State prediction system, member determination system, and state prediction method

Abstract

The present disclosure relates to a state prediction system, a member determination system, and a state prediction method. The state prediction system includes a controller. The controller is configured to acquire a current state value indicating a current state of a target individual in a group. The controller is configured to acquire a state propagation amount indicating an amount of a state propagated to the target individual from other people in the group through communication between the target individual and the other people. The controller is configured to predict a future state value indicating a future state of the target individual, in accordance with the acquired current state value of the target individual and the acquired state propagation amount.

Description

Technical Field

[0001] This invention relates to a state prediction system, a member determination system, and a state prediction method. Background Technology

[0002] Japanese Unexamined Patent Application Publication No. 2011-186521 (JP 2011-186521 A) discloses an emotion assessment device that assesses group emotions as emotions in interpersonal communication settings. In this technology, the group emotions are assessed based on information about the current emotions of individuals and a state transition model. Summary of the Invention

[0003] In the technology described in JP 2011-186521 A, it is possible to assess that the current group mood is deteriorating due to the deterioration of an individual's current mood, but no measures can be taken before the group mood deteriorates. Even if measures are taken, it is difficult to restore the atmosphere of the place from a state of deteriorating group mood. The inventors have recognized that predicting the future state of individuals is desirable in order to smoothly promote communication within a group.

[0004] This invention provides a technique for predicting the future state of individuals within a group.

[0005] A first aspect of the present invention relates to a state prediction system, which includes a controller. The controller is configured to acquire a current state value indicating the current state of a target individual in a group, acquire a state propagation amount indicating the amount of state propagated from other individuals to the target individual through communication between the target individual and other individuals in the group, and predict a future state value indicating the future state of the target individual based on the acquired current state value of the target individual and the acquired state propagation amount.

[0006] In the scheme, the controller can be configured to acquire a current state value indicating the current state of the other person, the current amount of communication between the target individual and the other person, and the degree of influence of the other person on the state of the target individual as information, and derive the state propagation amount from the other person to the target individual based on the acquired information.

[0007] In the scheme, the controller can be configured to derive the product of the current state value of the other personnel, the current amount of communication between the target individual and the other personnel, and the degree of influence of the other personnel on the state of the target individual, as the state propagation amount from the other personnel to the target individual, and to derive the sum of the current state value of the target individual and the state propagation amount from the other personnel to the target individual as the future state value of the target individual.

[0008] In the scheme, the controller can be configured to, for each of the plurality of other persons in the group, derive the product of the current state value of the other persons, the current amount of communication between the target individual and the other persons, and the degree of influence of the other persons on the state of the target individual, as the state propagation amount from the other persons to the target individual, and derive the sum of the current state value of the target individual and the state propagation amount from the plurality of other persons to the target individual as the future state value of the target individual.

[0009] In the scheme, the controller can be configured to, for each of a plurality of other persons in the group, acquire the state propagation amount from the other persons to the target individual, compare the acquired state propagation amounts from the plurality of other persons to the target individual, and based on the comparison results, suggest changes to the amount of communication between the target individual and the other persons to improve the future state value of the target individual.

[0010] In the scheme, the controller can be configured to predict the future state value of each of a plurality of individuals in the group, and based on the predicted future state value of each of the plurality of individuals, predict a future group state value indicating the future state of the group.

[0011] In the scheme, the controller can be configured to derive the target communication volume among the plurality of people to improve the predicted future group state value, and compare the current communication volume between two target people in the group with the target communication volume, and suggest changes to the communication volume based on the comparison result.

[0012] A second aspect of the present invention relates to a member determination system comprising a controller. The controller is configured to: temporarily define a plurality of individuals as candidates for forming a group; for each of the temporarily defined plurality of individuals, acquire a current state value indicating the current state of that individual; for each of the temporarily defined plurality of individuals, acquire a state propagation amount indicating a state quantity propagated from other individuals in the group to that individual, the state quantity being predicted based on past communication between that individual and the other individuals; for each of the temporarily defined plurality of individuals, predict a future state value indicating the future state of that individual based on the acquired current state value and the acquired state propagation amount to that individual; predict a future group state value indicating the future state of the group based on the predicted future state values ​​of each of the plurality of individuals; and determine the temporarily defined plurality of individuals as members of the group when the predicted future group state value satisfies predetermined conditions for a good state.

[0013] A third aspect of the present invention relates to a state prediction method executed by a computer. The state prediction method includes acquiring a current state value indicating the current state of a target individual in a group; acquiring a state propagation amount indicating the amount of state propagated from other individuals to the target individual through communication between the target individual and other individuals in the group; and predicting a future state value indicating the future state of the target individual based on the acquired current state value and the acquired state propagation amount.

[0014] According to these embodiments of the present invention, a technique is provided that can predict the future state of individuals within a group. Attached Figure Description

[0015] The features, advantages, and technical and industrial significance of exemplary embodiments of the present invention will now be described with reference to the accompanying drawings, wherein like reference numerals denote like elements, and wherein:

[0016] Figure 1 This is a diagram used to describe the functionality of the communication support system according to the first embodiment;

[0017] Figure 2 yes Figure 1 The following diagram illustrates the functions of the communication support system;

[0018] Figure 3 This is a diagram illustrating the configuration of the communication support system according to the first embodiment;

[0019] Figure 4 This is a graph showing an example of a member's current stress level and conversation volume;

[0020] Figure 5 It shows the basis Figure 4 A graph showing the current stress state value and the predicted future stress state value based on the dialogue volume;

