JP6428748B2 - Driving support system - Google Patents

Description

  The present invention relates to a driving support system.

  Humans receive emotions (also called sensibilities or emotions) that have received visual and auditory stimuli. There are Patent Document 1 and Patent Document 2 for evaluating visual sensitivity. Patent Document 1 displays an evaluation result of an image impression using a sensitivity term representing sensitivity. Japanese Patent Application Laid-Open No. 2004-228561 performs image evaluation reflecting a photographer's preference such as exposure level.

JP-A-9-16797 JP 2006-254107 A

  By the way, various driving assistances are provided in vehicles. As this driving support, it is conceivable to guide the driver's emotion toward the target emotional state. For example, when the driver has an unpleasant feeling, it is conceivable to apply a visual or auditory stimulus in a direction in which the driver's feeling becomes pleasant. In this case, it is difficult to appropriately select what is suitable as a sensory stimulus suitable for achieving the target emotional state. Specifically, when a certain sensory stimulus is applied, even if the driver's emotion can be brought close to a target emotional state such as a pleasant state, the driver becomes inactive on the contrary. May cause adverse effects.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a driving support system capable of appropriately applying sensory stimulation to be directed toward a target emotional state. .

In order to achieve the above object, the following solution is adopted in the present invention. That is, as described in claim 1,
A driving support system that guides a driver of a vehicle toward a target emotional state,
An information acquisition means for acquiring information about an evaluation object related to human senses;
A feature calculation unit configured to simulate hierarchical sensory information processing in the human brain, and to calculate feature data for each layer of the information acquired by the information acquisition unit;
Sensory statistic calculating means for calculating sensory statistic for the feature data calculated by the feature calculating means;
Sensory statistic determining means for determining a specific sensory statistic highly related to emotion among sensory statistics calculated by the sensory statistic calculating means;
Emotion information detection means for detecting emotion information related to the driver's emotion;
From the emotion information detected by the emotion information detection means, emotion determination means for determining the feeling of the driver,
Sensory statistic selection means for selecting a sensory statistic suitable for guiding the driver's emotion determined by the emotion determination means among the specific sensory statistic toward the target emotion state;
Providing a sensory stimulus so that the sensory statistic selected by the sensory statistic selecting unit is changed, and supporting means for supporting the driver's emotion toward the target emotion state;
It is supposed to be equipped with.

  According to the above solution method, it is possible to appropriately select a parameter for guiding to the target emotion state using the sensory statistic and appropriately guide to the target emotion state. In particular, sensory statistics are based on feature data obtained by simulating hierarchical sensory information processing in the human brain, and therefore appropriately reflect human emotions, which leads to the target emotional state. This is extremely suitable for preventing or suppressing a situation in which the driver's emotions are unnecessarily (inadvertently) inhibited when guiding to the target emotional state.

A preferred mode based on the above solution is as described in claim 2 and the following. That is,
The target emotional state is set in a predetermined area having two axes of pleasant / unpleasant and active / inactive as parameters (corresponding to claim 2). In this case, the region of the target emotional state as a preferable emotional state can be appropriately set by using two axes, which are general pleasant / unpleasant and active / inactive, as the driver's desired emotion.

The information about the evaluation target is at least one of information about vision and information about hearing,
Giving a sensory stimulus by the support means is at least one of a visual stimulus and an auditory stimulus,
(Corresponding to claim 3). In this case, it is possible to effectively guide to the target emotional state by using vision or hearing that has a great influence on human emotions.

An evaluation model is set that determines the evaluation value of emotion using sensory statistics as a parameter,
The sensory statistic selecting means determines, as the specific sensory statistic, a sensory statistic included in the evaluation model corresponding to an emotion for guiding toward the target emotional state.
(Corresponding to claim 4). In this case, a specific sensory statistic can be appropriately and easily selected using the evaluation model.

  A learning correction means for learning correction of the evaluation model is further provided (corresponding to claim 5). In this case, it is preferable to further optimize the evaluation model and more appropriately guide to the target emotion state.

