JP6727465B1 - Driver status determination device and driver status determination method - Google Patents

Abstract

The driver state determination device (100) includes a biological signal analysis unit (130) that calculates a characteristic amount used to estimate the driver's state from a biological signal of the driver of the vehicle (10), and a characteristic amount of the biological signal. Then, a driver state calculating unit (140) that calculates an estimated value of the driver's state and a driver that classifies the driver's state by comparing the estimated value of the driver's state with a predetermined threshold value. A state classifying unit (150) and an information presenting unit (160) that presents information to the driver according to the result of classifying the state of the driver. In the recording unit (141), has the driver's eyes closed degree, the route information of the vehicle (10), the in-vehicle environment information of the vehicle (10), and the operation of the vehicle (10) matched with the content of the information presentation performed? At least one of the information on whether or not and the estimated value of the driver's state are recorded in synchronization with each other, and the threshold updating unit (142) is based on the information recorded in the recording unit (141). Then, the threshold for classifying the driver's state is updated.

Description

The present invention relates to a driver state determination device that determines a driver's state from output data of a sensor mounted on a vehicle.

There are still many traffic accidents caused by the involuntary driving of vehicles. Therefore, based on the biological signal of the driver acquired by the sensor, for example, to detect an abnormal state such as a state in which the driver's fatigue accumulates or a state in which the driver cannot concentrate on driving, which is useful for avoiding aimless driving. Research is being done. In addition, since the driver's state (for example, drowsiness, fatigue, anxiety level, and concentration) has individuality (individual difference), the threshold value that is the criterion for determining the driver's state is A method of setting a different value for each driver has been studied (for example, Patent Document 1 below).

International Publication No. 2006/054542

According to the technique disclosed in Patent Document 1, by setting different thresholds for each driver in the information indicating the driver's state such as the frequency with which the driver changes his/her posture and the amount of change, the blood flow of the driver, I am trying to deal with sex. However, with regard to information such as the frequency of change in posture, the amount of variation, and blood flow, there are many events that cause changes in the value, and it takes time for the driver's state change to appear as posture and blood flow. Therefore, it is difficult to judge the individuality of the driver from such information. Further, Patent Document 1 does not describe a specific method for updating the threshold value.

The present invention has been made to solve the above problems, and provides a driver state determination device capable of estimating the state of a driver in consideration of the individuality of the driver. The purpose is to

The driver state determination device according to the present invention is a biometric sensor unit that acquires a biological signal of a driver of a vehicle, and a feature amount used for estimating the driver state from the biological signal by analyzing the biological signal. , A driver state calculation unit that calculates an estimated value of the state of the driver based on the feature amount of the biological signal, and an estimated value of the state of the driver are predetermined. A driver state classifying unit that classifies the state of the driver by comparing with a threshold value, an information presenting unit that presents information to the driver according to the classification result of the state of the driver, and the driver Of at least one of the degree of eye closure, route information of the vehicle, in-vehicle environment information of the vehicle, and information indicating whether or not the vehicle operation matching the content of the information presentation has been performed, and the driver's A recording unit that records the estimated value of the state in synchronization with each other, and the degree of eye closure of the driver, the route information of the vehicle, the in-vehicle environment information of the vehicle, and the information presentation that are recorded in the recording unit. Based on the regression coefficient of the regression line and the size of the intercept based on the correlation between the estimated value of the driver's state and at least one of the information as to whether or not the operation of the vehicle conforming to the content of comprises a threshold update unit for updating the threshold, wherein the biometric sensor section includes a plurality of biometric sensors having portions each measurement area do not overlap each other, an attitude sensor for acquiring the attitude information representing the attitude of the driver A biosensor selection unit that selects a biosensor in which the driver is present in the measurement range from the plurality of biosensors based on the posture information, and outputs a biosignal acquired by the selected biosensor, It is equipped with.

According to the driver state determination device according to the present invention, from the information of the driver's state and the degree of eye closure in the past, to determine the individuality of how to feel the driver's state, based on the determination result, The threshold for classifying states is updated. The state of the driver can be classified in consideration of the individuality of the driver, and the accuracy of determining the state of the driver can be improved.

Objects, features, aspects, and advantages of the present invention will become more apparent by the following detailed description and the accompanying drawings.

It is a figure which shows the structure of the driver condition determination apparatus which concerns on Embodiment 1. It is a figure which shows the structure of a biological sensor part. It is a figure which shows the structure of a vehicle behavior sensor part. It is a figure which shows the structure of a biological signal analysis part. It is a figure which shows the example of the filter bank which a filter bank production|generation part produces|generates. It is a figure which shows the example of the amplitude spectrum before a filter bank is applied. It is a figure which shows the example of the amplitude spectrum after a filter bank is applied. It is a figure which shows the example of the subjective evaluation in the data collection for model learning. It is a figure which shows the example of the calculation result of the driver state using the learned model. It is a figure which shows the example of the class classification of a driver's state. It is a figure for demonstrating the example of the determination method of the initial value of the threshold value which classifies a driver|operator's state. It is a figure which shows the example of the regression line which shows the estimated value of drowsiness and the degree of eye closure. It is a figure which shows the example of the threshold update criteria in a threshold update part. It is a figure which shows the structure of the driver condition determination apparatus which concerns on Embodiment 2. It is a figure which shows the structure of the driver condition determination apparatus which concerns on Embodiment 3. It is a figure which shows the structure of the driver condition determination apparatus which concerns on Embodiment 4. It is a figure which shows the structure of the driver condition determination apparatus which concerns on Embodiment 5. 28 is a flowchart showing an example of a procedure of selecting data as an update reference in the threshold update unit of the fifth embodiment. It is a figure showing the example of hardware constitutions of a driver condition judging device. It is a figure showing the example of hardware constitutions of a driver condition judging device.

