Tailgating sign estimation device, tailgating sign estimation method and program

Abstract

To effectively utilize an estimation result of a sign of tailgating of a self vehicle for preventing the tailgating of the self vehicle.SOLUTION: A tailgating sign estimation device 20 comprises: a tailgating sign determination section 236 for determining the presence / absence of a sign of tailgating of a self vehicle 1 based on a plurality of factors of the sign of the tailgating of the self vehicle 1 which is at least any one of information relating to a driver of the self vehicle 1, information relating to a peripheral vehicle and information relating to the self vehicle 1; and a tailgating sign level calculation section 235 for calculating a sign level of the tailgating of the self vehicle 1 by adding products of scores and weights of the plurality of factors of the sign of the tailgating of the self vehicle 1. In a case where the sign level of the tailgating of the self vehicle 1 calculated by the tailgating sign level calculation section 235 is larger than a threshold, the tailgating sign determination section 236 determines that there is a sign of tailgating of the self vehicle 1.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present disclosure relates to an aggressive driving sign estimating device, an aggressive driving sign estimating method, and a program. [Background technology]

[0002] Patent Document 1 describes a technology that detects the behavior of a vehicle relative to a vehicle ahead using a forward sensor and a vehicle speed sensor to detect whether the vehicle is tailgating. In the technology described in Patent Document 1, the forward sensor is realized as, for example, a camera, radar, or sonar, and detects the position and vehicle speed of a vehicle ahead that is located or traveling ahead of the vehicle on which the tailgating suppression device is implemented. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2020-052751 A Summary of the Invention [Problem to be solved by the invention]

[0004] Incidentally, the technology described in Patent Document 1 can detect that the vehicle is being aggressive, but cannot detect signs of aggressive driving of the vehicle (i.e., the stage before the vehicle starts aggressive driving). Therefore, the technology described in Patent Document 1 cannot prevent aggressive driving of the vehicle in advance.

[0005] In view of the above-mentioned points, the present disclosure aims to provide a tailgating sign estimation device, a tailgating sign estimation method, and a program that can utilize the estimated results of tailgating signs of the host vehicle to prevent tailgating of the host vehicle. [Means for solving the problem]

[0006] One aspect of the present disclosure is a tailgating sign estimation device that includes a tailgating sign determination unit that determines whether or not there are signs of tailgating of the host vehicle based on multiple factors of signs of tailgating of the host vehicle, which are at least one of information about the driver of the host vehicle, information about surrounding vehicles, and information about the host vehicle, and a tailgating sign degree calculation unit that calculates a sign degree of tailgating of the host vehicle by adding up the product of the scores and weights of each of the multiple factors of signs of tailgating of the host vehicle, and the tailgating sign determination unit determines that there are signs of tailgating of the host vehicle when the sign degree of tailgating of the host vehicle calculated by the tailgating sign degree calculation unit is greater than a threshold value. