JP2022520911A - How to determine product design that combines brain waves and eye movements with user similarity - Google Patents

Abstract

The present invention discloses a method for determining a product design that combines brain waves and eye movements with user similarity, and includes three parallel steps. (1) Collecting original EEG signals, removing artifacts and noise, β band and valence α band related to emotional arousal degree in EEG, and event-related potential ERPs potential when the determiner selects a solution. Extract changes. (2) Collect the degree of attention acquired based on the eye movement data. (3) Obtain the time for the decision-maker to select the solution, and acquire the degree of preference for the solution. All three of these data are combined to obtain implicit decision-making data, explicit decision-making data is obtained by the choice of the decision-maker, and the degree of relevance between the decision-making action data of different decision makers is calculated for user similarity. To get. This method collects explicit and implicit decision-making actions, establishes a decision-maker-solution relationship, calculates the nomination order, and forms a suitable set of solutions based on user similarity relationships. do. [Selection diagram] Fig. 1

Description

The present invention belongs to the technical field of a product design determination method, and specifically relates to a product design determination method that combines brain waves and eye movements with user similarity.

In an environment where market competition is fierce and product lifecycles are shortened, the design and development process should be simplified, scientific, regular and objective so that companies seek to gain a competitive advantage. , It is necessary to quickly find the optimum solution of the design solution through effective design decision means. During the product design decision process, complexity and uncertainty are one of the factors for long decision time and low efficiency. Design decisions are a multi-objective decision problem for multi-attribute groups, multi-attribute emphasizes opinion gathering and consensus, and multi-objective emphasizes the need for multidimensional elements to optimize solutions. .. In the background of the increasingly apparent problem of information overload, recommended systems are widely used as an important tool of information filtering technology, finding relevant domains of interest between different users with large amounts of data, Building a function of similarity between users and using this correlation to make real-time recommendations, this method solves product design decision problems based on a small amount of sample data and is the decision-maker's decision-maker. It can also solve multi-attribute problems. During the product development process, the decision-making action can be analyzed and evaluated for the goodness of the product design by using the brain wave band frequency and the gaze feedback of the human eye, based on the judgment of the whole solution to some extent, and the human cognitive evaluability and perception. Discover eye movement data when a decision-maker makes a decision based on wholeness, and acquire brain wave data based on various psychological states and emotional changes for various solutions during the process of human decision making. , Find out the collaborative choices of the decision makers by combining the above-mentioned user similarity, and solve the problem of multipurpose decision in the group-oriented decision process.

To overcome the shortage of prior art, the present invention provides a method for determining product design that combines brain waves and eye movements with user similarity, which utilizes user similarity and physiological measurement data to "" The purpose is to acquire "explicit decision-making action" and "implicit decision-making action", and to provide a new product design decision-making method by combining both.

In order to achieve the above object, the present invention uses the following technical solutions.

A product design decision method that combines brain waves and eye movements with user similarity. Through technologies such as similarity calculation, data processing, and correlation analysis, "explicit decision action" and "implicit decision action" The two action elements are extracted, and the actions of the two attributes have different weighting factors, and they are calculated to obtain the "determiner-solution" matrix. The method comprises the following steps.

(1) Enter multiple types of design solutions that meet the target product, establish a design sample library for the target product, and the system displays the solutions that cover the entire sample in a group in the form of photographs.

(2) The deciding person selects the solution for each group displayed by clicking the photo, the system records one selection time and the total selection time, and depending on the speed at which the deciding person selects the solution. The degree of preference P for the decision-maker's solution is shown.

(3) The decision-maker wears an eye tracker and an electroencephalograph at the same time during the selection process, and records eye movement data and electroencephalograph data.

(4) Based on the explicit decision-making action obtained by the decision-maker by selecting the solution, at the same time, the decision-maker obtains the implicit decision-making action based on the eye movement data, the brain wave data, and the speed of the solution-selection. -Establish an action matrix between solutions.