<Embodiment 1>
FIG. 1 is a diagram showing a configuration of a driver state determination device 100 according to the first embodiment of the present invention. As shown in FIG. 1, the driver state determination device 100 is mounted on a vehicle 10, and includes a biometric sensor unit 110, a vehicle behavior sensor unit 120, a biometric signal analysis unit 130, a driver state calculation unit 140, a recording unit 141, and The threshold update unit 142, the driver state classification unit 150, the information presentation unit 160, and the eye closure degree calculation unit 171 are provided.

The biometric sensor unit 110 acquires and outputs a biometric signal of the driver of the vehicle 10. The biological signal acquired by the biological sensor unit 110 is a time-series signal such as a heartbeat or a pulse wave. Examples of the biometric sensor unit 110 include an electrocardiographic sensor that electrically captures the contraction of the heart, and a photoelectric pulse wave sensor that captures the contraction of the heart by detecting the volume change of the blood vessel from the change of hemoglobin accompanying the pulsation. These are sensors that are generally used in contact with or in close proximity to the driver's skin, but are used without contacting the driver's skin, such as sensors that measure chest mutations using radio waves. It may be a sensor. As described above, various types can be considered for the biometric sensor unit 110, but any type may be used as long as it can output the biometric signal of the driver.

The biometric sensor unit 110 may be a single sensor, but in the present embodiment, the biometric sensor unit 110 configured as shown in FIG. 2 is used. As shown in FIG. 2, the biometric sensor unit 110 includes a first biometric sensor 111, a second biometric sensor 112, a posture sensor 113, and a biometric sensor selection unit 114.

The first biosensor 111 and the second biosensor 112 are both sensors that acquire and output a biosignal of the driver. However, the first biosensor 111 and the second biosensor 112 have different measurement areas.

The posture sensor 113 acquires posture information indicating the posture of the driver. As a method for the attitude sensor 113 to acquire the attitude information, for example, there is a method using a pressure sensor arranged on the seat surface or the back surface of the driver's seat, a motion capture using a camera, or the like. The method does not matter as long as it can be acquired.

The biometric sensor selection unit 114 determines which of the first biometric sensor 111 and the second biometric sensor 112 the driver is in based on the attitude information of the driver acquired by the attitude sensor 113. Then, the biometric sensor selection unit 114 selects and outputs the biometric signal output by one of the first biometric sensor 111 and the second biometric sensor 112 in which the driver is present within the measurement range.

As described above, the biometric sensor unit 110 of FIG. 2 realizes a wide measurement area by selectively using the first biometric sensor 111 and the second biometric sensor 112 having different measurement areas according to the posture of the driver. doing. Thereby, the biometric sensor unit 110 can acquire the biometric signal of the driver with high accuracy. Here, the example in which the biometric sensor unit 110 has two biometric sensors is shown, but the number of biometric sensors may be three or more.

The vehicle behavior sensor unit 120 observes the behavior of the vehicle 10 and outputs a vehicle behavior signal indicating the movement of the vehicle 10. As the vehicle behavior sensor unit 120, an acceleration sensor that measures the acceleration of the vehicle 10, a gyro sensor that measures the azimuth of the vehicle 10, or a combination of both can be used.

In this embodiment, the vehicle behavior sensor unit 120 having the configuration shown in FIG. 3 is used. As shown in FIG. 3, the vehicle behavior sensor unit 120 includes an acceleration sensor 121 that outputs an acceleration signal corresponding to the acceleration of the vehicle 10, and an acceleration signal analysis unit 122 that analyzes the data of the acceleration signal output by the acceleration sensor 121. Equipped with. In the present embodiment, the acceleration signal is used as a vehicle behavior signal that represents the behavior of the vehicle 10. As the acceleration sensor 121, there is a uniaxial, biaxial or triaxial sensor, but any of them may be used. When the acceleration sensor 121 is a multi-axis sensor, the acceleration sensor 121 outputs the acceleration signal of each axis independently.

The bio-signal analysis unit 130 analyzes the bio-signal of the driver output by the bio-sensor unit 110, and calculates the characteristic amount necessary for estimating the driver's state from the bio-signal. In the present embodiment, the biological signal analysis unit 130 having the configuration shown in FIG. 4 is used. As shown in FIG. 4, the biological signal analysis unit 130 of the present embodiment includes a signal shaping unit 131, a spectrum calculation unit 132, a filter bank generation unit 133, a filter bank application unit 134, and a cepstrum calculation unit 135.

The signal shaping unit 131 performs preprocessing for analyzing the biological signal acquired from the biological sensor unit 110. Specifically, the signal shaping unit 131 cuts out the biological signal with a determined frame size and applies a window function. In the present embodiment, the signal shaping unit 131 removes the noise component of the biological signal due to the movement of the vehicle 10, based on the data of the vehicle behavior signal (acceleration signal) acquired from the vehicle behavior sensor unit 120. If the systematic error of the biometric sensor unit 110 is obvious, the signal shaping unit 131 may also remove the noise component of the biometric signal with a bandpass filter.

The spectrum calculation unit 132 calculates the logarithmic amplitude spectrum of the biological signal by frequency analysis such as fast Fourier transform.