[0021] Figure 6 This is a diagram illustrating an example of the relationship between each of multiple dialogue volume patterns, future individual stress state values, and future group stress state values;

[0022] Figure 7 It is shown Figure 3 The flowchart of the communication support process of the communication support system;

[0023] Figure 8 This is a flowchart illustrating another example of communication support processing;

[0024] Figure 9 This is a diagram illustrating the member determination function of the communication support system according to the second embodiment;

[0025] Figure 10 yes Figure 9 The following diagram describes the membership determination function of the communication support system;

[0026] Figure 11 This is a diagram illustrating the configuration of the communication support system according to the second embodiment; and

[0027] Figure 12 It is shown Figure 11 The flowchart for member determination processing in the communication support system. Detailed Implementation

[0028] First Embodiment

[0029] Figure 1 This is a diagram illustrating the functionality of the communication support system according to the first embodiment. The communication support system supports... Figure 1 Successful communication occurs within group G1. Successful communication means that the individual members P1 through P4 (hereinafter also referred to as "individuals" or "persons") of group G1 communicate while maintaining a good state of mind. The number of members can be plural. An example of group G1 having an assembly will be illustrated below. The detailed processing of the system will be described later.

[0030] The communication support system periodically acquires the current state values ​​S1(t) to S4(t) indicating the current state of each member P1 to P4 of group G1, and the current communication quantities W12(t), W13(t), W14(t), W23(t), W24(t), and W34(t) between two members in each of multiple combinations of any two members drawn from group G1. The present time is taken as t. For example, the current state value S1(t) represents the current state of member P1, and the communication quantity W13(t) represents the current communication quantity between members P1 and P3. There is no communication between members P1 and P2, and their communication quantity W12(t) is zero. Figure 1 In the example, the magnitudes of the current state values ​​S1(t) to S4(t) are represented by the length of the bar chart, and the magnitudes of the AC quantities W13(t), etc., are represented by the thickness of the solid lines between the members.

[0031] Current state values ​​can be, for example, stress levels, mood levels, performance levels, values ​​indicating the degree of fatigue, or values ​​indicating the degree of happiness. A current state value can be obtained by substituting at least two of these values ​​into a predetermined calculation formula. The predetermined calculation formula can be, for example, a formula that calculates a sum by weighting each value. Current state values ​​can be detected using known technologies such as sensors. For example, mood values ​​can be obtained by analyzing images of members taken by a camera, performing facial expression analysis of members, and assessing members' emotions. Current state values ​​do not necessarily have to be detected values; they can also be values ​​reported by individual members.

[0032] The current state value can be a value ranging from negative to positive, where a larger positive value indicates a worse state and a smaller negative value indicates a better state. For example, when the current state value represents a stress state, a larger positive value indicates a worse state and a smaller negative value indicates a better state. It is also possible that a larger positive value represents a better state and a smaller negative value represents a worse state. For example, when the current state value represents happiness, a larger positive value represents a better state and a smaller negative value represents a worse state.

[0033] Communication volume is a numerical value representing the degree of interaction between two members per unit of time. Examples include the amount of conversation between two members per unit of time, the amount of time one member smiles at the other, the amount of time one member looks at the other, or the amount of time two members exchange gestures. Communication volume can be obtained by substituting at least two of these quantities into a predetermined formula. Communication volume is a value greater than or equal to zero. Communication volume can be detected using known technologies such as microphones, cameras, and sensors.

[0034] The communication support system obtains the state propagation quantity between two members in each of multiple combinations of any two members drawn from population G1, based on the current state values ​​S1(t) to S4(t) and the current communication quantities W12(t) to W34(t). The state propagation quantity indicates the amount of state propagated from one member to another through communication between the two members, and is a value ranging from negative to positive. The state propagation quantity from the first member to the second member can be proportional to, for example, the current state value of the first member and the current communication quantity between the members. The state propagation quantity from the first member to the second member can also be proportional to a coefficient representing the second member's sensitivity to the state of its partner. Therefore, for example, the state propagation quantity from member P1 to member P3 is likely to be different from the state propagation quantity from member P3 to member P1. Since the communication quantity W12(t) between members P1 and P2 is zero, it is assumed that the state propagation quantities from member P1 to member P2 and from member P2 to member P1 are zero.

[0035] For each of members P1 through P4, the communication support system predicts a future state value indicating the member's future state based on the member's current state values ​​S1(t) through S4(t) and the state propagation amounts from multiple other members to that member. The future is represented by time t+1. For example, the future state value S1(t+1) is a numerical value representing the future state of member P1 and can be obtained by adding the state propagation amounts from member P2 to P1, from member P3 to P1, and from member P4 to P1 to the current state value S1(t) of member P1. Future state values ​​S2(t+1) through S4(t+1) can also be obtained in the same way.

[0036] The communication support system predicts the group state value Sm, which indicates the future state of group G1, based on future state values ​​S1(t+1) to S4(t+1). Figure 1 In the example, assume that the future group state value Sm is greater than the threshold Th. When the group state value Sm is greater than the threshold Th, the communication support system suggests to members P1 to P4 that the communication amount be changed so that the future group state value Sm is less than the threshold Th.

[0037] For example, the communication support system recommends increasing the amount of communication between members P1 and P2, between members P1 and P3, between members P2 and P3, and between members P3 and P4, while recommending reducing the amount of communication between members P1 and P4, and between members P2 and P4.

