a tendency to the square image. The body ratio and streamline level are in direct proportion, wherein the longer body ratio tends to the streamline. The results for tail are similar to body ratio, wherein the characteristic of longer tail is affecting the streamline image, which is mutual evidence with the result of the loner body ratio regarding the streamline image. Moreover, the longer and sharper tail, which is similar to sport bike, has a meaning
of streamline image as well. It is found obviously in the image space after the principle component analysis procedure that the motorcycle with rear handle and flat seat cushion tend to the practical image, which has a partial difference with results of practical images only related to tail from the “Quantification I” category. The reasons might be that the most of rear handles are equipped onto the flatter tails, but there are at least four tail variables in the tail style cluster, so there is a possibility that influence of handle feature has been weakened due to there are only two variables (with or without). If comparing the quantifiable design feature against adjective glossary in the Regression Analysis with results of “Quantification I”, it is found that both rear handle against the “practical-decorative” image and body, plated proportion against the “stability-handy” image are various upon the Regression Analysis. However, its partial correlation value is lower (about 0.55) instead, and the reason might be probable that only the quantifiable characteristics would be examined upon the Regression Analysis procedure, and the remaining qualitative characteristics might be even more obvious after the Regression Analysis procedure. 3.4.2 Verifying From the result of Quantification I, it shows the influence tendency of design feature tower each image and this chapter is to verify the accuracy of previous research result. The experiment method is the same as previous image questionnaire, and the researchers choose the other three types of motorcycle which is similar to original sample group and investigate the same interviewees. The returned questionnaires are 90 and it can get 9 sets score of actual testees’ form image toward three motorcycles. Besides, according to the previous result of Quantification I, the researchers separated the form of those three motorcycles into nine design features. From the sum of each feature score, the researchers could gain the image tendency of that motorcycle. This research used 7-Level Assessment Scale and to level 4 means no image feeling and the researchers add image tendency in level 4 to get predicted image scores. Comparing the actual score and predicted score and took the higher difference between actual and predicted score to divide by actual score. The researchers got the inaccuracy rate and accuracy rate (1-inaccuracy rate) and found out the
predicted accuracy rate of two groups averaged 80%. The accuracy rate of four items, “textural-rough”, “practical-decorated”, “steady-handy” and “taste-vulgar” reached above 95% and “classical-alternative” and “recreative-formal” was lower 70%. As to another predicted model, the accuracy rates were very low and the prediction underestimated the positive feelings of testees. The average positive image scores were high and only “practical-decorated” this set had the accuracy rate at 97%. The researchers presumed that this motorcycle is the high price model of Italy famous company and although there was no mark in appearance, the testees still could recognize the type which influenced the experiment result. 3.4.3 Summar In this part, the researchers are carrying on cluster analysis and Quantification I analysis on the results from the image examination of the Method of Semantic Differential (SD), where the style of existing straddling type motorcycles on the markets would be classified into clusters, and it is available to observe roughly the primary clusters for the straddling type motorcycle on the markets as well as primary characteristic for each cluster. However, it is still impossible to find out the causal relationship between form features and images. Therefore, the researchers need to compensate it by results of the “Quantification I” Analysis process, and then find out the influence of design features upon images by means of the partial correlation value of each design feature against different image. If compared the difference in results of adjective glossary against comprehensive evaluation in the Multiple Regression Analysis with results of Quantification I; it is found that both “classic-alternative”, “unique-popular” are the primary image glossaries upon the Multiple Regression Analysis. However, the results from the “Quantification I” Analysis revealing the design features with classical image included the little wind goggles, shorter body ratio and more plated decorative parts…etc. The design features such as, cowling, longer body ratio, short tail, water-drop shape and streamline fuel tank and double side cylinders type engine have the unique image, which is almost identical to design features for final overall preference evaluation, where results resulted from these two statistic analyses is mutual evidence. In light of the final overall comprehensive evaluation
revealed by the experiments, it is found that the teenager groups prefer more handy and common design features, which is contrary to the general concepts on the young people, such as chasing modernization and individualization etc. The color of frame has no obvious image influence toward the whole image. The wind goggle has obvious influence of the positive image like streamline, textural, classical, steady, recreative and taste while the long body ratio and high plated accessories ratio tend to valuable adjectives. As to the installing method of exhaust pipe, most of teenagers prefer side-inclined installation, instead of the advanced and anti-looking design features. Regarding to the tail style, the loner tail has a higher score in every positive image which is the same in overall evaluation. On the contrary, the young teenagers dislike advanced short tail and anti-looking of hippie type tail. The rear handle only has a bigger influence against image of classic and stability, and it is a big difference from the image space initially wherein the motorcycle with a rear handle tends to an axial result of practical image. The anti-looking fuel tank has a positive image overall and only water-dropped shape with ancient streamline is disliked. As to the engine, the double side cylinders engine used in European motorcycle has an overwhelming positive images, which has textural, stability and unique images, but it tends to a fancy decoration image in “practical-decorative” adjective glossaries, therefore it is not acceptable to the young people. Most of the young teenagers in the experiment were males and the rate between male and female, female only occupied 3.8% (only 7 people). As a result, the researchers did not discuss the difference of gender toward image. From the analysis result of Quantification I, the researchers could understand that each image was influenced by different form design factors. According to these form design factors, the designer could design the straddling type motorcycles could more conform to the images. From this result, it can prove different form brings different image. The designers should know the causal relation between form and image and then design the products fit in with users requiring image. 3.5 The Extraction for Upper Contour Curves of Motorcycle From above result, the researchers found the influence
of head, tail, fuel tank and body ratio features toward the whole image. Also, the researchers could know the main influence to straddling type motorcycles image is the upper contour curve of motorcycle (Figure 16). Due to the difference of head style is large and the type of contour is complicated, the researchers distinguish the features according to the difference of accessories. In this study, the researchers tried to utilize the Principle Component Analysis to destruct the principle profile segments of characteristics from the body curve.