The filter bank generation unit 133 generates band pass filters having different resolutions in the frequency domain between the portion corresponding to the biological signal band and the portion corresponding to the noise band. As a method of estimating a driver's fatigue, drowsiness, and anxiety level from a biological signal, there is a method based on information of the driver's heart contraction. Since the biological signal indicating the contraction of the heart can be observed as a time-series signal, it is considered to capture its characteristics by frequency analysis. Although it is necessary to increase the sampling frequency in order to increase the resolution of the sensor, if this is done, the area occupied by the biomedical signal will be small in the frequency range of the acquired data. Therefore, it is necessary to sensitively perceive information corresponding to a biological signal. Therefore, the filter bank generation unit 133 generates triangular filter groups having different pass bands as shown in FIG.

The filter bank application unit 134 applies the filter bank generated by the filter bank generation unit 133 to the logarithmic amplitude spectrum calculated by the spectrum calculation unit 132. At this time, the length of the logarithmic amplitude spectrum and the length of each filter bank must be the same. For example, when the filter bank is applied to the amplitude spectrum of the graph of FIG. 6, the amplitude spectrum after application of the filter bank is as shown in the graph of FIG. This makes it possible to extract features of the biomedical signal in the frequency domain with high density while extracting features corresponding to noise with low density.

The cepstrum calculation unit 135 converts the logarithmic amplitude spectrum obtained by the filter bank application unit 134 into a cepstrum region by a discrete cosine transform or the like to calculate a characteristic amount necessary for estimating the driver's state. At this time, information corresponding to low frequencies in the logarithmic amplitude spectrum is converted into the cepstrum region. The number of samples to be extracted may be any number. Here, it is assumed that the operation of the cepstrum calculator 135 ends after calculating the cepstrum of the biological signal scale. However, the operation of the cepstrum calculation unit 135 is not limited to this. For example, velocity information obtained by first-order differentiation of the cepstrum or acceleration information obtained by second-order differentiation of the cepstrum may be used as the additional information of the feature amount. Good. Further, the cepstrum calculation unit 135 does not necessarily have to calculate the feature amount from the biological signals of a single frame, but may calculate it from the biological signals of a plurality of connected frames.

The driver's state calculation unit 140 calculates an estimated value of the driver's state based on the feature amount data obtained as a result of the analysis of the biological signal. At this time, the driver state calculation unit 140 uses the data of the behavior of the vehicle 10 obtained by analyzing the acceleration signal by the acceleration signal analysis unit 122 from the data output by the acceleration sensor 121 in the vehicle behavior sensor unit 120 as the characteristic amount. May be added to the data. The estimated value of the driver's state can be calculated by a technique such as regression analysis in which the driver's state is the objective variable and the feature amount is the explanatory variable.

In the present embodiment, driver state calculation unit 140 calculates an estimated value of driver drowsiness. FIG. 8 is an example of subjective evaluation in data collection for model learning of regression analysis. For example, as shown in FIG. 8, the driver state calculation unit 140 performs regression analysis on drowsiness on a scale of 0 to 100, where 0 is a state where no drowsiness is felt and 100 is a state where the person is likely to sleep at the limit of drowsiness. Thus, the drowsiness of the driver is estimated from the biological signal. An example of the calculation result of the driver's state using the learned model is shown in FIG. The driver state calculation unit 140 outputs the estimated value of the driver's drowsiness at a constant cycle as shown in FIG. 9.

Here, the driver's state estimated by the driver state calculation unit 140 is drowsiness, but the present invention is not limited to this, and the driver state calculation unit 140 may estimate the driver's fatigue, for example. The estimation method is not limited to the regression analysis, and may be a known pattern recognition or machine learning method.

The degree of eye closure calculation unit 171 acquires an image of the face of the driver taken by a camera installed in the vehicle 10, analyzes the image, and calculates and outputs the degree of eye closure of the driver. The degree of eye closure is a time series signal expressed as the reciprocal of the ratio of the distance between the upper and lower eyelids when the eyes are fully open to the distance between the upper and lower eyelids at a certain time. Is. For example, if the degree of eye closure when the eyes are completely closed is 100% and the degree of eye closure when the eyes are completely open is 0%, the instantaneous peak at which the time-series signal of the degree of eye closure becomes 100%. If it appears, it indicates a blink. By analyzing the time series signal of the degree of eye closure, the driver's state such as drowsiness can be determined. The image acquired by the degree of eye closure calculation unit 171 is not limited to a still image, and may be a moving image.

The recording unit 141 synchronizes the data of the estimated value of the driver's state (here, drowsiness) calculated by the driver state calculating unit 140 and the data of the driver's eye closing degree output by the eye closing degree calculating unit 171 with each other. Let me record it.

The driver state classification unit 150 classifies the driver state by comparing a predetermined threshold value with the estimated value of the driver state calculated by the driver state calculation unit 140. Here, the driver state classification unit 150 classifies the current driver state into classes according to the degree of drowsiness. FIG. 10 is an example of class classification of the driver's state, and the drowsiness of the driver is classified into three classes of “C1: normal”, “C2: feel drowsiness”, and “C3: feel strong drowsiness”. Has been done. In FIG. 10, the threshold values between the classes are the boundary between the classes C1 and C2 (the threshold value between C1 and C2) when the learning data for building the model used in the driver state calculation unit 140 is collected, and It is the average value of the results of the subject's subjective evaluation regarding the boundary between the class C2 and the class C3 (threshold between C2 and C3). FIG. 11 shows the answer result of each subject. Here, an example is shown in which the driver's state (drowsiness) is classified into three classes, but it may be classified into more classes or into two classes depending on the purpose.

