JPH03201169A - Designing device - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Objective of the Invention] (Industrial Application Field) This invention provides a design for displaying a vehicle design on an output means such as a CRT while confirming the image a person has of the vehicle. Regarding equipment.

(Prior art) This type of design device includes, for example, March 1988; Published by the Institute of Electronics, Information and Communication Engineers;
(There is something like the one shown in Figure 16 described in , 71N0.3pp, 245-247J. This is a computer memory that stores the appearance of a house, the entrance, a Western-style room, a Japanese-style room, a kitchen, a bath, etc.) A knowledge database containing housing design elements, evaluation terms such as luxurious, spacious, and spacious as adjectives expressing feeling, knowledge that defines the relationships between evaluation terms, and knowledge that defines the relationships between these evaluation terms and housing design elements. This is memorized, and a human operates an evaluation terminology input means such as a keyboard to enter, for example, a limited term for housing design such as "Japanese-style room," which represents a design element that is a component of a house, and also inputs the term "Japanese-style room" that represents a design element that is a component of a house. When you input the evaluation term "spacious" that describes the feeling in your image into a computer, the computer uses the input evaluation term from the knowledge database to infer and extract appropriate Japanese-style room design elements and displays them on the screen. It is configured to display images.

(Problems to be Solved by the Invention) By the way, if the technical evaluation terminology of the above-mentioned design device is adapted to the vehicle, it can be put to practical use as a vehicle design device. With such a design device, when a single evaluation term is input, design elements corresponding to it are inferred and extracted 100%, so there is no problem, but when multiple evaluation terms are input, the difference between the evaluation terms For example, sporty (8
There is a problem in that it is not possible to make subtle expressions for specifying the interior of a vehicle, such as 0%) and 20%.

Therefore, an object of the present invention is to provide a design device that can input subtle expressions that suit various human senses.

[Structure of the Invention] (Means for Solving the Problems) In order to solve the above problems, the present invention, as shown in FIG. an input means CL2 for inputting a term; a calculation means CL3 for inferring a design from the knowledge database based on the input evaluation term;
It is equipped with an output means CL4 for outputting this inferred design, a weighting means CL5 for weighting the input evaluation terms when there is a plurality of these terms, and a correction means for correcting the inference by the calculation means CL3 by this weighting. Means C
L6 is provided.

(Operation) When an evaluation term is input through the input means CL2, a design is inferred from the knowledge database CL1 based on the input evaluation term, and this inferred design is output to the output means CL4.
is output by When there is a plurality of input evaluation terms, and this input is a delicate expression, and it is necessary to weight the terms, the weighting means CL5 performs this, and the correction means CL6 adjusts the calculation means CL3 according to this weighting. The output means CL4 corrects the inference based on the above and outputs an output according to the delicate input of expression.

(Example) This invention will be described in detail below.

FIG. 2 shows a schematic configuration diagram of one embodiment, and this design apparatus is composed of a computer 1, a keyboard 3, and a CRT 5.

The computer 1 includes a memory M as a knowledge database CLI and a CPU 7 as a calculation means CL3.

Here, the memory M stores evaluation terms 9, knowledge 11 that defines relationships between evaluation terms, vehicle design elements 13, and knowledge 15 that defines relationships between evaluation terms and design elements.

The evaluation term 9 is composed of adjective words as shown in the evaluation term column of Table 1 in FIG.

Knowledge 11 that defines the relationship between evaluation terms is knowledge that was obtained by collecting evaluation terms 9 as shown in Table 1 and analyzing the results of a preliminary experiment to define the relationship between evaluation terms. This relationship means, for example, that evaluation term 9 is defined by quantification theory I.
Data is obtained by dividing factors into factors 1 to 10 based on human senses through multivariate analysis such as ``class'' or ``class'', and the degree of relationship between each evaluation term and each factor is obtained as a factor loading (numerical value). It can be said that terms with close numerical loadings on each factor are closely related and similar. Knowledge 11 that defines the relationship between evaluation terms selects the most similar term when the input evaluation term is not in the evaluation term group used in the sensory evaluation experiment described later when a design device is required. It is used to create a factor axis, which will be described later.

The vehicle design element group 13 is divided into categories for each item, such as the number of meters in the vehicle interior, as shown in Table 2 of FIG.

Knowledge that defines the relationship between evaluation terms and vehicle design elements 1
No. 5 was obtained as a result of the sensory evaluation experiment as shown in Table 2 above. Specifically, items of evaluation terms 9 and vehicle design elements 11 are selected, vehicle design elements corresponding to each item are shown to an unspecified number of people, and each evaluation term 9, which is a group of adjective words, is evaluated. Feelings received from vehicle design elements are collected, and the accumulated results are analyzed as partial regression coefficients (correlation coefficients) using multivariate angle analysis such as quantification theory type I or type II.