(5) The user similarity between the determiners is calculated based on the action matrix, the order of superiority of the solution means is obtained, and the solution means decision matrix D = user similarity × action matrix.

In the target product design sample library in step (1), the total number of samples in the sample library S is n, and Sn = {s ₁ , s ₂ , ..., s _j }, s _j indicates the first solution. During the solution determination process, we recommend a solution that covers the entire sample, the solution is randomly displayed in groups on the system screen in the form of a photo, and the number of sample samples of the solution shown on the screen each time is r. r is generally 3-5, and p / r is only an integer.

In the step (2), the total number of decision makers U is q, and Uq = {u ₁ , u ₂ , ..., _ui }, _ui indicates the first decision maker. In the solution selection of each group, the deciding person sets the selected solution to 1 and the unselected solution to 0, and establishes an explicit determination matrix Ex ( _ui , s _j ) of the person and the solution. do. Here, the total number of solutions to be set to 1 corresponding to each determiner is n / r.

In the step (2), the time required for the predetermined decision-maker u _i to select the solution j is set to t _cij , that is, the selection time of the single group, and the total selection time is set to T _ci , and the decision-maker The preference function P ( _ui , s _j ) for the determinant's solution is shown at the speed of selecting the solution.

P _ij = 1 / (t _cij / T _ci ).

To acquire the eye movement data of the determiner in the step (3), it is necessary to collect two types of eye movement actions, eye movement and gaze, and the method of recording the eye movement data is a single group. The degree of attention to the determinant's solution is measured by the ratio of the gaze time of the single solution to the total gaze time of the solution, and the gaze frequency is the same. The single plan gaze time t _ij , the single plan total gaze time T _eij , the single plan gaze frequency f _eij , and the single plan total gaze frequency F _eij . The method calculates only the gaze time of the single solution in one group of solutions and the solution with the highest gaze frequency of the single solution, i.e., the gaze time of the maximum single solution and the gaze frequency of the maximum single solution. Is extracted, so the number of record solutions is n / r. Establish eye movement determination matrix E ( _ui , s _j ) for humans and solutions.

E _ij = (t _ej / T _eij ) + (f _eij / F _eij )

In step (3), to acquire the EEG data of the determiner, it is necessary to perform artifact removal and digital filtering processing on the collected original EEG signal, and the method of recording the EEG data is as follows. .. Collect original EEG signals to remove artifacts containing ocular and myoelectric potentials, use digital filtering to remove noise, and extract β-bands for EEG emotional arousal and α-bands for valence And, the potential change of the event-related potential ERPs at the time of selection by the determiner is extracted. dt is a differentiation in which the independent variable is set to time t, and f _α and f _β are frequency values related to the emotion valence α band and the emotion arousal β band of EEG, respectively, and the EEG determination matrix B (human and solution) ( _ui , s _j ) is established.

The explicit decision action in step (4) is shown in the explicit decision matrix Ex ( _ui , s _j ) and the implicit decision action Im ( _ui , s _j ).

Im ( _ui , s _j ) = (E, B, P)

Acquire the action matrix C.

前記ステップ（５）において、ユーザ類似性計算方法は以下のとおりであり、２人の決定者ｕ_１とｕ_２は、行為マトリックスＣによってユーザ類似性Ｓｉｍを計算し、Ｃ（ｕ_１）で決定者ｕ_１がポジティブフィードバックを備えた前の１０個の解決手段集合を示し、Ｃ（ｕ_２）は決定者ｕ_２がポジティブフィードバックを備えた前の１０個の解決手段集合であり、ユーザ類似性要素はＳｉｍであり、

ｕ_１とｕ_２を二つずつ比較することにより、ユーザ類似性ＳｉｍのマトリックスＤ（Ｄ＝Ｓｉｍ×Ｃ）を取得する。

In the step (5), the user similarity calculation method is as follows, and the two determiners u ₁ and u ₂ calculate the user similarity Sim by the action matrix C and determine by C (u ₁ ). Person u ₁ shows the previous set of 10 solutions with positive feedback, and C (u ₂ ) is the previous set of 10 solutions with determinant u ₂ with positive feedback, user similarity. The element is Sim,

By comparing u ₁ and u ₂ two by two, the matrix D (D = Sim × C) of the user similarity Sim is obtained.