Fig. 16 Upper Curves of Straddling Type Motorcycle
To gain reliable segments, the primaru step is to make the curve into quantification. Firstly, the researchers used AUTO-CAD software to draw the upper contour of motorcycle and all the head leaded to right and set the end of tail as the origin. Besides, the researchers enlarged the 25 tested samples to same length by two-dimension and equally divided fuel tank and tail each into 25 parts and recorded total 51 coordinates and count the slope between two lines. The upper contour curve of motorcycle body was mainly constituted by fuel tank and seat cushion and the researchers divided it into 25 parts each and select coordinates to count the slope which could get dimensionless data. There were 25 data of slope each for fuel tank and seat cushion (Table 7) and total 50 sets data which could present the upper curve of straddling type motorcycle. Besides, the researchers used principle component analysis to reduce the segments.
slope YY XX
12
12
Table 7 Fuel Tank Segmental Curve Slope of Sample Slope 1 Slope 2 Slope 3 Slope 4 Slope 5 3.017 1.481 1.099 0.9115 0.6415
Slope 6 Slope 7 Slope 8 Slope 9 Slope 10 0.5455 0.4575 0.339 0.215 0.1185
Slope 11 Slope 12 Slope 13 Slope 14 Slope 15 0.0545 0.014 -0.016 -0.046 -0.0805
Slope 16 Slope 17 Slope 18 Slope 19 Slope 20
-0.119 -0.1575 -0.1865 -0.2205 -0.258
Slope 21 Slope 22 Slope 23 Slope 24 Slope 25 -0.31 -0.4 -0.752 -1.554 -1.8055
In principle component analysis, it is unavoidable that some variables present negative linear relationship and the researchers chopped off these variables unconditionally and inserted into the closest segment (Table 8).
Table 8. Table of Principle Component Analysis
3.5.1 The Result of Characteristic Curves Extraction The searchers compared each segment slope among each sample by means of the Principle Component Analysis, and decreased the co-linear slopes with similar variation into one principle component, and acquired 11 principle components which are all larger than 1(Table 9). Moreover, the 50 spacing intervals would be divided into 11 segments in total (Figure 17), where each segment would act as one single component except 25 and 26 which are connected to fuel tank of tail seat (Table 10). The variables remained in the component are all successive segments, in which the tail seat cushion has 6 segments (Figure 18) and the fuel tank curve has 5 segments (Figure 19).
Fig. 17 Segments Diagram of Upper Curves of Motorcycle
Fig. 18 Segments Diagram of Tail Seat Feature Curve
Table 9. Classified Table of Tail Seat Feature
Classified Percentage Explain Feature 1 32% Tail lantern style and the end of
back seat cushion Feature 2 20% Main curve of back seat cushion Feature 3 16% The transforming curve segment
of front and back seat cushion Feature 4 12% Later half part of front seat
cushion Feature 5 16% Main part of front seat Feature 6 4% The end of seat cushion
connecting to fuel tank
Table 10. Classified Table of Fuel Tank
Classified Percentage Explain Feature 7 4% The connecting part of the fuel
and the end of seat cushion Feature 8 8% Lean part of user’s lower body Feature 9 28% The curve raises gently to the top
of fuel tank Feature 10 44% Main curve of the top of fuel tank Feature 11 16% The end segment between the end
of fuel tank and frame 3.5.2 Summary While utilizing a great deal of curve data, it is applicable to extract the principle components by the factor decreasing method; therefore, the researchers extract slopes for comparison among samples, and then extracts the feature curves of products, which is capable to rapidly extract the feature profiles from the products without form specification. Thus, it is applicable on products with irregular geometric profile as well.
4. Conclusion and Suggestions 4.1 Preliminary Results This study proceeded from the collection and sieve of glossaries and samples to the experimental designs and analyses, where the general conclusion could be summarized as the following items: 1. The recognition of interviewees on the destruction for
product form of a motorcycle could be found based on the results of attractiveness questionnaires and the researchers found out that the top five attractive product forms are: exhaust pipe, body ratio, fuel tank, tail portion and engine. There was a very high proportion lying on the exhaust pipe.