As shown in Table 2, the items include the number of meters installed on the instrument panel, the instrument panel and meter cluster, the meter cluster and center console, the thickness of the door, the overhang of the center cluster, and the armrest. This study focused on each of the following.

The number of meters is classified into five categories. Namely, the first category has one large meter, the second category has one large meter and two small meters, the third category has one large meter and three small meters, and the third category has two large meters. and one small meter, and a fifth category has two large meters and two small meters.

Items focused on instrument panels and meter clusters are divided into two categories. That is, there are a first category in which the instrument panel and the meter cluster are integrated, and a second category in which the instrument panel and the meter cluster are separated.

The third item, which focuses on the meter cluster and center console, is divided into two categories. In other words, the first model that combines the meter cluster and center console.
category, and a second category in which the meter cluster and center console are separated.

Items that focus on the thickness of the door are divided into two categories. That is, the first category has thick doors, and the second category has thin doors.

Items that focus on the overhang of the center cluster are classified into two categories. That is, the first category has a center cluster overhang, and the second category has no center cluster overhang.

Below, items focusing on armrests and other items are classified into categories.

The keyboard 3 constitutes an input means CL2 for evaluation terms, and also constitutes a weighting means CL5 for weighting the evaluation terms when there are a plurality of input evaluation terms. For example, if you want to obtain a design with a subtle expression such as "I can concentrate somewhat and it is quite bright", use the keyboard 3 to input the evaluation terms "I can concentrate" and "It is bright", and the former is 20%, the latter is 80%, etc. This is what you enter.

The CPU 7 infers the design from the knowledge database based on the input evaluation terms and the flowchart shown in FIG.

The CRT 5 constitutes an output means CL4 for outputting the inferred design and displays the design as an image.

According to the above configuration, the evaluation terms 9 selected in advance, the knowledge 11 that defines the relationship between the evaluation terms conducted in the preliminary experiment, the vehicle design elements 13, and the relationship between the evaluation terms obtained in the sensitivity experiment and the vehicle design elements are Prescribed knowledge 15 and the like are stored in the memory M of the computer 1. In this state, when the keyboard 3 is operated to input an adjective as an evaluation term, each vehicle design element 13 is inferred and extracted from the knowledge database by the action of the CPU 7 of the computer 1.

Here, a description will be given based on the flowchart of FIG. 5 assuming that the evaluation term 9 is "somewhat able to concentrate and quite bright".

In step S1, it is assumed that "I can concentrate a little and it's quite bright" is input, which represents the feeling that humans have. In this case, the adjective as evaluation term 9 is “
Select two options: "I can concentrate" and "Bright" and input them using keyboard 3. Then, input the ratio between each evaluation term 9 as a percentage so that the total becomes 100%.

In other words, the feeling of "somewhat able to concentrate" is, for example, 2.
0% is determined, and 80% is determined as a feeling of "quite bright", and each is input using the keyboard 3. Using the plurality of evaluation terms 9 weighted and applied in this way, design elements for the vehicle interior are inferred from the knowledge database in step S2, and in step S3,
The design elements extracted by inference are output and displayed as a design image on the CRT 5.

Figure 6 shows the inference process in step S2.
a) shows part of the knowledge database showing the relationship between evaluation terms, items, and categories shown in Figure 4. Based on the input of evaluation terms in step S1, evaluation terms are circled in Table 3. As shown in the box, "I can concentrate" is the evaluation term most closely related to the input evaluation term 9.
"Bright" is selected. Partial regression coefficients corresponding to the selected evaluation terms are then extracted as circled in Table 3.

The partial regression coefficient of each evaluation term extracted in this way is multiplied by the weight given as a percentage in step Sl, and the result is obtained as shown in Table 4 of FIG. 6(b).

Table 5 in Figure 6(C) is the sum of the partial regression coefficients between these evaluation terms for each category of each item.
The evaluation terms and partial regression coefficients shown in Table 5 indicate the relationship between the input term information and the design element.

Then, from among the partial regression coefficients obtained as shown in Table 5, the category showing the maximum value for each item is extracted as indicated by an * mark, and the image is displayed on the CRT as the vehicle interior design as shown above. It is something. Therefore, subtle expressions such as ``somewhat concentrated and quite bright'' can be entered as evaluation terms, and a design that is closer to the image that humans have can be extracted and displayed. For this reason,
It is possible to input various images that the user has at a dealer etc. and display a design according to the input.
It is possible to select a car design closer to the user's image.

In addition, if a single term is input as an evaluation term, the item corresponding to the largest partial regression coefficient marked with an asterisk (*) among the partial regression coefficients of each item category corresponding to the evaluation term from Figure 4 and Table 2. , categories are inferred and extracted, and images are displayed on the CRT 5 as described above.