The present invention has the following features.

(1) The design decision method based on user similarity simplifies a complicated product analysis process by utilizing the result of solution screening, and provides an innovation method and decision support for the design decision of a company or a group.

(2) In the present invention, the implicit choice obtained by combining the display choice of the determiner with the physiological measurement data improves the accuracy of the decision based on cognition.

(3) The present invention can reduce the subjective influence of the determiner and accurately expand the scope of design by combining quantitative and qualitative determination modes.

It is a general flowchart of a product decision. It is a detailed flowchart of the method. This is the home page of the product decision platform. This is the start screen for product decision. It is a decision data display screen. It is a result output screen of the product decision.

The present invention discloses a method for determining a product design that combines brain waves and eye movements with user similarity, and the reliability of the evaluation result of the solution is low due to human cognitive ambiguity and subjective influence during the design determination process. To solve the problem. The new method of design decision provided by the present invention is applied to the evaluation and testing stage of a design proposal, and further screening process is performed by utilizing a set of solutions in a design sample library. From the screening principle, this method applies to the design of consumable products. The method involves three parallel steps, the first step, collecting the original EEG signal, removing artifacts such as electrooculogram and myoelectric potential, removing noise using digital filtering, and β regarding the degree of valence in EEG. The band and the valence α band are extracted, and the potential change of the event-related potential ERPs when the determiner selects the solution is extracted. The second step, eye movement, gaze and gaze, three types of eye movement behavior data are collected. In the third step, the time for the decision-maker to select the solution is acquired and the degree of preference is acquired. The data of these three steps are combined to obtain implicit decision-making data, explicit decision-making data is obtained by the choice of the decision-maker, and the degree of relevance between the decision-making action data of different decision makers is calculated by the user. Get similarities. This method collects explicit and implicit decision-making actions, establishes a decision-maker-solution relationship, calculates the nomination order of solutions based on this correlation, and is based on user similarity relationships. Form a suitable set of solutions.

Next, the present invention will be further described with reference to the drawings and examples.

A product design decision method that combines brain waves and eye movements with user similarity. Through technologies such as similarity calculation, data processing, and correlation analysis, "explicit decision action" and "implicit decision action" The two action elements are extracted, and the actions of the two attributes have different weighting factors, and they are calculated to obtain the "determiner-solution" matrix. The method comprises the following steps.

1. 1. After logging in, the operator enters the homepage of the system for determining the product design plan, clicks "New experiment" to add the name of the experiment, and adds a design plan that meets the target product, that is, a design sample of the target product. After establishing the library, the system displays the solution covering the entire sample to the group in the form of a photograph, and then the number of single-group samples displayed on the screen, the minimum experiment time for the single group, the number of decision makers of the experiment participants, etc. Experimental parameters are set and shown in FIG.

2. 2. Under the guidance of the operator, each decision-maker reads "Notes" on the right side of the system homepage, and conducts a preliminary experiment to select a satisfactory solution as shown in FIG. After that, the decision-maker wears an eye tracker and an electroencephalograph at the same time with the cooperation of the operator, and records the eye movement data and the electroencephalograph data. The decision-maker enters basic personal information and clicks "Start Experiment" as shown in FIG.

3. 3. When the timekeeping of the experiment begins, the determiner clicks on the photo to select the solution that best matches the decision goal display and displays the solution in the group if the minimum experiment time limit is required. The system records one selection time and the total selection time, and indicates the degree of preference P for the solution by the decision-maker depending on the speed at which the decision-maker selects the solution. After the last group display is finished, the experiment will automatically end the timing.