  The sensory statistics are assumed to be autocorrelation and cross-correlation of the feature data (corresponding to claim 6). In this case, a specific sensory statistic is provided. In particular, auto-correlation and cross-correlation accurately indicate the relevance of a large number of feature data, and thus more accurately and precisely show human sensitivity than feature data itself. As a result, sensory statistics This is extremely preferable in appropriately supporting the guidance to the target emotional state using the quantity.

  According to the present invention, it is possible to appropriately apply a sensory stimulus that is directed toward a target emotion state.

The figure which shows the example of a control system for performing driving assistance. The figure which shows the example of a setting of the area division which uses two axis | shafts of pleasant / unpleasant and active / inactive as a parameter. The block diagram which shows the example of a control system which performs the setting of an evaluation model. The figure which simulates the hierarchical auditory information processing in a human brain, and shows the calculation of the autocorrelation and the cross correlation of the feature data of each hierarchy, and each feature data. The figure for demonstrating the method of determining an evaluation model. Explanatory drawing which shows the example of determination of a cross correlation. The flowchart which shows the example of control of this invention. The flowchart which shows the example of control of this invention.

  First, an overview of driving assistance will be described with reference to FIG. In FIG. 1, S1 is a biosensor that detects a driver's electrocardiogram, heartbeat, and the like. A driver's (driver) emotion is detected by the driver emotion detection unit S2 that receives the output from the biosensor S1. Thereafter, the driver emotion determination unit S3 determines the driver's emotion. The determination of the emotion by the driver emotion determination unit S3 is where the driver's emotion is located in the emotion map shown in FIG. In order to detect the driver's emotion, a camera that captures the driver's face can be used instead of or in addition to the biosensor S1.

  The emotion map of FIG. 2 is divided into nine regions vertically and horizontally, with two axes of pleasant / unpleasant and active / inactive as parameters (the enhancement of each region is indicated by a broken line). Specifically, for the active / inactive parameters, A1 is set in the inactive direction and A3 is set in the active direction, with A2 being an intermediate value. That is, A1 and below are in an inactive state, A3 and above are in an active state, and an intermediate state is between A1 and A3.

  As for the comfort / discomfort parameters, B2 is set as an intermediate value, B1 is set in the discomfort direction, and B3 is set in the comfort direction. That is, B1 and below are in an uncomfortable state, B3 and above are in a comfortable state, and a state between B1 and B3 is an intermediate state. In the embodiment, among the nine regions, the region where the pleasant / unpleasant level is larger than B1 and the active / inactive level is between A1 and A3 (region 2-2). And the region 3-2) are set as a preferable emotion state set as the target emotion state.

  In FIG. 1, S5 is a sensory information control unit, which controls a display unit S6 related to visual stimuli and a speaker unit S7 related to auditory stimuli. In particular, the sensory information control unit S5 controls the display unit S6 and the speaker unit S7 so as to change a parameter related to stimulus application set by the sensory information control parameter setting unit S4. And this parameter is selected using the evaluation model selected by evaluation model selection part S8 mentioned later. The selection of the evaluation model by the sensory information control parameter setting unit S4 and the evaluation model selection unit S8 will be described later including the setting of the evaluation model.

  Next, with reference to FIG. 4, calculation of feature data in each layer based on processing simulating auditory information processing in the human brain, and moment and cross-correlation for each feature data will be described. That is, starting from the input of sound data to the cochlea (basement membrane), the brainstem thalamic process is finally performed, and the first layer process, the second layer process, Hierarchical processing is performed, and FIG. 2 shows such processing in a simulated manner. It should be noted that the numbers in FIG. 2 surrounded by circles are displayed with circles 1, 2, etc. (for example, circle 1 means the number 1 surrounded by circles).