FIG. 7 shows a flowchart according to another embodiment.

In this embodiment, the method of inputting the evaluation term 9 is different. That is, the result of the factor axis obtained by the factor analysis of the evaluation term 9 is displayed as an image on the CRT 5, and the weighted evaluation term is input by plotting it on the image.

First, in step Sll, a factor axis related to the image of the design desired by the user is selected. In this factor axis selection, as shown in the upper column of step Sll, the term representing the factor axis is CRT.
5, and select as many numbers as you like using the keyboard 3 or the like while looking at the displayed screen. For example, step S11 shows a case where the factor axes of "beauty", "dynamism", and "luxury" are selected.

Once the factor axes have been selected, an image is displayed on the CRT 5 with each factor axis serving as a coordinate axis, as shown in the lower column of step Sll. In this case, the standard evaluation term (black circle) and car name A are used to make it easier for users to plot on the screen.
, BSC, D, etc. are displayed at the same time. Therefore, the user inputs these terms or the image of the actual vehicle by plotting them on the coordinates as shown in double force, and the weighting of each axis of the input point is calculated from the coordinate value of each axis. It is something that

To explain this with reference to FIG. 8, when plotted as shown by the double force in FIG. 8, the weighting is such that the aesthetic factor is 0.
5. The dynamic factor is 0.3. Once each factor axis representing the evaluation term 9 is weighted in this manner, inference and extraction of each design element is performed in step S12 in the same manner as in the case of FIG.

This inference process is shown in FIG. FIG. 9 shows the relationships between factor axes, items, and categories instead of the relationships between evaluation terms, items, and categories.

Table 6 in Figure 9(a) shows the results obtained using Quantification Theory Type 1 (described later) for the relationship between the factor scores of each factor axis obtained through factor analysis, items, and categories (described later). This figure shows the correspondence between items and categories for the three factor axes input. This Table 6 is stored in the memory M as a knowledge database.

Then, the weighting of the three factor axes obtained as above, that is, the aesthetic factor is 0.5, the dynamic factor is 0.3, and the luxury factor is 0, is multiplied by the partial regression coefficient of each axis. figure(
b) is obtained as Table 7.

Next, the partial regression coefficients of each axis are calculated as the sum of each category of each item, and are obtained as shown in Table 8 of FIG. 9(C). Table 8 shows the relationship between the user's image plotted on the layout as described above and each item and category. Infer and extract the categories with 1. The result is displayed as an image on the CRT 5 as the vehicle interior design by the output of step S3.

Therefore, even if the user cannot think of specific evaluation terms according to the image, it is possible to directly input the image on the factor axes that represent the design of the vehicle interior, making it easier to display the design. can.

Next, the creation of the factor axis used in step 5.11 above will be described.

FIG. 10 shows a control flowchart. First, in step 521, a partial regression coefficient matrix X=(x+k) (Table 2) indicating the relationship between evaluation terms and each item and category in the vehicle design elements is created regarding the knowledge database. Read as data. Note that X-tx1hl indicates the partial regression coefficient of the j-th item and category in the i-th adjective.

Next, in step S22, the factor loading matrix Y-(Fzl (Table 1)) between the evaluation term and the factor axis is read as data.
represents the factor loading of the th factor axis.

Then, in step 823, the partial regression coefficients of items and categories for all evaluation terms are weighted with the factor loadings corresponding to each factor axis, and an inference formula for calculating the sum is calculated by matrix multiplication of (Y7) x. A partial regression coefficient matrix 2'' (Z+*l) that can quantify the axes, items, and categories can be obtained as shown in Table 9 in Figure 11. It shows partial regression coefficients for items and categories first.

In step S11, a factor axis is formed based on the correspondence between the factor axis created in this way and each item and category, and is displayed as an image. Table 6 in FIG. 9(a) is a partial extraction of Table 9.

FIG. 12 shows a flowchart according to another embodiment of factor axis creation. In this flowchart, steps S31 to 836 are added to the flowchart in FIG.
This is done by narrowing down the terms that represent each factor axis for the factor loading matrix Y.

That is, the i-1th adjective is selected in step S31, and the maximum factor loading Y is selected in step S32.
, the factor loading amount of the j-th factor and axis that is the max of , is determined.

Next, in step 533, all factors other than the factor loadings of the selected j-th factor axis are set to zero. For example, Table 1 in Figure 3
On the other hand, in Table 10 of Figure 13, the factor loading of the first factor axis for the first adjective is 0.99, which is the maximum.
All other factor loadings are replaced with 0. Then, in step S34, the value of Yl is written into Y', and in step S35 it is determined whether all evaluation terms have been processed. If processing has not been completed for all evaluation terms, i+1 is performed in step S36 to process the next evaluation term, and the processes from steps S32 to 835 are repeated.