4. As shown in FIG. 5, the record of the experimental data is read, and the eye movement data and the electroencephalogram data acquired by the experiment have a one-to-one correspondence with each other depending on the time. Based on the explicit decision-making action acquired by the decision-maker by selecting the solution, at the same time, the implicit decision-making action is acquired by the eye movement data, the brain wave data, and the speed of solution selection, and the decision-maker-solution means. Establish an action matrix between.

5. The user similarity between the decision makers is calculated based on the action matrix, the ranking of the superiority of the solution is obtained, and the solution decision matrix D = user similarity × action matrix. The computational process takes place in the background of the system, where Figure 6 is the group of optimal solutions for decisions displayed in the form of photographs, and the product displayed in the leftmost photo is the solution for the optimal decision for the experiment. It is a means.

In the target product design sample library in step (1), the total number of samples in the sample library S is n, and Sn = {s ₁ , s ₂ , ..., s _j }, s _j indicates the first solution. During the solution determination process, recommendations are made in the form of solutions that cover the entire sample, the solutions are randomly grouped on the system screen in the form of photographs, and the number of sample solutions shown on the screen each time is r. In general, r is 3 to 5, and p / r can only be an integer.

In the step (2), the total number of decision makers U is q, and Uq = {u ₁ , u ₂ , ..., _ui }, _ui indicates the first decision maker. In the solution selection of each group, the deciding person sets the selected solution to 1 and the unselected solution to 0, and establishes an explicit decision matrix Ex ( _ui , s _j ) for humans and solutions. do. Here, the total number of solutions to be set to 1 corresponding to each determiner is n / r.

In the step (2), the time required for the determinant u _i to select the solution j is set to t _cij , that is, the selection time of the single group, the total selection time is set to T _ci , and the decision maker solves the problem. The preference function P ( _ui , s _j ) for the determinant's solution is shown at the speed of selecting the means.

P _ij = 1 / (t _cij / T _ci ).

To acquire the eye movement data of the determiner in the step (3), it is necessary to collect two types of eye movement actions, eye movement and gaze, and the method of recording the eye movement data is a single group. The degree of attention to the determinant's solution is measured by the ratio of the gaze time of the single solution to the total gaze time of the solution, and the gaze frequency is the same. The single plan gaze time t _ij , the single plan total gaze time T _eij , the single plan gaze frequency f _eij , and the single plan total gaze frequency F _eij . The method calculates only the gaze time of one group of solutions and the solution with the highest gaze frequency of the single solution, i.e., the gaze time of the maximum single solution and the gaze frequency of the maximum single solution. Therefore, the number of recording solutions is n / r. Establish eye movement determination matrix E ( _ui , s _j ) for humans and solutions.

E _ij = (t _ej / T _eij ) + (f _eij / F _eij )

In step (3), to acquire the EEG data of the determiner, it is necessary to perform artifact removal and digital filtering processing on the collected original EEG signal, and the method of recording the EEG data is as follows. .. Collect original EEG signals to remove artifacts containing ocular and myoelectric potentials, use digital filtering to remove noise, and extract β-bands for EEG emotional arousal and α-bands for valence And, the potential change of the event-related potential ERPs at the time of selection by the determiner is extracted. dt is a differentiation in which the independent variable is set to time t, and f _α and f _β are frequency values related to the emotion valence α band and the emotion arousal β band of EEG, respectively, and the EEG determination matrix B (human and solution) ( _ui , s _j ) is established.

The explicit decision action in step (4) is shown in the explicit decision matrix Ex ( _ui , s _j ) and the implicit decision action Im ( _ui , s _j ).

Im ( _ui , s _j ) = (E, B, P)

Acquire the action matrix C.