In this way, it is possible to classify the driver's state by setting the threshold value between the classes. However, as can be seen from the response results of the subjects shown in FIG. (Individual difference). In order to deal with this individuality, it is effective to correct and update the threshold value between classes for each driver.

The threshold updating unit 142 determines the individuality regarding how the driver feels the state from the past information on the driver's state and the degree of eye closure recorded in the recording unit 141, and based on the determination result, the driver The state classifying unit 150 updates the threshold value for classifying the state of the driver. Hereinafter, as an example of the operation of the threshold value updating unit 142, a method of updating the above-mentioned threshold values for the three classes of drowsiness will be described.

The threshold updating unit 142 extracts a sample from the data recorded in the recording unit 141 and obtains a regression line showing the relationship between the estimated value of drowsiness and the degree of eye closure. FIG. 12 is an example of a regression line obtained from the latest 20 samples. In the graph of FIG. 12, the horizontal axis represents the drowsiness estimation value calculated by the driver state calculation unit 140, and the vertical axis represents the driver's eye closure degree calculated by the eye closure degree calculation unit 171.

Since there is a strong correlation between the degree of eye closure and drowsiness, the threshold updating unit 142 uses the correlation to adapt the threshold to the driver. For example, the regression line of FIG. 12 has a small regression coefficient (slope of the regression line) and a small intercept. This means that the driver state calculation is performed even when the driver has a low degree of eye closure, that is, a state where he/she does not feel much drowsiness. It means that the estimated value of the drowsiness calculated by the unit 140 is likely to be large, and has the individuality. When the driver has such an individuality, for example, in a system that issues an alarm when the driver enters the state of class C2, the alarm is issued even though the driver does not actually feel drowsiness. Person feels uncomfortable. Therefore, when the regression line as shown in FIG. 12 is obtained, the threshold updating unit 142 updates the threshold between C2 and C3 to a value larger than the current value to prevent an erroneous warning from being issued. To do.

Based on the same idea as this, the threshold updating unit 142 updates the threshold value between C2 and C3 to a value smaller than the current value when the regression coefficient is large and the intercept is large, and when the regression coefficient is large and the intercept is small, C1 to C2. The inter-threshold value is updated to a value larger than the present value, and when the regression coefficient is small and the intercept is large, the C1-C2 threshold value is updated to a smaller value than the present value. The operation of the above threshold updating unit 142 is summarized as shown in FIG.

When the regression coefficient has a negative value, the estimated value of the driver's state is unstable, and thus the process of the driver's state classification unit 150 may be stopped. Further, FIG. 12 shows an example of obtaining the regression line using the latest 20 samples, but the number of samples and the interval may be arbitrary values.

The information presenting unit 160 presents information to the driver according to the result of classification of the current driver's state by the driver state classifying unit 150. The method of presenting the information may be a visual method in which an image such as an icon corresponding to the driver's state is displayed on the instrumental panel of the vehicle 10 or the center display, or a speaker provided in the vehicle 10 changes the state of the driver. It may be an audible one that outputs a corresponding voice message or alarm.

In the case of the above-mentioned three-class classification of drowsiness, when the driver's drowsiness falls into class C2, the information presenting unit 160 outputs a voice message from the speaker and displays an icon on the instrumental panel to display the driver. When the drowsiness of the person becomes Class C3, it is possible to give an alarm from the speaker and display the appearance image of the nearest rest point on the center display and the route to the rest point on the center display.

As described above, in the driver state determination device 100 according to the first embodiment, the threshold update unit 142 determines the individuality regarding how the driver feels the state from the information of the past driver state and the degree of eye closure. Then, based on the determination result, the driver state classification unit 150 updates the threshold value for classifying the driver's state. Therefore, the driver state classification unit 150 can make a detailed determination in consideration of the driver's individuality when classifying the driver's state. Therefore, the accuracy of determining the driver's condition can be improved.

Although the biometric sensor unit 110 and the vehicle behavior sensor unit 120 are arranged in the driver state determination device 100 in FIG. 1, the biometric sensor unit 110 and the vehicle behavior sensor unit 120 are used in the driver state determination device. It may be externally attached to 100. If the behavior (movement) of the vehicle is not added to the calculations performed by the biological signal analysis unit 130 and the driver state calculation unit 140, the vehicle behavior sensor unit 120 may be omitted.

<Second Embodiment>
FIG. 14 is a diagram showing the configuration of the driver state determination device 100 according to the second embodiment. The configuration of the driver state determination device 100 of FIG. 14 is obtained by replacing the eye closure degree calculation unit 171 with a route information acquisition unit 172 in the configuration of FIG. The route information acquisition unit 172 acquires position information, route information, traffic information, etc. of the vehicle 10 from the navigation system of the vehicle 10.

In the second embodiment, the recording unit 141 mutually compares the estimated state data of the driver calculated by the driver state calculation unit 140 and the route information data of the vehicle 10 acquired by the route information acquisition unit 172 with each other. Record synchronously. Further, the threshold updating unit 142 determines the individuality regarding how the driver feels the state from the information recorded in the recording unit 141, and based on the determination result, the driver state classification unit 150 causes the driver to classify. The threshold for classifying the state of is updated. Since other configurations and operations of driver state determination device 100 are the same as those in the first embodiment, description of the same components as those in the first embodiment will be omitted here.

For example, it is assumed that the driver state calculation unit 140 calculates an estimated value of anxiety degree as an estimated value of the driver's state. In this case, the driver state classifying unit 150 classifies the driver's anxiety level based on the threshold value updated by the threshold value updating unit 142, and the information presenting unit 160 determines the driver's anxiety level classification result. Present information to the driver.