  In FIG. 4, sound data is input to the input unit 1 configured by a microphone. The input sound data is classified by frequency with respect to the sound data by the Fourier transform unit 2 (first layer processing). Although FIG. 2 shows the case where it divides into two, high frequency and low frequency, for simplification, it is actually divided into about 20-30. The processing in the first layer corresponds to the processing in the cochlear basement membrane.

  In the processing of the second layer, the feature data (segmented frequencies) obtained in the first layer are subjected to non-linearly compressed envelope extraction processing (extraction of feature data indicated by ○ 1, ○ 2), and thereafter The characteristic data indicated by ◯ 3 and ◯ 4 are acquired by the non-linear compression processing (corresponding to processing in which strong sounds are weakened and weak sounds are strengthened). The processing in the second layer corresponds to processing based on the amplitude characteristics of the cochlear basement membrane and the brainstem thalamus.

  In the processing of the third layer, the Fourier transform units 3 and 4 further classify the feature data ○ 3 and ○ 4 in the second layer according to the frequency (corresponding to the modulation frequency for the feature data in the second layer). ). The feature data calculated in the third hierarchy is indicated by ○ 5 to ○ 8. The processing in the third hierarchy corresponds to the processing in the brainstem thalamus.

For each of the above-described feature data ○ 1 to ○ 8, the moment and the cross correlation are calculated.
“Moment” is defined by any one of average, variance, kurtosis, and skewness as basic statistics. When the square value of “Moment” is used depending on the hierarchy, the term “Power” is used separately. In FIG. 4, what is indicated by M2 or M3 is Moment or Power, and what is indicated by C0, C1, and C2 is cross-correlation. More specifically, M2 is a moment for ○ 3 and ○ 4 that are feature data in the second hierarchy, and M3 is a power for ○ 5 to ○ 8 that is feature data in the third hierarchy. is there.

  C0 is a cross-correlation between ◯ 3 and ◯ 4, which are feature data of the second hierarchy. C1 indicates a cross-correlation between ○ 5 and ○ 7 and a cross-correlation between ○ 6 and ○ 8, and is a cross-correlation between different frequencies divided by the processing of the first layer. C2 indicates the cross-correlation between ○ 5 and ○ 6, and the cross-correlation between ○ 7 and ○ 8, and the cross-correlation between the same frequency when divided by the processing of the first layer. Become.

  Next, an example of setting an evaluation model will be described with reference to FIG. In the following description, it is assumed that the sensitivity to be evaluated is “uncomfortable” as an example. That is, for example, a case where an unpleasant level (pleasure level) given to a listener by engine sound is evaluated in a plurality of stages will be described as an example.

  First, in FIG. 5, what is indicated as G1, G2,... Gn indicates a large number of different (for example, 500 to 1000) sound data to be evaluated for auditory sensitivity. Auditory sensitivity evaluation values (e.g., evaluation values at a plurality of stages with respect to the uncomfortable level) are associated with the large number of sound data. Note that the discomfort level evaluation value is an average value obtained by averaging evaluation values by a plurality of experts who perform evaluation. In FIG. 3, thick vertical lines are lines that partition individual moments and cross-correlations shown in FIG. And the thin vertical line within the range enclosed by the thick vertical line partitions each data (correlation coefficient) which shows Moment and a correlation level.

  In FIG. 5, “2nd” indicates that the feature data is in the second layer, and “3rd” indicates that the feature data is in the third layer. In addition, “Across” means cross-correlation between different frequencies among the frequencies classified in the first layer, and “Within” means frequencies of the frequencies classified in the first layer. This means cross-correlation between the same frequencies.

  In FIG. 3, six types of correlation are set. That is, in order from the left side to the right side, “2nd Moment”, “2nd Across”, “3rd Power,“ 3rd Across ”,“ 3rd Within (Real) ”, and“ 3rd Within (Image) ”. Note that “3rd Within” is subjected to Hilbert transform, and its real part is “3rd Within (Real)” and its imaginary part is “3rd Within (Image)”.