[0038] Figure 2 yes Figure 1The following diagram describes the functionality of the communication support system. Assume that each member P1 through P4 changes their communication quantity based on a suggestion from the communication support system. Therefore, the future state values ​​S1(t+1) through S4(t+1) newly predicted by the communication support system differ from those of the members. Figure 1 The value in. And Figure 1 The values ​​in the threshold Th are different, and the future group state value Sm is also less than the threshold Th. Therefore, compared with... Figure 1 Compared to the previous state, members are more likely to maintain a good state during the gathering, and thus can more smoothly promote communication within group G1.

[0039] The following section details an example where the current state value and future state value are stress state values, and the communication volume is dialogue volume. Stress state values ​​are, for example, values ​​ranging from negative to positive, where larger positive values ​​represent stronger stress, and smaller negative values ​​represent greater relaxation. The stress states represented by positive and negative values ​​can be reversed. Dialogue volume is a value above zero.

[0040] Figure 3 This is a diagram illustrating the configuration of an communication support system 1 according to a first embodiment. The communication support system 1 includes microphones 2, cameras 4, sensors 6, storage devices 7, output devices 8, and processing devices 10. Although not shown, the communication support system 1 has multiple microphones 2, multiple cameras 4, and multiple sensors 6. The communication support system 1 can also be referred to as a "state prediction system." The processing device 10 is an example of a controller in this invention.

[0041] Microphone 2 acquires the conversations of multiple members P1 to Pk (k being an integer greater than 2) constituting group G1, and provides the acquired voice data to processing device 10. Camera 4 captures images of members P1 to Pk, and provides the captured image data to processing device 10. Sensor 6 is attached to the body of each of members P1 to Pk, detects the heart rate of each member P1 to Pk, and provides the detected heart rate data to processing device 10. Output device 8 includes at least one of, for example, a display capable of outputting images and a voice output device capable of outputting voice, to output various information. Output device 8 may be included in a mobile terminal such as a smartphone owned by each of members P1 to Pk.

[0042] The processing device 10 includes a first analysis unit 12, a second analysis unit 14, a first acquisition unit 16, a second acquisition unit 18, an individual state prediction unit 20, a group state prediction unit 22, an export unit 24, and a suggestion unit 26. The processing device 10 can be, for example, a personal computer, a smartphone, a server device, etc.

[0043] The configuration of the processing device 10, in terms of hardware, can be implemented by a computer's CPU, memory, or other LSI, and in terms of software, can be implemented by a program loaded in memory; however, here, functional blocks implemented through the cooperation of hardware and software are depicted. Therefore, those skilled in the art will understand that functional blocks can be implemented in various forms by hardware alone, by software alone, or by a combination thereof.

[0044] The second analysis unit 14 has a facial image recognition function and identifies which registered user's facial image is represented by multiple facial images captured by camera 4. Storage device 7 stores the feature values ​​of the registered user's facial image. The second analysis unit 14 compares the feature values ​​of the registered user's facial image stored in storage device 7 with the feature values ​​of the facial image in the image data, performs authentication processing on the facial image of each member P1 to Pk, and determines whether each member P1 to Pk is a registered user. When the second analysis unit 14 determines that each member P1 to Pk is a registered user, the second analysis unit 14 provides the first analysis unit 12 and the second acquisition unit 18 with the identification information about the registered user, which serves as the identification information for each member P1 to Pk. Information for designating multiple members, such as for administrators, can be input to the input unit (not shown) of the processing device 10.

[0045] The second analysis unit 14 performs speech analysis on the speech data provided by the microphone 2 and image analysis on the image data provided by the camera 4, periodically detecting the amount of dialogue between members Pi and Pj per unit time up to the present (i, j≤k). For example, the longer the dialogue time between members Pi and Pj per unit time, the larger the amount of dialogue Wij(t) is derived by the second analysis unit 14. The second analysis unit 14 provides the detected amount of dialogue Wij(t) to the second acquisition unit 18.

[0046] The second analysis unit 14 has a speaker recognition function and identifies which registered user's voice data is represented by the voice data provided from the microphone 2. The voice templates of registered users are registered in the storage device 7, and the second analysis unit 14 compares the provided voice data with the voice templates stored in the storage device 7 to identify the speaker. The second analysis unit 14 identifies the position of each of members P1 to Pk from the image data and evaluates which member the speech is directed at based on the direction the facial image of the member speaking is facing. The second analysis unit 14 may have natural language processing capabilities and can analyze the dialogue between members, and evaluate which members are talking to each other based on the analysis results. Known techniques can be used to detect the amount of dialogue Wij(t).

[0047] The first analysis unit 12 periodically detects the pressure state values ​​S1(t) to Sk(t) of each of the members P1 to Pk by analyzing the heart rate data provided by the sensor 6, and provides the detected pressure state values ​​S1(t) to Sk(t) to the first acquisition unit 16 and the second acquisition unit 18. Known techniques can be used to detect the pressure state values. It should be noted that the sensor 6 may not be required, and the level of pressure reported by each of the members P1 to Pk may be periodically input into the input unit (not shown) of the processing device 10 by an administrator or the like via digital or voice input. In this case, the first analysis unit 12 detects the pressure state values ​​S1(t) to Sk(t) by analyzing the input information.

[0048] The first acquisition unit 16 and the second acquisition unit 18 perform the following processing on each of the members P1 to Pk of the group G1.

[0049] The first acquisition unit 16 acquires the current pressure state value Si(t) of the target member Pi detected by the first analysis unit 12. This operation is an example of the first acquisition unit 16 acquiring the current state value Si(t) indicating the current state of the target individual Pi in the group G1.