The threshold updating unit 142 refers to the data recorded in the recording unit 141 and confirms the estimated value of the driver's anxiety degree when the vehicle is currently traveling on the route in the past. If the estimated value of the anxiety level of the driver has increased when the driver has traveled the current route in the past, the threshold updating unit 142 performs a process of reducing the threshold value for classifying the anxiety level.

If the current traveling route is the first traveling route and the past data regarding the route is not present in the recording unit 141, as an alternative to that, traveling to the currently traveling route among other traveling routes in the past is performed. Data of routes having similar conditions may be used.

Further, for example, it is assumed that the driver state calculation unit 140 calculates an estimated value of carelessness as an estimated value of the driver's state, and the driver state classification unit 150 classifies the driver's carelessness. .. In this case, when the route information acquisition unit 172 acquires the traffic congestion occurrence information of the route currently being traveled, the threshold updating unit 142 sets the threshold of the carelessness classification in order to prevent a rear-end collision accident due to the carelessness of the driver. You may perform processing, such as making it small.

According to the driver state determination device 100 according to the second embodiment, when the driver state classification unit 150 classifies the driver's state, the driver's individuality which is known from the traveling history of the driver is taken into consideration. Judgment will be made.

<Third Embodiment>
FIG. 15 is a diagram showing the configuration of the driver state determination device 100 according to the third embodiment. The configuration of the driver state determination device 100 of FIG. 15 is obtained by replacing the configuration of FIG. 1 with the degree of eye closure calculation unit 171 replaced by an in-vehicle environment information acquisition unit 173. The in-vehicle environment information acquisition unit 173 acquires in-vehicle environment information such as air-conditioning set values, in-vehicle temperature, humidity, and carbon dioxide concentration from the air conditioning system of the vehicle 10.

In the third embodiment, the recording unit 141 synchronizes the data of the estimated value of the driver's state calculated by the driver state calculation unit 140 and the data of the in-vehicle environment information acquired by the route information acquisition unit 172 with each other. Let me record it. Further, the threshold updating unit 142 determines the individuality regarding how the driver feels the state from the information recorded in the recording unit 141, and based on the determination result, the driver state classification unit 150 causes the driver to classify. The threshold for classifying the state of is updated. Since other configurations and operations of driver state determination device 100 are the same as those in the first embodiment, description of the same components as those in the first embodiment will be omitted here.

For example, it is assumed that the driver state calculating unit 140 calculates an estimated value of drowsiness and poor physical condition as an estimated value of the driver's state. In this case, the driver state classification unit 150 classifies the driver's drowsiness and poor physical condition based on the threshold value updated by the threshold value updating unit 142. The information presenting unit 160 also presents information to the driver according to the classification result of the drowsiness of the driver and the degree of poor physical condition.

Generally, it is known that when the carbon dioxide concentration increases, the human body is adversely affected such as fatigue, headache, and drowsiness. Therefore, the threshold updating unit 142 takes out a sample from the data recorded in the recording unit 141 to obtain a regression line, and in accordance with the regression coefficient and the size of the intercept of the regression line, similar to FIG. Update thresholds for classifying states.

Further, the threshold updating unit 142 may update the threshold for classifying the driver's drowsiness or heat stroke based on the current temperature and humidity data acquired by the in-vehicle environment information acquiring unit 173.

According to the driver state determination device 100 according to the third embodiment, when the driver state classification unit 150 classifies the driver's state, a detailed determination is made in consideration of the in-vehicle environment of the vehicle 10. Become.

<Embodiment 4>
FIG. 16 is a diagram showing the configuration of the driver state determination device 100 according to the fourth embodiment. The configuration of the driver state determination device 100 of FIG. 16 is obtained by replacing the eye closure degree calculation unit 171 with a driver reaction determination unit 174 in the configuration of FIG. 1. The driver reaction determination unit 174 indicates the operation of the vehicle 10 performed by the driver after the information presentation unit 160 presents the information to the driver (that is, after the estimated value of the driver's state exceeds the threshold value). The information is acquired and it is determined whether or not the operation of the vehicle 10 that matches the content of the information presentation has been performed, that is, whether the driver has accepted the content of the information presentation by the information presentation unit 160.

For example, it is assumed that the driver state calculation unit 140 calculates an estimated value of concentration as an estimated value of the driver's state. At this time, if the driver's state classification unit 150 determines that the driver's concentration is low and the information presenting unit 160 presents the driver with guidance to a rest point, the eye-closure degree calculating unit 171 determines After that, information on the operation of the vehicle 10 performed by the driver is acquired. Then, the degree of eye closure calculation unit 171 determines whether the driver has operated the vehicle 10 according to the content of the information presentation, that is, whether the driver has moved the vehicle 10 to the presented rest point.

In the fourth embodiment, the recording unit 141 includes the data of the estimated value of the driver's state calculated by the driver state calculating unit 140 and the determination result of the driver reaction determination unit 174 (the driver accepts the content of the information presentation. Whether or not the data) is recorded in synchronization with each other. Further, the threshold updating unit 142 determines the individuality regarding how the driver feels the state from the information recorded in the recording unit 141, and based on the determination result, the driver state classification unit 150 causes the driver to classify. The threshold for classifying the state of is updated. Since other configurations and operations of driver state determination device 100 are the same as those in the first embodiment, description of the same components as those in the first embodiment will be omitted here.

For example, when the driver's state classifying unit 150 determines that the driver's concentration has decreased and the information presenting unit 160 has presented information, but the driver did not operate the vehicle 10 accordingly. In fact, it can be considered that the driver did not feel a decrease in concentration. Therefore, if the threshold updating unit 142 refers to the data recorded in the recording unit 141 and confirms that the driver did not accept the information presentation regarding the decrease in concentration in the past, the threshold updating unit 142 suppresses excessive information presentation. First, the threshold value for classifying the driver's concentration is increased.