  Among the moments and cross-correlations shown in FIG. 4, in particular, “2nd Across” (cross-correlation between different frequencies divided in the first layer for the feature data in the second layer) indicates the uniformity of amplitude fluctuation. It becomes the parameter shown. Further, “3rd Power” (the square value of the moment for the feature data in the third hierarchy) is a parameter indicating the amount of amplitude fluctuation (power or energy). Furthermore, “3rd Within” (cross-correlation between the same frequencies sorted in the first layer for the feature data in the third layer) indicates an amplitude variation pattern (that is, how the sound changes).

  Here, with reference to FIG. 6, for example, a method of determining the correlation value of “2nd Across” will be described. In FIG. 6, the vertical axis and the horizontal axis indicate frequencies separated in the first layer (frequency is, for example, 52 Hz to 8844 Hz). Correlation values (correlation coefficients) are acquired in a large number of unit areas divided vertically and horizontally. In FIG. 6, the double circle indicates that the correlation value is extremely high, the single circle indicates that the correlation value is lower than that of the double circle, but the correlation value is greater than or equal to a predetermined value, and the cross indicates that the correlation value is zero And white indicates no data.

  In FIG. 6, an area in which unit areas having a correlation value equal to or greater than a predetermined value are assembled to form a group is selected. Is done. Note that the final correlation value is similarly determined for other cross-correlations.

  In FIG. 5 again, the moment and the cross-correlation are calculated for each sound data. One sound data has many values for the moment and the cross-correlation. For each Moment and each cross-correlation, principal component analysis (for example, five dimensions) is performed, and data extraction (data narrowing down) is performed as shown by column A in FIG. Further, as shown by the B column in FIG. 5, those having a certain correlation or more with respect to the subjective evaluation value (for example, a parameter that greatly affects the evaluation of the discomfort level) are selected from the A column.

  Based on the parameters shown in the column A acquired by the principal component analysis, the expression (1) indicating the evaluation model is determined as shown in Equation 1.

  (1) In the formula, Y is an evaluation value for each sound data, a0 to am, b1 to bm, c (suffix is omitted here) is a coefficient (constant), and X1 to Xm are These are the moment (or power) and cross-correlation values in the column A shown in FIG. Initially, the above coefficients are unknown. By applying the subjective evaluation value Y and the moments and cross-correlation values X1 to Xm for the sound data G1 to Gn to the equation (1), each coefficient is determined by a regression method. In equation (1), which is a regression model, each coefficient is associated with a specific one of a number of moments and cross-correlation values.

  Equation (1) is determined so as to include the moment and the cross-correlation value indicated by the extracted column A. From this equation (1), leaving only the parameters that have a great influence on the evaluation of, for example, the discomfort level as shown in column B of FIG. 3 (by deleting parameters that do not affect the evaluation), The model formula is determined as the following formula (2). That is, in the equation (2), the suffix n is smaller than the suffix m in the equation (1).

  Here, the above equation (1) includes a second-order term and an interaction term (the third term on the right side excluding the initial value a0) for all the first-order terms. Actually, from the viewpoint of simplifying the evaluation model as much as possible, only a specific first-order term that has a predetermined influence or more on the subjective evaluation has a second-order term and an interaction term. It is preferable that the state after such processing is performed is Equation (2). Equation (2) is rewritten so that the suffix of each coefficient is continuous after deleting the coefficients for the unnecessary parameters from Equation (1) (Equations (1) and (2)). In the formula, the same suffix coefficient does not always indicate the same content).

  When an uncomfortable level evaluation of certain sound data is performed, the moment (Power) and the cross-correlation value are calculated for the certain sound data as shown in FIG. 5 and the calculation result is applied to the above equation (2). An evaluation value for a certain sound data is calculated with a numerical value (also visualized by quantification). It should be noted that the validity of the evaluation model in the equation (2) is expressed by a plurality of sound data other than the sound data G1 to Gn shown in FIG. 5 (Moments and cross-correlation values calculated from them) and a subjective evaluation value corresponding thereto. You can also verify using.