[0050] The second acquisition unit 18 acquires a state propagation quantity for each of the other persons in group G1. This state propagation quantity indicates the state quantity propagated from other persons Pj to target member Pi through communication between target member Pi and other persons Pj in group G1. Specifically, for each of the multiple other persons in group G1, the second acquisition unit 18 acquires the current stress state value Sj(t) of other person Pj detected by the first analysis unit 12, the current dialogue quantity Wij(t) between target member Pi and other persons Pj detected by the second analysis unit 14, and the degree of influence λ of other persons Pj on the state of target member Pi. The degree of influence λ is a value specific to target member Pi. For each of the multiple other persons in group G1, the second acquisition unit 18 derives the product of the acquired current stress state value Sj(t) of other person Pj, the current dialogue quantity Wij(t) between target member Pi and other persons Pj, and the degree of influence λ of other persons Pj on the state of target member Pi, as the state propagation quantity from other persons Pj to target member Pi.

[0051] The operation is exemplified as follows: For each of the multiple other persons in group G1, the second acquisition unit 18 acquires the current state value Sj(t) indicating the current state of the other person Pj, the current communication amount Wij(t) between the target individual Pi and other persons Pj, and the degree of influence λ of other persons Pj on the state of individual Pi, and derives the product of the acquired current state value Sj(t) of other persons Pj, the current communication amount Wij(t) between individual Pi and other persons Pj, and the degree of influence λ of other persons Pj on the state of individual Pi as the state propagation amount from other persons Pj to individual Pi.

[0052] The storage device 7 pre-stores the degree of influence λ unique to each registered user in association with the registered user's identification information. The value of the degree of influence λ is zero or higher. Since the degree of influence λ varies from individual registered users, it is set for each registered user. The second acquisition unit 18 acquires the identification information of the target member and retrieves the degree of influence λ associated with the acquired identification information from the storage device 7.

[0053] The degree of influence λ can be set for each combination of individuals and other people. In this way, the state propagation can also reflect the different degrees of influence of the partner's state on the individual, depending on whether the partner is skilled or clumsy, thus enabling the acquisition of more accurate state propagation.

[0054] The degree of influence λ can be varied based on factors such as the physical condition of the members. Regarding the degree of influence λ, it can be input into the processing device 10 by an administrator or similar entity, storing the values ​​reported by each member P1 to Pk before the meeting in the storage device 7, or storing the values ​​derived from sensor (not shown) detection values ​​in the storage device 7. For example, the degree of fatigue or concentration of the members can be periodically detected by sensors, and the processing device 10 can periodically derive the degree of influence based on the detection results. Alternatively, the sleep time and average sleep time reported by the members the previous day can be input into the processing device 10, and the processing device 10 can derive the degree of influence based on the ratio of the previous day's sleep time to the average sleep time.

[0055] For each of members P1 to Pk, the individual state prediction unit 20 predicts the future stress state value Si(t+1) of the target member Pi based on the current stress state value Si(t) of the target member Pi obtained by the first acquisition unit 16 and the state propagation amount from other personnel to the target member Pi obtained by the second acquisition unit 18. That is, the individual state prediction unit 20 predicts the future state values ​​S1(t+1) to Sk(t+1) of each member P1 to Pk in the group G1. Specifically, for each of members P1 to Pk, the individual state prediction unit 20 sets the sum of the current stress state value Si(t) of the target member Pi and the state propagation amount from multiple other personnel to the target member Pi as the future stress state value Si(t+1) of the target member Pi.

[0056] The operation is exemplified as follows: For each of the individuals P1 to Pk in the group G1, the individual state prediction unit 20 predicts the future state value Si(t+1) that indicates the future state of individual Pi based on the current state value Si(t) of individual Pi and the state propagation amount from multiple other individuals to individual Pi.

[0057] Figure 4 Examples of the current stress status values ​​and dialogue volume for members P1 through P3 are shown. Figure 5 It shows according to Figure 4 The current stress level and the predicted future stress level based on the volume of conversations.

[0058] exist Figure 4 In the example, assume that the current stress state value S1(t) of member P1 is "15", the current stress state value S2(t) of member P2 is "18", and the current stress state value S3(t) of member P3 is "10". Assume that the current dialogue volume W12(t) between members P1 and P2 is "1.0", the current dialogue volume W13(t) between members P1 and P3 is "0.2", and the current dialogue volume W23(t) between members P2 and P3 is "0".

[0059] In this example, the future pressure state value S1(t+1) of member P1 is represented by the following equation (1) as described above.

[0060] S1(t+1)=S1(t)+S2(t)×W12(t)×λ12+S3(t)×W13(t)×λ13 Equation (1)

[0061] Here, λ12 is the degree of influence of member P2 on the state of member P1, and λ13 is the degree of influence of member P3 on the state of member P1. As mentioned above, λ12 and λ13 can be equal. S2(t)×W12(t)×λ12 is the state propagation quantity from member P2 to member P1. S3(t)×W13(t)×λ13 is the state propagation quantity from member P3 to member P1.

[0062] For example, assuming λ12 = 0.1 and λ13 = 0.2, S1(t+1) becomes "17.2". In this way, the future stress state value S1(t+1) of member P1 can be predicted based on the current stress state values ​​S1(t) to S3(t) of members P1 to P3, as well as the dialogue amounts W12(t) and W13(t). Since the future state value of an individual is predicted based on the state propagation amount, it is possible to predict the future state value of an individual while taking into account the influence of communication with other people.