According to the driver state determination device 100 according to the fourth embodiment, the threshold value for classifying the driver state is updated based on whether or not the driver has accepted the content of the information presentation. Therefore, when the driver's state classification unit 150 classifies the driver's states, detailed determination according to the driver's preference is performed.

<Embodiment 5>
FIG. 17 is a diagram showing the configuration of the driver state determination device 100 according to the fifth embodiment. The configuration of the driver state determination device 100 of FIG. 17 is different from that of FIG. 1 in the route information acquisition unit 172 described in the second embodiment, the in-vehicle environment information acquisition unit 173 described in the third embodiment, and the embodiment. The driver reaction determination unit 174 described in Section 4 is added.

In the fourth embodiment, the recording unit 141 outputs the output data of the driver reaction determination unit 174 (estimated value of the driver's state), the output data of the eye closure degree calculation unit 171 (driver's eye closure degree), and route information. Output data of the acquisition unit 172 (position information and route information of the vehicle 10, traffic information, etc.), output data of the in-vehicle environment information acquisition unit 173 (in-vehicle environment information), and output data of the driver reaction determination unit 174 (driving (A determination result of whether or not the person has accepted the content of the information presentation) is recorded in synchronization with each other. Further, the threshold updating unit 142 determines the individuality regarding how the driver feels the state from the information recorded in the recording unit 141, and based on the determination result, the driver state classification unit 150 causes the driver to classify. The threshold for classifying the state of is updated. Since other configurations and operations of driver state determination device 100 are the same as those in the first embodiment, description of the same components as those in the first embodiment will be omitted here.

In order to accurately reflect the individuality of the driver in the classification of the driver's state performed by the driver state classifying unit 150, the threshold updating unit 142 causes the driver's reaction determining unit 174 to drive the driver based on the determination result. It is preferable to update the threshold for classifying the state of the person. However, if the information presenting section 160 has not presented any information due to inappropriate setting of the threshold value or the like, the driver reaction determining section 174 has not made any determination. Therefore, it is impossible to update the threshold value based on the determination result of the driver reaction determination unit 174.

Therefore, in the fifth embodiment, the threshold updating unit 142, when the output data of the driver reaction determination unit 174 is not recorded in the recording unit 141, the eye closing degree calculation unit 171, the route information acquisition unit 172, the in-vehicle environment information. The threshold value for classifying the driver's state is updated based on any output data of the acquisition unit 173. At this time, it is desirable that the threshold value is updated by taking advantage of each of the output data of the eye closure degree calculation unit 171, the route information acquisition unit 172, and the in-vehicle environment information acquisition unit 173.

An example of the operation of the threshold updating unit 142 is shown. In the present embodiment, when updating the threshold for classifying the driver's state, the threshold updating unit 142 follows the flow of FIG. 18 and calculates the degree of eye closure calculation unit 171, route information acquisition unit 172, in-vehicle environment information acquisition unit. 173 or the output data of the driver reaction determination unit 174 is selected to update the threshold value.

As described above, in order to accurately reflect the individuality of the driver in the classification of the driver's state, it is preferable that the threshold value be updated based on the output data of the driver reaction determination unit 174. Therefore, the threshold update unit 142 first confirms whether or not the output data of the driver reaction determination unit 174 is recorded in the recording unit 141 (step ST101). However, if the output data of the driver reaction determination unit 174 is recorded (YES in step ST101), the threshold value is updated based on the data (step ST102). However, if the output data of the driver reaction determination unit 174 is not recorded in the recording unit 141 (NO in step ST101), the following process is performed.

For example, if the vehicle 10 is traveling on a route on which the driver has traveled, the output data of the route information acquisition unit 172 regarding that route is recorded in the recording unit 141, and the threshold value is based on that data. By updating, the driver's individuality can be reflected in the threshold value. However, if the vehicle 10 is traveling on a route that the driver has not traveled before, the output data of the route information acquisition unit 172 regarding the route is not recorded in the recording unit 141, and thus the degree of eye closure calculation unit 171 or It is necessary to use the output data of the in-vehicle environment information acquisition unit 173.

Therefore, the threshold updating unit 142 determines whether or not the route currently being traveled is a route with which the driver has traveled, that is, the output data of the route information acquisition unit 172 regarding the currently traveling route is recorded in the recording unit 141. It is confirmed whether or not there is any (step ST103). If the route currently being traveled is a route on which the driver has traveled (YES in step ST103), the threshold value is updated based on the output data of the route information acquisition unit 172 (step ST104).

If the route the vehicle 10 is currently traveling on is a route that the driver has not traveled to (NO in step ST103), the threshold updating unit 142 determines whether the current time is night and whether the vehicle 10 is traveling in the tunnel. It is confirmed whether it is inside (step ST105).

Since the degree of eye closure calculation unit 171 calculates the degree of eye closure from the image of the driver's face taken by the camera, it generally depends on the performance and specifications of the camera, but generally, a clear image such as at night or in a tunnel is taken. In such a situation, it is difficult to calculate the degree of eye closure. Therefore, the threshold updating unit 142 updates the threshold based on the output data of the in-vehicle environment information acquiring unit 173 recorded in the recording unit 141 during the nighttime or during the tunnel traveling (YES in step ST105) (step ST106). ). Further, when neither at night nor in a tunnel (NO in step ST105), the threshold updating unit 142 updates the threshold based on the output data of the eye closure degree calculating unit 171 recorded in the recording unit 141 (step ST107). ..