In the formula (2), for example, when the vehicle engine sound is evaluated as pleasant / uncomfortable, the coefficient to be used can be extremely reduced as a0, a1, a2, a3, such as n = 3. Other than the coefficient a0 indicating the basic value, the coefficient a1 is for "2nd Acros", the coefficient a2 is for "3rd Power", and the coefficient a3 is for "3rd Wihin". The evaluation value obtained by using the evaluation model corresponding to the expression (2) defined by the coefficients a0, a1, a2, and a3 is extremely R ² = 0.97 in relation to the subjective evaluation value. It was confirmed to show a high correlation.

  Next, an overall control system example for setting an evaluation model for performing sensitivity evaluation will be described with reference to FIG. In FIG. 3, reference numeral 11 denotes an input unit for sound data to be evaluated for auditory sensibility, and is constituted by a microphone, for example. As described above, the feature data is calculated by the sound feature calculator 12 for the input sound data, and the autocorrelation and the cross-correlation are calculated by the auditory statistic calculator 13 based on the feature data.

  On the other hand, a plurality of evaluation models for auditory sensitivity are set. The evaluation model A shown in FIG. 3 evaluates the level of pleasantness / discomfort, for example, for engine sound evaluation. The evaluation model B is for evaluating the active / inactive level in the engine sound evaluation, for example. In addition, there are many evaluation models for road noise evaluation and the like.

  In each evaluation model, parameters (evaluation axes) corresponding to the type of sound source to be evaluated and the type of sensitivity are set by selection in the evaluation parameter selection unit P. For example, in the evaluation of the discomfort level for the engine sound, the irregularity of the sound and the degree of difference in sound intensity are selected as parameters (moment and cross-correlation are performed so as to detect the parameters).

  A plurality of evaluation models can be set according to the difference in the types of sensibilities to be evaluated. Various models other than the above can be set as the evaluation model. For example, an evaluation model for a sound when closing a vehicle opening / closing member (door, bonnet, etc.), an evaluation model for a brake sound, and an environmental sound around a building. Various models such as an evaluation model can be set. Although it is possible to set one evaluation model that integrates the type of sound source and the type of sensitivity to be evaluated, it is possible to set an evaluation model for each type of sound source and for each type of sensitivity to be evaluated. Therefore, it is preferable for more appropriate sensitivity evaluation.

  In FIG. 5, personal information (for example, sex, age, occupation, etc.) is input in the personal information input unit 21, and the input personal information is pre-processed in the personal information pre-processing unit 22 (for example, age is classified into a plurality of levels). Is recorded (stored) in the personal information recording unit 23.

  The biological information input by the biological sensor input unit 24 and the subjective sensitivity evaluation value input by the subjective evaluation unit 25 are input to the actual measured sensitivity estimation unit 26 to estimate the actual measured sensitivity value. The actual measured sensitivity value is recorded (stored) in the actual measured sensitivity value recording unit 27.

  Among the sound sources input by the sound source information input unit 28, the evaluation target is specified by the evaluation target specifying unit 29. Then, the evaluation target (sound data) specified by the evaluation target specifying unit is recorded (stored) in the evaluation target recording unit 30. The calculation result in the auditory statistic calculation unit 13 is recorded (stored) in the auditory statistic recording unit 31.

  Each of the recording units 23, 27, 30, and 31 described above constitutes the database unit D. An evaluation model is set based on the data in the database part D. Specifically, the data in the database unit D is input to the evaluation parameter selection unit P, and an appropriate parameter (evaluation axis) is selected according to the type of sensitivity to be evaluated (column B in FIG. 5). The evaluation model is set according to the selected parameter (the evaluation model shown in the equation (2) is set).