[0063] For example, when the stress state value Sj(t) of a member Pi's communication partner Pj is negative, the state propagation from partner Pj becomes negative. Therefore, member Pi's future stress state value Si(t+1) becomes lower than the current stress state value Si(t). That is, when the partner is in a relaxed state, the member's stress state can be improved by the member being influenced by the partner's state through communication with the partner.

[0064] For each of members P1 to Pk, the individual state prediction unit 20 can predict the future state value Si(t+1) of the target member Pi based on the current state value of the target member Pi and the state propagation amount from one other person Pj to the target member Pi. That is, if the state propagation amount from multiple other people to the target member Pi is greater than 0, all state propagation amounts need not be considered. In this case, one other person Pj can be, for example, the member with the largest absolute value of the state propagation amount to the target member Pi. There may be more than two other people. Even without using state propagation amounts with relatively small absolute values, it is possible to suppress the deterioration of the prediction accuracy of future state values ​​by using state propagation amounts with the largest absolute values. In addition, the processing load can be reduced.

[0065] The group state prediction unit 22 predicts the group G1 stress state value Sm, which indicates the future state of the group G1, based on the future stress state values ​​S1(t+1) to Sk(t+1) of each member P1 to Pk predicted by the individual state prediction unit 20. The group state prediction unit 22 sets the statistical value of each future stress state value S1(t+1) to Sk(t+1) of members P1 to Pk as the group G1 stress state value Sm. The statistical value is, for example, the average value.

[0066] The operation is exemplified as follows: the group state prediction unit 22 predicts the group state value Sm, which indicates the future state of the group G1, based on the future state values ​​S1(t+1) to Sk(t+1) of each individual P1 to Pk predicted by the individual state prediction unit 20.

[0067] When the predicted future stress state value Sm of group G1 is greater than the threshold Th, the derivation unit 24 derives the target dialogue quantity Wij_m(t) between members P1 to Pk in group G1, thereby improving the predicted future stress state value Sm of group G1. The threshold Th can be appropriately set through experiments or simulations.

[0068] The operation is exemplified as follows: the derivation unit 24 derives the target communication quantity between individuals P1 to Pk in the group G1, such that when the predicted future state value Sm of the group G1 meets the predetermined conditions for the adverse state, the predicted future group state value Sm is improved.

[0069] The derivation unit 24 virtually sets multiple modes with different dialogue volumes. For each mode, based on the current stress state values ​​S1(t) to Sk(t) of each member P1 to Pk, it derives the future stress state values ​​S1(t+1) to Sk(t+1) of each member P1 to Pk, and derives the average of the derived future stress state values ​​S1(t+1) to Sk(t+1) as the stress state value Sm of the group G1. In this way, it is possible to evaluate which mode of dialogue volume results in a smaller stress state value Sm for the group G1.

[0070] The export unit 24 sets the mode of dialogue volume with the minimum stress state value Sm of group G1 as the target dialogue volume Wij_m(t) between members P1 and Pk. The export unit 24 can also set the mode of dialogue volume with the stress state value Sm of group G1 equal to or lower than a predetermined value as the target dialogue volume Wij_m(t) between members P1 and Pk.

[0071] Figure 6 An example is shown showing the relationship between the future stress state values ​​S1(t+1) to Sk(t+1) of an individual and the future stress state value Sm of the group G1 in each of several dialogue volume patterns. In this example, the dialogue volume patterns cover all combinations of values ​​such that the dialogue volumes Wij(t) differ by 0.1.

[0072] exist Figure 6In the example, it is assumed that the stress state value Sm of group G1 is minimized in dialogue mode 1. Therefore, the dialogue volume Wij(t) of mode 1 is set as the target dialogue volume Wij_m(t). Thus, for example, the target dialogue volume W12_m(t) between members P1 and P2 is 0.1, the target dialogue volume W13_m(t) between members P1 and P3 is 0.1, and the target dialogue volume W14_m(t) between members P1 and P4 is 0.1.

[0073] When the stress state value Sm of group G1 is greater than the threshold Th, for each of the multiple combinations of two members Pi and Pj in group G1, the suggestion unit 26 compares the current dialogue volume Wij(t) of the two members with the target dialogue volume Wij_m(t). If the current dialogue volume Wij(t) is less than the target dialogue volume Wij_m(t), the suggestion unit 26 suggests increasing the dialogue volume; if the current dialogue volume Wij(t) is greater than the target dialogue volume Wij_m(t), the suggestion unit 26 suggests decreasing the dialogue volume, for example, by having the output device 8 output an image or voice indicating the suggestion content.

[0074] The operation is illustrated in the following example: When the future group state value Sm is greater than the threshold Th, the suggestion unit 26 compares the current communication quantity Wij(t) with the target communication quantity Wij_m(t) for two target individuals Pi and Pj in group G1, and suggests changing the communication quantity based on the comparison result. In this way, it is possible to suggest changes to the communication quantity between two individuals in group G1, so that when the state is predicted to deteriorate, the future state of group G1 is unlikely to deteriorate. Therefore, it enables members to discover what measures should be taken to improve the future state of group G1 better than predicted.

[0075] When the stress state value Sm of group G1 is lower than the threshold Th, it is recommended that unit 26 not change the dialogue volume.

[0076] The processing device 10 periodically repeats the above processing.

[0077] Figure 7 It is shown Figure 3 The flowchart shows the communication support process of Communication Support System 1. This process begins when the meeting starts.