In the step (5), the user similarity calculation method is as follows, and the two determiners u ₁ and u ₂ calculate the user similarity Sim by the action matrix C and determine by C (u ₁ ). Person u ₁ shows the previous set of 10 solutions with positive feedback, and C (u ₂ ) is the previous set of 10 solutions with determinant u ₂ with positive feedback, user similarity. The element is Sim.

By comparing two by two, the matrix D of the user similarity Sim is obtained.

D = Sim × C

The above is merely a preferred embodiment of the present invention, and it should be noted that those skilled in the art can make a plurality of improvements and modifications without departing from the principle of the present invention. Modifications should also be considered as the scope of protection of the present invention.

(Additional note)
(Appendix 1)
It is a product design decision method that combines brain waves and eye movements with user similarity, and the subjective choice of the decision-maker is used as explicit decision data, and the decision-maker's eye movement data, brain wave data, and degree of preference are implicit. Determining data, calculating user similarity with different decision levels, achieving final product decisions,
(1) Steps to establish a sample library of target product design means,
(2) The deciding person selects a solution sample and obtains an explicit decision matrix Ex by the explicit decision act.
(3) The system selects the selection time and total selection time of the single group sample, shows the preference function P depending on the speed, and the decider wears the eye tracker and the electroencephalograph at the same time during the selection process, and the eye movement data. And the step of recording the brain wave data, acquiring the eye movement determination matrix E and the brain wave determination matrix B, and acquiring the implicit action matrix Im = (E, B, P).
(4) The relationship between the determiner and the solution sample, that is, the step of establishing the matrix C = (Ex, Ex, Ex) + Im, and
(5) A step of calculating a matrix of user similarity Sims between decision makers by the action matrix C, and
(6) The final solution determination matrix D is a method comprising a step of user similarity and joint result with the action matrix: D = Sim × C.

(Appendix 2)
It is a method for determining a product design that combines brain waves and eye movements described in Appendix 1 with user similarity.
In step (2), in the solution selection of each group, the decision-maker sets the selected solution to 1 and the unselected solution to 0, and sets the explicit determination matrix Ex ( _ui ,) of the person and the solution. A method characterized in that s _j ) is established, where ui indicates a determinant of the _i and s _j indicates a solution of the j.

(Appendix 3)
It is a method for determining a product design that combines brain waves and eye movements described in Appendix 1 with user similarity.
In step (3), the eye movement determination matrix _Eij is _{Eij = (teij /} T _ej ) + ( _feij / F _ij ).
A method characterized by having a single plan gaze time t _ej , a single plan total gaze time T _eij , a single plan gaze frequency f _eij , and a single plan total gaze frequency F _eij .

であり、
決定者ｕ_ｉが解決手段ｓ_ｊを選択する時間はｔ_ｃｉｊであり、即ち、シングルグループ選択時間であり、ｄｔは独立変数が時間ｔにされる微分であり、ｆ_α、ｆ_βはそれぞれＥＥＧの感情価数αバンドと感情喚起度βバンドに関する周波数値であることを特徴とする方法。 (Appendix 4)
It is a method for determining a product design that combines brain waves and eye movements described in Appendix 1 with user similarity.
In step (3), the EEG determination matrix _Bij

And
The time for the determiner u _i to select the solution s _j is t _cij , that is, the single group selection time, dt is the differentiation of the independent variable to time t, and f _α and f _β are EEG, respectively. A method characterized in that it is a frequency value relating to an emotional valence α band and an emotional arousal β band.

(Appendix 5)
It is a method for determining a product design that combines brain waves and eye movements described in Appendix 1 with user similarity.
In step (3), the function P ( _ui , s _j ) of the degree of preference for the solution s _j of the determinant _ui is shown at the speed at which the determinant selects the solution.
P _ij = 1 / (t _cij / T _ci ),
Here, t _cij is a time for selecting the solution s _j selected by the determiner u _i , that is, a single group selection time, and T _c i is a total selection time.