  The recorded content in the recording unit 30 is used as a large number of sound data G1 to Gn shown in FIG. 5, and the recorded content in the recording unit 27 is used as a subjective evaluation value for each sound data G1 to Gn. . The personal information input unit 21 is for correcting the evaluation value according to personal information (for example, gender or age group), and may be omitted. The recording unit 31 is for learning correction of the evaluation model when a new sensibility evaluation is performed, and may be omitted when the learning correction is not performed.

  In order to make the evaluation model more appropriate, when the subjective evaluation value differs from the evaluation value based on the evaluation model for the sound data acquired by the evaluation target sound source acquisition unit 11, the acquired sound data is converted to the sound data shown in FIG. The evaluation model (equation (2) corresponding to this) can be learned and corrected by including in the data and the subjective evaluation value by the evaluation specialist.

  In the description of FIGS. 3 to 6, the description has been given focusing on auditory sensibility, but an evaluation model can be similarly set for visual sensibility. That is, when simulating human visual information processing, classification by frequency (that is, resolution) is performed in the first layer, direction selection processing is performed in the second layer, and direction signal strength processing is performed in the third layer. From FIG. 5 using the autocorrelation (correlation value indicating the value based on the moment) and cross-correlation (correlation value indicating) for each feature data of the third layer An evaluation model for visual sensibility is set by the same method as described above.

  Referring again to FIG. 1, the evaluation model selection unit S8 displays an evaluation model suitable for guiding the emotion of the driver of the vehicle toward the target emotion state (region 2-2 or 3-2) shown in FIG. Select from the multiple evaluation models shown in 3. For example, when it is determined that the current emotional state of the driver is, for example, the region 2-1 in FIG. 2, an evaluation model suitable for guiding the emotion in the inactive direction is selected ( Selection of evaluation model to evaluate activity / inactivity).

  The sensory information control parameter setting unit S4 selects parameters (X1, X2, etc. in the equation (2)) indicating autocorrelation or cross-correlation in the evaluation model selected as described above. In other words, the autocorrelation or cross-correlation parameter in the selected evaluation model has a great influence on the driver's feelings in the active / inactive direction, but has a great influence on the pleasant / unpleasant direction. It will be difficult to give.

  The sensory information control unit S5 controls the display unit S6 and the speaker unit S7 so as to change the parameter selected as described above. It can be known from the evaluation value obtained from the evaluation model how much the change direction and change amount of the parameter affect the driver's emotion. As a result, the sensory information control unit S5 can control the display unit S6 and the speaker unit S7 with an appropriate control direction (for example, an increasing / decreasing direction of sound volume) and an appropriate control amount (for example, sound pressure). This is also preferable in preventing control overshoot for the display unit S6 and the speaker unit S7 (for example, transition from the emotional state in the region 2-1 to the region 2-3 beyond the region 2-2). Prevention).

  FIG. 7 and FIG. 8 are flowcharts showing control examples for guiding the driver's emotion toward the target emotion state, which will be described below. In the following description, Q indicates a step.

  First, after data is input in Q1, the current emotional state of the driver, that is, the degree of pleasure / discomfort (level) and the degree of activity / inactivity (level) are determined in Q2. Thereafter, in Q3, it is determined whether or not the activity is equal to or greater than a predetermined value A3. If the determination in Q3 is YES, it is determined in Q4 whether the comfort level is B1 or less. When the determination of Q4 is YES, the driver's current emotion is in the area 1-1 in FIG. At this time, in Q5, priority is given to the pleasant axis, and support for lowering discomfort (support for increasing comfort) is executed. In addition, support for lowering activity (increasing inactivity) may be performed.

  When the determination in Q4 is NO, the driver's current emotion is in the region 2-1 or 3-1 in FIG. At this time, in Q6, priority is given to the active axis, and support for lowering the activity is executed.

  If the determination in Q3 is NO, it is determined in Q7 whether or not the activity is A1 or less. When the determination in Q7 is YES, it is determined in Q8 whether or not the comfort level is B1 or less. When the determination in Q8 is YES, the current emotional state of the driver is in the region 1-3. At this time, in Q9, priority is given to the pleasant axis, and support for lowering discomfort (support for increasing comfort) is executed. In addition, support for increasing the activity (decreasing the inactivity) may be performed.

  When the determination in Q8 is NO, the current emotional state of the driver is in the region 2-3 or the region 3-3. At this time, in Q10, priority is given to the active axis, and support for increasing the activity (decreasing the inactivity) is executed.

  If the determination in Q7 is NO, it is determined in Q11 whether or not the comfort level is B1 or less. When the determination in Q11 is YES, the current emotional state of the driver is in the region 1-2. At this time, in Q12, priority is given to the pleasant axis, and support for reducing discomfort (increasing comfort) is executed.

  When the determination in Q11 is NO, the current emotional state of the driver is in the region 2-2 or the region 3-2 and is the same as the target emotional state. Therefore, at this time, the process shifts to Q13 and it is only determined that the emotional state is appropriate for driving, and support for emotional transition is not executed.

  In the control execution for emotion transition in Q5, Q6, Q9, Q10, and Q12 described above, as described above, an evaluation model suitable for that is selected, and the parameters included in this evaluation model change. Done. Hereinafter, this point will be described with reference to the flowchart of FIG.

  First, in Q21 in FIG. 8, an evaluation model corresponding to the direction in which the sensitivity is shifted is selected (selection in S8 in FIG. 1). Next, in Q22, parameters in the selected evaluation model are extracted. After that, in Q23, support for guiding to the target emotion state is executed with the control content that changes the selected parameter.

  Next, a specific example of auditory or visual support for guiding to a target emotion state will be described. Auditory can provide the following support. As support for lowering unpleasantness, lowering the volume of the alarm sound, raising the threshold value for generating an alarm (changing in a direction that is difficult to alarm), selecting a pleasant sound as a background from the speaker unit S7 It can flow. Further, as support for improving pleasure, a pleasant sound can be selected and the background or engine sound adjusted to be pleasant can be played from the speaker unit S7.

  In addition, as a support for lowering the degree of activity on hearing, music with a calming effect can be played from the speaker unit S7 (the same music can also be calmed down with a reduced rhythm). Further, as support for raising the activity level, music having an uplifting effect can be played from the speaker unit S7 (the same music can also be raised by raising the rhythm).

  The following assistance can be provided visually. As support for reducing discomfort, the visibility of uncomfortable visual fields can be reduced. That is, by adopting a variable window frame structure to narrow the field of view, to reduce the transmittance of the window glass, to narrow the orientation range of the headlight, for example, it is possible to make it difficult to see a complicated landscape on the road shoulder. In addition, unpleasant visual features can be weakened (for example, the reflectance of side mirrors and rearview mirrors is reduced, or the illuminance of vehicle interior lighting is reduced). As support for improving pleasure, for example, a navigation device can be used to provide route guidance for viewing a desired landscape.

As support for lowering the degree of activity, for example, it can be changed to vehicle interior lighting with a calming effect (it can be cold-colored lighting or lower illuminance). In addition, as support for increasing the activity level, it can be changed to vehicle interior lighting having an uplifting effect (warm-colored lighting or illuminance is increased).
To change the visual stimulus, use projection mapping to partially limit the field of view, partially change the pattern atmosphere and illuminance in the passenger compartment, project images in front of the driver, etc. You can also.

  The target emotional state can be induced not only by hearing and vision but also by vehicle control or other control. For example, in order to improve comfort, when a favorite landscape (environment) is obtained, the driving mode can be changed to a direction in which the behavior of the vehicle becomes sensitive (for example, improvement in accelerator sensitivity and sensitivity of the steering wheel). Similarly, in order to entertain the scenery, the vehicle speed is reduced, the vehicle speed is controlled (depending on the distance between the following vehicles) and the pumping brake is automatically applied. In addition, it is possible to urge the user to enjoy the scenery by automatically turning on the hazard lamps, and to propose a nearby parking place by the navigation device. In addition, the olfaction can be used to increase (decrease) the degree of pleasure and activity (preparing a plurality of types of scents and supplying the scent selected according to the situation into the passenger compartment). Furthermore, it is also possible to induce the target emotional state by tactile sense, for example, changing the hardness of the steering handle or the hardness of the seat. Of course, any two or more of the various methods described above can be appropriately combined to induce the target emotional state.