[0078] The second analysis unit 14 designates k members P1 to Pk (k≤N) (step S10), and when no instruction to end processing is input ("No" in step S12), the second analysis unit 14 detects the current dialogue volume Wij(t) (i, j≤k) between members Pi and Pj (step S14). The first analysis unit 12 detects the current stress state values ​​S1(t) to Sk(t) of each member P1 to Pk (step S16), and the personal state prediction unit 20 derives the future stress state values ​​S1(t+1) to Sk(t+1) of each member P1 to Pk based on the current dialogue volume Wij(t) and the stress state values ​​S1(t) to Sk(t) (step S18).

[0079] The group state prediction unit 22 derives the average value Sm of the future pressure state values ​​S1(t+1) to Sk(t+1) of each member P1 to Pk (step S20), and when the average value Sm is equal to or lower than the threshold Th ("No" in step S22), the process returns to step S12.

[0080] When the average value Sm is greater than the threshold Th ("Yes" in step S22), the derivation unit 24 specifies the pattern of the target dialogue quantity Wij_m(t) that minimizes the average value Sm (step S24), and derives Wij_s(t) = Wij(t) - Wij_m(t) (step S26).

[0081] Suggestion unit 26 urges members Pi and Pj who satisfy Wij_s(t)<0 to speak (step S28), and urges members Pi and Pj who satisfy Wij_s(t)>0 to remain silent (step S30), then returns to step S12. When the instruction to end processing is input ("Yes" in step S12), the processing ends.

[0082] The methods for determining which member is recommended to change their conversation volume are not limited to the examples above.

[0083] Figure 8 This is a flowchart illustrating another example of communication support processing. Instead... Figure 7 Steps S28 and S30 are executed. Figure 8 The actions in steps S42 to S50.

[0084] exist Figure 7 After step S26, the derivation unit 24 derives Wij_s_m(t) = max(|Wij_s(t)|) (step S42). max(|Wij_s(t)|) indicates that Wij_s(t) has the largest absolute value. That is, the derivation unit 24 sets Wij_s(t) with the largest absolute value as Wij_s_m(t).

[0085] When X<Wij_s_m(t)<Y ("Yes" in step S44), the process returns to step S12. In this case, the suggestion unit 26 does not suggest changing the conversation volume. X is a lower limit value and Y is an upper limit value, which can be appropriately determined through experiments or simulations. For example, X is a negative value and Y is a positive value. That is, when Wij_s_m(t) is close to zero, the suggestion unit 26 does not suggest changing the conversation volume.

[0086] When X<Wij_s_m(t)<Y is not satisfied ("No" in step S44) and Wij_s_m(t)≤X is satisfied ("Yes" in step S46), the suggestion unit 26 urges member Pi and member Pj to speak (step S48), and the process returns to step S12. When Wij_s_m(t)≤X is not satisfied ("No" in step S46), the suggestion unit 26 urges member Pi and member Pj to remain silent (step S50), and the process returns to step S12.

[0087] In this example, the change of conversation volume is only suggested for two members having the maximum absolute value of the difference between the current conversation volume and the target conversation volume. Therefore, a suggestion that can improve the future group state value Sm can be made while minimizing the number of members to be suggested.

[0088] The operation of step S44 can be omitted, and the determination of step S46 can be performed after step S42. In this case, the determination in step S46 is changed: when Wij_s_m(t)<0 (step S46), the suggestion unit 26 urges member Pi and member Pj to speak (step S48), and the process returns to step S12. When Wij_s_m(t)>0 (step S46), the suggestion unit 26 urges member Pi and member Pj to remain silent (step S50), and the process returns to step S12. When Wij_s_m(t)=0 (step S46), the process returns to step S12. Even in this modified example, a suggestion that can improve the future group state value Sm can be made while minimizing the number of members to be suggested.

[0089] Furthermore, in Figure 7 the process of , in step S28, the suggestion unit 26 may urge member Pi and member Pj satisfying Wij_s(t)<X to speak, and in step S30, the suggestion unit 26 may urge member Pi and member Pj satisfying Wij_s(t)>Y to remain silent. X and Y are the aforementioned lower limit value and upper limit value. In this modified example, members to be suggested can be limited to those predicted to contribute to the improvement of the future group state value Sm to a relatively large extent.

[0090] As described above, the following example is given, in which the mode of setting the dialogue volume that minimizes the pressure state value Sm of group G1 to the target dialogue volume between members P1 and Pk is set when the future pressure state value Sm of group G1 is greater than the threshold Th. However, it is possible to suggest changing the dialogue volume without deriving the target dialogue volume.

[0091] Specifically, recommendation unit 26 compares the state propagation amounts from multiple other individuals in group G1 to the target member, and based on the comparison results, recommends changing the amount of communication between the target member and other members, thereby improving the predicted future state value of the target member. Recommendation unit 26 specifies the positive and maximum state propagation amount from other individuals to the target member and recommends reducing the amount of communication between that member and other individuals. By reducing the amount of communication between that member and other individuals, compared to reducing the previous amount of communication, the state propagation amount from other individuals to that member can be smaller, and the future state value of that member can be smaller.

[0092] Recommendation unit 26 can specify a negative and minimum state propagation amount from other personnel to the target member, and recommend increasing the amount of communication between that member and other personnel. By increasing the amount of communication between that member and other personnel, compared to increasing the amount of communication before, the state propagation amount from other personnel to that member can be reduced, and the future state value of that member can be reduced.

[0093] In a variant example, the suggestion unit 26 can decide whether a suggestion exists for more than one member, and may not decide for all members. In this case, the member targeted for the decision can be pre-registered in the storage device 7 by an administrator or the like, and can be a member whose influence λ on the member's state is equal to or greater than a predetermined value, or a member whose influence λ on the member's state is the greatest. The second acquisition unit 18 can acquire only the state propagation amount of the member targeted for the decision, and the individual state prediction unit 20 can predict only the future stress state value of the member targeted for the decision.

[0094] In the variant example, it is suggested to change the amount of communication between the target member and other people, thus enabling the member to discover what actions should be taken to improve their future state. In this way, the future stress state value Sm of group G1 can be improved. Furthermore, since the target dialogue volume is not derived, the processing is simplified.

[0095] Second Embodiment

[0096] In the second embodiment, a process is added to the first embodiment to determine the members of group G1 from multiple individuals so that the future group state value indicates a good state. The differences from the first embodiment will be described in detail below.

[0097] Figure 9 This diagram illustrates the member determination function of the communication support system according to the second embodiment. Before the meeting begins, the communication support system temporarily designates four members P1 to P4 from, for example, seven individuals P1 to P7, as candidates to form group G1. At this time, these seven individuals P1 to P7 do not need to be gathered at a communicable distance.

[0098] The communication support system acquires the current state values ​​S1(t) to S4(t) of each member P1 to P4, and predicts the state propagation quantity based on the past communication quantities W12, W13, W14, W23, W24, and W34 between the two members in each of multiple combinations of any two members drawn from the population G1. The method for deriving the state propagation quantity is the same as that in the first embodiment, and past communication quantities can be used instead of current communication quantities.

[0099] For each member P1 to P4, the communication support system predicts a future state value indicating the member's future state based on the member's current state values ​​S1(t) to S4(t) and the state propagation from multiple other members to that member. The communication support system predicts the future group state value Sm based on the future state values ​​S1(t+1) to S4(t+1). Figure 9 In the example, assume the future group state value Sm is greater than the threshold Th. In this case, members P1 through P4 are inappropriate, and therefore the communication support system temporarily assigns other members. The communication support system repeats the temporary assignment of members until the future group state value Sm becomes equal to or lower than the threshold Th.

[0100] Figure 10 yes Figure 9 The following diagram describes the member determination function of the communication support system. The communication support system temporarily designates four candidates, P1, P2, P5, and P6, from individuals P1 to P7, to form group G1. The future state values ​​S1(t+1), S2(t+1), S5(t+1), and S6(t+1) newly predicted by the communication support system differ from... Figure 9 The value in. And Figure 9 The values ​​in the data are different, assuming the future group state value Sm is also less than the threshold Th. Therefore, the communication support system formally determines members P1, P2, P5, and P6 as members of group G1 and notifies them. Upon receiving the notification, members P1, P2, P5, and P6 gather together to begin a meeting. Figure 9 Compared to members P1 to P4, each of members P1, P2, P5, and P6 is able to easily maintain a good state during the assembly and can smoothly facilitate communication within group G1. After the assembly begins, the communication support system performs the processing of the first embodiment.

[0101] Figure 11 This diagram illustrates the configuration of the communication support system 1A according to the second embodiment. The communication support system 1A may also be referred to as a "member determination system." Besides... Figure 3 In addition to the configuration shown, the processing device 10A also includes a temporary setting unit 30 and a determination unit 32. An example will be described where the current state value and future state value are pressure state values ​​and the communication quantity is a dialogue quantity. The processing device 10A is an example of a controller in this invention.

[0102] Temporary setting unit 30 temporarily sets multiple individuals as candidates to form group G1. First acquisition unit 16 acquires the current stress state value of each of the temporarily set multiple individuals.

[0103] The second acquisition unit 18 acquires the past dialogue volume between each of the temporarily set plurality of personnel and other personnel in group G1, and acquires the state propagation volume, which indicates the state volume propagated from other personnel to that personnel, and the state volume is predicted based on the acquired past dialogue volume. The storage device 7 stores the past dialogue volume between two registered users for each of the plurality of combinations of any two registered users. The second acquisition unit 18 acquires the past dialogue volume from the storage device 7.

[0104] For each of the temporarily set personnel, the personal state prediction unit 20 predicts the future stress state value of that personnel based on the current stress state value of that personnel and the state propagation amount of that personnel.

[0105] The group state prediction unit 22 predicts the future stress state value Sm of group G1 based on the predicted future stress state value of each of the multiple individuals.

[0106] When the predicted future pressure state value Sm of group G1 meets the predetermined conditions for a good state, the determination unit 32 determines the temporarily set multiple people as multiple members of group G1, and causes the output device 8 to output information for specifying the determined multiple members.

[0107] When the future stress state value Sm of group G1 does not meet the predetermined conditions for a good state, the temporary setting unit 30 temporarily resets the multiple individuals who make up the candidates of group G1, and the first acquisition unit 16, the second acquisition unit 18, the individual state prediction unit 20, the group state prediction unit 22 and the determination unit 32 perform the above-mentioned processing again.

[0108] Figure 12 It is shown Figure 11It is a flowchart of member determination processing for the communication support system 1A. A temporary setting unit 30 temporarily sets k members from N people (k<N) (step S60). A second acquisition unit 18 acquires the past conversation volume between these members (step S62), and a first analysis unit 12 detects the current pressure state value of each member (step S64). An individual state prediction unit 20 derives the future pressure state value of each member based on the past conversation volume and the current pressure state value (step S68).

[0109] A group state prediction unit 22 derives the average value Sm of the future pressure state values of the k members (step S70), and when the average value Sm is greater than the threshold Th ("YES" in step S72), the process returns to step S60. When the average value Sm is equal to or lower than the threshold Th ("NO" in step S72), a determination unit 32 determines the temporarily set members as official members of the gathering (step S74), and the process ends.