Once the above processing is completed for all evaluation terms, a new factor loading matrix Y as shown in Table 10 of Figure 13 will be created.
' is created and stored in memory M.

Then, in step 323, weighting calculations similar to those described above are performed, and a partial regression coefficient matrix 2-(Z+h) that can quantify each factor axis, item, and category can be obtained. As a result, a table similar to FIG. 11 is obtained, although the numerical values are different.

FIG. 14 shows a flowchart relating to the creation of a correspondence table used when samples such as actual vehicles are simultaneously displayed on the coordinates of the factor axis as in step S11 of FIG. 7.

First, in step S41, the factor score matrix "r-(TH)" regarding the factor scores for each factor axis is read. Note that ij indicates the factor score for the j-th prisoner of the i-th sample. That is, as shown in FIG. In Table 12 of b), a factor score of 0.95 is read for factor 1 of the first sample 1.This table 12 shows the evaluation terms and factor loadings of the factor axes as well as the actual vehicle samples through factor analysis. This is obtained as the factor score of the factor axis.Next, step S43
A correspondence table between evaluation samples and design elements is read in. This correspondence table is shown in Table 11 of FIG. 15(a), and is a correspondence table showing what kind of items and categories a sample such as an actual vehicle is made up of. Then, in step S44, a multivariate analysis method such as quantification theory type 1 is applied to these read data to quantify each factor axis and the item category of each sample.

The results are shown in Table 13 shown in FIG. 15(e).

That is, for actual vehicle sample 1, the score related to the first factor, that is, the aesthetic factor in this example, is 0°16. In this way, scores are determined for all factors for all samples. Through this inference process, it is possible to quantify the factor axes that represent the design evaluation image and the sample items and categories.
It is possible to display car A, car B, car C, car D, etc. in step Sll in the figure. Note that the weighting can be determined from the beginning, or the weighting can be done in the selected order. As input data, engine displacement, vehicle price, age, etc. can also be added. This invention can be applied not only to vehicle interior design but also to exterior and other designs, and can also be applied to housing and other designs.

[Effects of the Invention] As is clear from the above, it is possible to weight evaluation terms regardless of the structure of the present invention, it is possible to input a subtle image of a design, and a design that more closely corresponds to the image can be output. be able to.

[Brief explanation of drawings]

Fig. 1 is a block diagram of the present invention, Fig. 2 is a block diagram according to an embodiment, Fig. 3 is a table 1 regarding knowledge that defines the relationship between evaluation terms, and Fig. 4 is a table of sensory evaluation experiment results. 2. FIG. 5 is a flowchart according to one embodiment, FIG. 6 is a diagram showing the inference process from family 3 to Table 5, FIG. 7 is a flowchart according to another embodiment, and FIG. 8 is a coordinate diagram of the factor axis. Figure 9 shows the inference process from family 6 to Table 8, Figure 10 is a flowchart for creating a correspondence table for creating factor axes, Figure 11 shows each factor axis and item,
Table 9 of knowledge that defines relationships with categories; Figure 12 is a flowchart of another example for obtaining a correspondence table related to factor axis creation; Figure 13 is Table 10 of knowledge that defines relationships between evaluation terms.
, Fig. 14 is a flowchart related to creating a correspondence table when plotting samples on the factor axis, Fig. 15 shows Tables 11 to 13 related to the inference process, and Fig. 16 shows a block diagram of a conventional example. . CLI...Knowledge database Cl3...Input means Cl3...Calculation means Cl4...Output means Cl3...Weighting means Cl6...Correction means Figure 5 Dynamic factor Figure 8 Figure 10 Figure 14 Drawing procedure amendment (method) Indication of the case 2゜Name of the invention 3゜Person making the amendment Address related to the case (residence) Name 4゜Representative flat-bottomed year 1999 Patent Application No. 341129 design equipment

Claims (2)

[Claims] (1) A knowledge database of evaluation terms expressing the design and the relationship between design elements, an input means for inputting the evaluation terms, an arithmetic means for inferring the design from the knowledge database using the input evaluation terms, and a calculation means for inferring the design from the knowledge database using the input evaluation terms; an output means for outputting the design; a weighting means for weighting the input evaluation terms when there is a plurality of them; and a correction means for correcting the inference by the calculation means using the weighting. design equipment. (2) A knowledge database of relationships in which factor loadings are determined for each of multiple sensory factors between slogan terms expressing design, and relationships in which partial regression coefficients are determined between evaluation terms and design elements, and A design device comprising an inference unit that weights factor loadings and associates each factor with a design element by the summation thereof.