  Although the embodiments have been described above, the present invention is not limited to the embodiments, and appropriate modifications can be made within the scope of the claims. In the vehicle, in order to simplify the equipment as much as possible, only the control system shown in FIG. 1 and each part 11 to 13 of FIG. 3 are mounted on the vehicle, and the other control system (FIG. The portion shown in FIG. 4) can be configured in an information center that communicates with the vehicle. In addition, the evaluation model acquired at the information center can be stored in the storage unit (database) of the vehicle by communication as appropriate. It is also possible to set a plurality of evaluation models divided by age, gender, etc. (Evaluation models corresponding to young females, male seniors, etc. are individually set). The evaluation model may be set with at least any two types of moment, autocorrelation, and cross-correlation for feature data (for example, an evaluation model for hearing is set based on moment and cross-correlation On the other hand, a visual evaluation model is set by autocorrelation and cross-correlation, or by moment and cross-correlation, or by moment, auto-correlation and cross-correlation). In particular, for vision, variations in brightness, symmetry, uniformity, and the like can be acquired by autocorrelation. Of course, the object of the invention is not limited to what is explicitly stated, but also implicitly includes providing what is substantially preferred or expressed as an advantage.

  The present invention is preferable in guiding the vehicle driver to a desirable target emotional state.

S1: Living body sensor S2: Driver emotion detection unit S3: Driver emotion determination unit S4: Emotion upper control parameter setting unit S5: Sensory information control unit S6: Display unit (giving visual stimulus)
S7: Speaker part (stimulation on hearing)

Claims (6)

A driving support system that guides a driver of a vehicle toward a target emotional state,
An information acquisition means for acquiring information about an evaluation object related to human senses;
A feature calculation unit configured to simulate hierarchical sensory information processing in the human brain, and to calculate feature data for each layer of the information acquired by the information acquisition unit;
Sensory statistic calculating means for calculating sensory statistic for the feature data calculated by the feature calculating means;
Sensory statistic determining means for determining a specific sensory statistic highly related to emotion among sensory statistics calculated by the sensory statistic calculating means;
Emotion information detection means for detecting emotion information related to the driver's emotion;
From the emotion information detected by the emotion information detection means, emotion determination means for determining the feeling of the driver,
Sensory statistic selection means for selecting a sensory statistic suitable for guiding the driver's emotion determined by the emotion determination means among the specific sensory statistic toward the target emotion state;
Providing a sensory stimulus so that the sensory statistic selected by the sensory statistic selecting unit is changed, and supporting means for supporting the driver's emotion toward the target emotion state;
A driving support system characterized by comprising: In claim 1,
The driving support system, wherein the target emotional state is set in a predetermined region having two axes of pleasant / unpleasant and active / inactive as parameters. In claim 1 or claim 2 ,
The information about the evaluation target is at least one of information about vision and information about hearing,
Giving a sensory stimulus by the support means is at least one of a visual stimulus and an auditory stimulus,
A driving support system characterized by that.
In any one of Claims 1 thru | or 3,
An evaluation model is set that determines the evaluation value of emotion using sensory statistics as a parameter,
The sensory statistic selecting means determines, as the specific sensory statistic, a sensory statistic included in the evaluation model corresponding to an emotion for guiding toward the target emotional state.
A driving support system characterized by that. In claim 4,
A driving support system, further comprising learning correction means for learning correction of the evaluation model. In any one of Claims 1 thru | or 5,
The driving assistance system, wherein the sensory statistic is an autocorrelation and a cross-correlation of the feature data.

Images