According to the present embodiment, the thresholds for classifying the driver's state are updated by taking advantage of the advantages of each of the first to fourth embodiments. Therefore, when the driver's state classification unit 150 classifies the driver's states, more detailed determination is performed.

<Hardware configuration example>
19 and 20 are diagrams each showing an example of the hardware configuration of the driver state determination device 100. Each function of the components of the driver state determination device 100 is realized by the processing circuit 50 shown in FIG. 19, for example. That is, the driver state determination device 100 obtains the biological signal of the driver of the vehicle and analyzes the biological signal to calculate the characteristic amount used for estimating the driver's state from the biological signal, and the characteristic of the biological signal. Based on the amount, calculate the estimated value of the driver's state, by comparing the estimated value of the driver's state with a predetermined threshold, classify the driver's state, the classification result of the driver's state In accordance with the information provided to the driver, the driver's degree of eye closure, vehicle route information, vehicle interior environment information, and information on whether or not the vehicle operation matching the content of the information presentation has been performed. A processing circuit 50 is provided for updating the threshold value based on the past information of at least one of them and the estimated value of the driver's state. The processing circuit 50 may be dedicated hardware, or may be a processor (Central Processing Unit (CPU), a processing device, an arithmetic device, a microprocessor, a microcomputer, a computer) that executes a program stored in a memory. It may be configured by using a DSP (also called a Digital Signal Processor).

When the processing circuit 50 is dedicated hardware, the processing circuit 50 is, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), or an FPGA (Field-Programmable). Gate Array), or a combination of these. Each function of the components of the driver state determination device 100 may be realized by an individual processing circuit, or those functions may be collectively realized by one processing circuit.

FIG. 20 shows an example of the hardware configuration of the driver state determination device 100 when the processing circuit 50 is configured using the processor 51 that executes a program. In this case, the functions of the components of the driver state determination device 100 are realized by software (software, firmware, or a combination of software and firmware). The software and the like are written as a program and stored in the memory 52. The processor 51 realizes the function of each unit by reading and executing the program stored in the memory 52. That is, the driver state determination device 100, when executed by the processor 51, obtains a biological signal of the driver of the vehicle and analyzes the biological signal to estimate the driver's state from the biological signal. A process of calculating a feature amount to be used, a process of calculating an estimated value of the driver's state based on the feature amount of the biological signal, and comparing the estimated value of the driver's state with a predetermined threshold value, A process of classifying the state of the driver and presenting information to the driver according to the result of classifying the state of the driver, the degree of eye closure of the driver, route information of the vehicle, in-vehicle environment information of the vehicle, and information presentation The process of updating the threshold value based on the past information of at least one of the information indicating whether or not the vehicle is operated in accordance with the above information and the estimated value of the driver's state is the result. A memory 52 for storing the program to be executed. In other words, it can be said that this program causes a computer to execute the procedure and method of the operation of the components of the driver's state determination device 100.

Here, the memory 52 is a nonvolatile memory such as a RAM (Random Access Memory), a ROM (Read Only Memory), a flash memory, an EPROM (Erasable Programmable Read Only Memory), or an EEPROM (Electrically Erasable Programmable Read Only Memory). Volatile semiconductor memory, HDD (Hard Disk Drive), magnetic disk, flexible disk, optical disk, compact disk, mini disk, DVD (Digital Versatile Disc) and its drive device, or any storage medium used in the future. May be.

Heretofore, a configuration has been described in which the functions of the components of the driver state determination device 100 are realized by either hardware or software. However, the configuration is not limited to this, and a configuration in which some of the components of the driver state determination device 100 are realized by dedicated hardware and another of the other components is realized by software or the like may be used. For example, for some of the constituent elements, the function is realized by the processing circuit 50 as dedicated hardware, and for some of the other constituent elements, the processing circuit 50 as the processor 51 executes a program stored in the memory 52. The function can be realized by reading and executing.

As described above, the driver state determination device 100 can realize each function described above by hardware, software, or the like, or a combination thereof.

It should be noted that, in the present invention, the respective embodiments can be freely combined, or the respective embodiments can be appropriately modified or omitted within the scope of the invention.

Although the present invention has been described in detail, the above description is illustrative in all aspects, and the present invention is not limited thereto. It is understood that innumerable variants not illustrated can be envisaged without departing from the scope of the invention.

Reference numeral 113 posture sensor, 114 biosensor selection unit, 120 vehicle behavior sensor unit, 121 acceleration sensor, 122 acceleration signal analysis unit, 130 biosignal analysis unit, 131 signal shaping unit, 132 spectrum calculation unit, 133 filter bank generation unit, 134 filter Bank application unit, 135 cepstrum calculation unit, 140 driver state calculation unit, 141 recording unit, 142 threshold value updating unit, 150 driver state classification unit, 160 information presentation unit, 171 eye closure degree calculation unit, 172 route information acquisition unit, 173 In-vehicle environment information acquisition unit, 174 driver reaction determination unit, 50 processing circuit, 51 processor, 52 memory.