[0110] In the member determination processing, k may be a predetermined constant value, or k may be changed when the average value Sm is greater than the threshold Th ("YES" in step S72). When k is changed when the average value Sm is greater than the threshold Th ("YES" in step S72), the temporary setting unit 30 may increase k by a predetermined number or decrease k by a predetermined number, and then execute the operation of step S10.

[0111] According to this embodiment, a plurality of members that are expected to have a relatively improved future state of the group G1 through communication can be determined from a plurality of people.

[0112] The present invention has been described above based on embodiments. It should be noted that the embodiments are merely examples, and those skilled in the art should understand that various modified examples of combinations of components and processes can be made therefrom, and these modified examples are also within the scope of the present invention.

[0113] For example, it is not necessary to provide the derivation unit 24 and the suggestion unit 26, and it is not necessary to suggest changing the conversation volume. In this case, the first acquisition unit 16 causes the output device 8 to output the individual's current state value so that the individual can specify the current state value, and the individual state prediction unit 20 causes the output device 8 to output the individual's future state value. For example, the output device 8 may Figure 1Displayed as an image. Since it is possible to predict whether an individual's state will deteriorate in the future based on the output current state value and future state value, it is possible to specify whether communication measures need to be taken for that individual. Therefore, it becomes easy to maintain an individual's future state in a good condition. Alternatively, the individual state prediction unit 20 can cause the output device 8 to output the individual's future state value, such that the individual can only specify a future state value if the individual's future state value is greater than a threshold. Furthermore, the group state prediction unit 22 can cause the output device 8 to output future group state values, and can cause the output device 8 to output future group state values ​​only if the future group state value is greater than a threshold. Since it is possible to predict whether the state of group G1 will deteriorate in the future based on the future group state value, it is possible to specify whether communication measures need to be taken for group G1.

Claims

1. A state prediction system, characterized in that... Includes a controller, the controller being configured to Retrieve the current state value of the target individual within the indicated group. The state propagation quantity is the amount of state information that is propagated from other individuals to the target individual through communication between the target individual and other individuals in the group. Based on the current state value of the target individual and the obtained state propagation amount, predict the future state value indicating the future state of the target individual. The system acquires information including the current state value indicating the current state of the other personnel, the current communication volume between the target individual and the other personnel, and the degree of influence of the other personnel on the target individual's state. Based on the acquired information, the state propagation amount from the other personnel to the target individual is derived.

2. The state prediction system according to claim 1, characterized in that, The controller is configured to The product of the current state values ​​of the other personnel, the current communication volume between the target individual and the other personnel, and the degree of influence of the other personnel on the state of the target individual is derived as the state propagation quantity from the other personnel to the target individual. The sum of the current state value of the target individual and the state propagation amount from the other individuals to the target individual is derived as the future state value of the target individual.

3. The state prediction system according to claim 1, characterized in that, The controller is configured to For each of the other individuals in the group, the product of the other individual's current state value, the current communication volume between the target individual and the other individuals, and the degree of influence of the other individuals on the target individual's state is derived as the state propagation quantity from the other individuals to the target individual. The sum of the current state value of the target individual and the state propagation amount from the plurality of other individuals to the target individual is derived as the future state value of the target individual.

4. The state prediction system according to any one of claims 1 to 3, characterized in that, The controller is configured to For each of the other individuals in the group, obtain the state propagation amount from the other individuals to the target individual, and The state propagation amounts from the plurality of other persons to the target individual are compared, and based on the comparison results, it is suggested to change the amount of communication between the target individual and the other persons to improve the future state value of the target individual.

5. The state prediction system according to any one of claims 1 to 3, characterized in that, The controller is configured to Predict the future state value of each of the multiple individuals in the group, and Based on the predicted future state value of each of the plurality of individuals, a future group state value indicating the future state of the group is predicted.

6. The state prediction system according to claim 5, characterized in that, The controller is configured to Derive the target communication volume among the multiple individuals to improve the predicted future group state value, and The current amount of communication between two target individuals in the group is compared with the target amount of communication, and a change in the amount of communication is recommended based on the comparison result.

7. A member determination system, characterized in that... Includes a controller, the controller being configured to For now, several individuals will be designated as candidates to form the group. For each of the temporarily defined personnel, obtain the current state value indicating the current state of that personnel. For each of the temporarily defined plurality of individuals, obtain a state propagation quantity indicating the state quantity propagated from other individuals in the group to that individual, the state quantity being predicted based on past communication volume between that individual and the other individuals. For each of the temporarily defined plurality of personnel, based on the current state value of the personnel and the state propagation amount of the personnel, a future state value indicating the future state of the personnel is predicted. Based on the predicted future state value of each of the plurality of individuals, a future group state value indicating the future state of the group is predicted, and When the predicted future group state value meets the predetermined conditions for a good state, the temporarily defined multiple individuals will be identified as members of the group.

8. A state prediction method executed by a computer, characterized in that it comprises: Obtain the current state value of the target individual in the indicated group; The state propagation quantity is obtained as the state quantity that is propagated from other persons to the target individual through communication between the target individual and other persons in the group. Based on the current state value of the target individual and the state propagation amount obtained, predict the future state value that indicates the future state of the target individual; The information obtained includes the current state value indicating the current state of the other persons, the current amount of communication between the target individual and the other persons, and the degree of influence of the other persons on the state of the target individual. as well as Based on the acquired information, the state propagation amount from the other personnel to the target individual is derived.

Images