2. According to the “Method of Semantic Differential (SD)” and other related analysis method on the image glossaries related to straddling type motorcycle from the young people, the researchers found that the image structure of existing straddling type motorcycle consisted of three principle adjective glossaries, such as: "evaluation", "individual", and "formal".
3. The design features for product form of the existing straddling type motorcycles can be classified into six principle categories by the Cluster Analysis method.
4. Generally speaking, the “noble-cheap” image consisting of factors for the existing straddling type motorcycle are mainly subjected to the following characteristics. The “unique-usual” image consisting of factors are mainly subjected to the following characteristics. The “streamline-square” image consisting of factors are mainly subjected to the following characteristics.
5. The longer body ratio and the higher proportion for plated parts having a tendency toward adjective glossaries of valuation. Besides, as to the installing method for exhaust pipe, most of teenagers prefer side-inclined installation which is popular currently. Regarding to the tail style, the loner tail has a higher score in every positive image and is also very popular in overall evaluation by the young people. Instead, they kept away the advanced short tail and anti-looking of hippie type tail. The rear handle only has a better tendency against image of classic and stability; however, while it is a big difference from the image space initially, wherein the motorcycle with a rear handle tends to an axial result of practical image. Although the anti-looking fuel tank has a positive
image overall, it is not popular among young people, and as to the engine, the double side cylinders engine used by the European designed motorcycle has a total overwhelming positive image, which has the textural, stability, unique images, but it tends to a fancy decoration image in “practical-decorative” adjective glossaries. Therefore, it is not acceptable to the young people.
6. According to the analysis on 8 image factor sets related to the evaluation of "favor-disfavor", it is found that only "classic-alternative” and “unique-popular" showing significance against the comprehensive evaluation. It presents that the young people prefer unique and popular form motorcycles but not with courageous and alternative design, which is different from the general concepts of rebellion and unique characters. But instead, they might be affected by their colleagues, where they prefer the more classic and mature designs. Above conclusions are verified by the results from the “Quantification I” Analysis.
7. The researchers utilize the method of Principle Component Analysis to divide the feature line segments and try to destruct the feature line segments of non-geometric curve, and at last found the rear seat cushion could be destructed into 6 segments of characteristic curve, including: "tail lantern and the end of back seat”, “the major portion of rear seat cushion”, “curve for the turning portion between front/rear seat cushion”, “backrest portion of front seat”, “the major portion of front seat”, and “end of seat cushion connected to fuel tank” etc. As to the portion of fuel tank which could be divided into 5 segments of characteristic curve, including: “connecting end of fuel tank and seat cushion”, “lower supporting portion of rider”, “curve extending slightly to top of fuel tank”, “major curve of fuel tank”, and “end line segment in-between the end of fuel tank and frame”. There are total 11 feature segments in upper portion of whole straddling type motorcycle.
In light of the above conclusions, it is available to be used in database provided for designing the classification of data. In addition, it is verified that the factor decreasing method could be utilized in extracting the principle component, which would greatly facilitate the related designing for the product without any form specification.
4.2 Future Study and Recommendations According to the above-mentioned research results and limitations, thereby the present study discloses the several recommendations for the future follow-up researches: 1. If the researchers might be able to perform the
investigation for samples in the 1:1 plane or entity, so that it will have an improvement on the accuracy for the results of the whole research.
2. The interviewees picked up upon the proceeding of study are almost male with interests of motorcycle, but if there is a possibility to put in more appropriate population variation to act as the sub-issue for studying, and in this way, the study and discussions would cover a broader scope provided for a much more comprehensive discussion.
3. Looking thoroughly into the influence caused by the feature line segments and images, which could be studied further in near future.
5. Acknowledgement This study was helped by SooChow University Motor association (SCUMOTOR,http://scumotor.com.tw:8080/). This help is gratefully acknowledged. 6. Reference 1. Taiwan Transportation Vehicle Manufacturers
Association, ” Taiwan Vehicle's Industry's Overview”, Vehicle Industry Monthly, 114(2003).
2. Chan, C. S., Can style be measured? Design Studies. 21(3): 277-291(2000).
3. Ishihara, S., Ishihara, K., Nagamachi, M. and Matsubara, Y., “An automatic builder for a Kansei Engineering expert system using self-organizing neural networks”, International Journal of Industrial Ergonomics, 15(1): 13-24(1995).
4. International Organization of standardization (ISO) 3833:1977, “Road vehicles-Types-Terms and definitions” (1976).
5. Nagamachi, M., Kansei Engineering: A new ergonomic consumer-oriented technology for product development, International Journal of Industrial Ergonomics, 15(1): 3-11(1995).
6. Tanouea, C., Ishizakab, K., and Nagamachic, M., Kansei engineering: A study on perception of vehicle interior image, International Journal of Industrial Ergonomics, 19( 2): 115-128(1997).