Claims (7)

A biometric sensor unit that acquires a biometric signal of a vehicle driver,
By analyzing the biological signal, a biological signal analysis unit that calculates a feature amount used to estimate the state of the driver from the biological signal,
Based on the feature amount of the biological signal, a driver state calculation unit that calculates an estimated value of the state of the driver,
By comparing the estimated value of the state of the driver with a predetermined threshold value, a driver state classification unit that classifies the state of the driver,
An information presentation unit that presents information to the driver according to the classification result of the state of the driver,
At least one of the degree of eye closure of the driver, route information of the vehicle, in-vehicle environment information of the vehicle, and information on whether or not the vehicle operation matching the content of the information presentation has been performed, A recording unit that records the estimated value of the driver's state in synchronization with each other,
Information regarding whether or not the driver's eye closure, the route information of the vehicle, the in-vehicle environment information of the vehicle, and the operation of the vehicle matching the content of the information presentation , recorded in the recording unit. A threshold updating unit that updates the threshold based on the regression coefficient and the size of the intercept of a regression line based on the correlation between at least one of the estimated values of the driver's state and
Equipped with
The biological sensor unit,
A plurality of biometric sensors with portions each measurement area do not overlap each other,
A posture sensor for obtaining posture information representing the posture of the driver,
Based on the posture information, a biosensor selecting the biosensor in which the driver exists in the measurement range from the plurality of biosensors, and outputting a biosignal acquired by the selected biosensor, and a biosensor selection unit,
With
Driver status determination device. A biometric sensor unit that acquires a biometric signal of a vehicle driver,
By analyzing the biological signal, a biological signal analysis unit that calculates a feature amount used to estimate the state of the driver from the biological signal,
Based on the feature amount of the biological signal, a driver state calculation unit that calculates an estimated value of the state of the driver,
By comparing the estimated value of the state of the driver with a predetermined threshold value, a driver state classification unit that classifies the state of the driver,
An information presentation unit that presents information to the driver according to the classification result of the state of the driver,
At least one of the degree of eye closure of the driver, route information of the vehicle, in-vehicle environment information of the vehicle, and information on whether or not the vehicle operation matching the content of the information presentation has been performed, A recording unit that records the estimated value of the driver's state in synchronization with each other,
Information regarding whether or not the driver's eye closure, the route information of the vehicle, the in-vehicle environment information of the vehicle, and the operation of the vehicle matching the content of the information presentation , recorded in the recording unit. Based on the regression coefficient of the regression line and the size of the intercept based on the correlation between at least one of the above and the estimated value of the driver's state, and considering the individual difference in the way of feeling the driver's state, the threshold value is set to A threshold updating unit for updating,
A driver condition determination device including. Further comprising a vehicle behavior sensor unit that outputs a vehicle behavior signal that represents the movement of the vehicle,
The biological signal analysis unit analyzes the biological signal in consideration of the vehicle behavior signal,
The driver state determination device according to claim 1 or 2. The vehicle behavior signal is an acceleration signal of the vehicle,
The driver condition determination device according to claim 3. The biological signal analysis unit performs frequency analysis of the biological signal, and in the process of the frequency analysis, with respect to the logarithmic amplitude spectrum of the biological signal, the resolution in a frequency band including the biological signal and other frequency bands. Apply different filter banks of,
The driver state determination device according to any one of claims 1 to 4. The biometric sensor unit of the driver state determination device acquires the biometric signal of the driver of the vehicle,
The biological signal analysis unit of the driver state determination device, by analyzing the biological signal, to calculate a feature amount used to estimate the state of the driver from the biological signal,
The driver state calculation unit of the driver state determination device, based on the feature amount of the biological signal, calculates an estimated value of the driver state,
The driver state classification unit of the driver state determination device, by comparing the estimated value of the driver state with a predetermined threshold value, to classify the driver state,
The information presenting unit of the driver state determination device presents information to the driver according to the classification result of the state of the driver,
Whether the threshold update unit of the driver state determination device has operated the vehicle in accordance with the degree of eye closure of the driver, route information of the vehicle, in-vehicle environment information of the vehicle, and the content of the information presentation. At least one of information on whether or not the driver's state and the estimated value of the driver's state, obtained from past information , the degree of eyesight of the driver, route information of the vehicle, in-vehicle environment information of the vehicle, and The regression coefficient of the regression line and the size of the intercept based on the correlation between at least one of the information indicating whether the operation of the vehicle that matches the content of the information presentation is performed and the estimated value of the driver's state. Based on the updated threshold,
The biological sensor unit,
Based on the posture information indicating the posture of the driver, a biosensor in which the driver is present in the measurement area is selected from a plurality of biosensors having different measurement areas, and the selected biosensor is selected. Obtains the biological signal of the driver output by,
Driver status judgment method. The biometric sensor unit of the driver state determination device acquires the biometric signal of the driver of the vehicle,
The biological signal analysis unit of the driver state determination device, by analyzing the biological signal, to calculate a feature amount used to estimate the state of the driver from the biological signal,
The driver state calculation unit of the driver state determination device, based on the feature amount of the biological signal, calculates an estimated value of the driver state,
The driver state classification unit of the driver state determination device, by comparing the estimated value of the driver state with a predetermined threshold value, to classify the driver state,
The information presenting unit of the driver state determination device presents information to the driver according to the classification result of the state of the driver,
Whether the threshold update unit of the driver state determination device has operated the vehicle in accordance with the degree of eye closure of the driver, route information of the vehicle, in-vehicle environment information of the vehicle, and the content of the information presentation. At least one of information on whether or not the driver's state and the estimated value of the driver's state, obtained from past information , the degree of eyesight of the driver, route information of the vehicle, in-vehicle environment information of the vehicle, and The regression coefficient of the regression line and the size of the intercept based on the correlation between at least one of the information indicating whether the operation of the vehicle that matches the content of the information presentation is performed and the estimated value of the driver's state. Based on the individual differences in how the driver feels the state, the threshold value is updated.
Driver status judgment method.

Images