ことを特徴とする方法。 (Appendix 6)
It is a method for determining a product design that combines brain waves and eye movements described in Appendix 1 with user similarity.
In the step (5), the user similarity calculation method is as follows, and the two determiners u ₁ and u ₂ calculate the user similarity Sim by the action matrix C and determine by C (u ₁ ). Person u ₁ shows the previous set of 10 solutions with positive feedback, and C (u ₂ ) is the previous set of 10 solutions with determinant u ₂ with positive feedback, user similarity. The element is Sim,

A method characterized by that.

Claims (6)

It is a product design decision method that combines brain waves and eye movements with user similarity, and the subjective choice of the decision-maker is used as explicit decision data, and the decision-maker's eye movement data, brain wave data, and degree of preference are implicit. Determining data, calculating user similarity with different decision levels, achieving final product decisions,
(1) Steps to establish a sample library of target product design means,
(2) The deciding person selects a solution sample and obtains an explicit decision matrix Ex by the explicit decision act.
(3) The system selects the selection time and total selection time of the single group sample, shows the preference function P depending on the speed, and the decider wears the eye tracker and the electroencephalograph at the same time during the selection process, and the eye movement data. And the step of recording the brain wave data, acquiring the eye movement determination matrix E and the brain wave determination matrix B, and acquiring the implicit action matrix Im = (E, B, P).
(4) The relationship between the determiner and the solution sample, that is, the step of establishing the matrix C = (Ex, Ex, Ex) + Im, and
(5) A step of calculating a matrix of user similarity Sims between decision makers by the action matrix C, and
(6) The final solution determination matrix D is a method comprising a step of user similarity and joint result with the action matrix: D = Sim × C. The method for determining a product design in which the electroencephalogram and eye movements according to claim 1 are combined with user similarity.
In step (2), in the solution selection of each group, the decision-maker sets the selected solution to 1 and the unselected solution to 0, and sets the explicit determination matrix Ex ( _ui ,) of the person and the solution. A method characterized in that s _j ) is established, where ui indicates a determinant of the _i and s _j indicates a solution of the j. The method for determining a product design in which the electroencephalogram and eye movements according to claim 1 are combined with user similarity.
In step (3), the eye movement determination matrix _Eij is _{Eij = (teij /} T _ej ) + ( _feij / F _ij ).
A method characterized by having a single plan gaze time t _ej , a single plan total gaze time T _eij , a single plan gaze frequency f _eij , and a single plan total gaze frequency F _eij . The method for determining a product design in which the electroencephalogram and eye movements according to claim 1 are combined with user similarity.
In step (3), the EEG determination matrix _Bij

And
The time for the determiner u _i to select the solution s _j is t _cij , that is, the single group selection time, dt is the differentiation of the independent variable to time t, and f _α and f _β are EEG, respectively. A method characterized in that it is a frequency value relating to an emotional valence α band and an emotional arousal β band. The method for determining a product design in which the electroencephalogram and eye movements according to claim 1 are combined with user similarity.
In step (3), the function P ( _ui , s _j ) of the degree of preference for the solution s _j of the determinant _ui is shown at the speed at which the determinant selects the solution.
P _ij = 1 / (t _cij / T _ci ),
Here, t _cij is a time for selecting the solution s _j selected by the determiner u _i , that is, a single group selection time, and T _c i is a total selection time. The method for determining a product design in which the electroencephalogram and eye movements according to claim 1 are combined with user similarity.
In the step (5), the user similarity calculation method is as follows, and the two determiners u ₁ and u ₂ calculate the user similarity Sim by the action matrix C and determine by C (u ₁ ). Person u ₁ shows the previous set of 10 solutions with positive feedback, and C (u ₂ ) is the previous set of 10 solutions with determinant u ₂ with positive feedback, user similarity. The element is Sim,

A method characterized by that.

Images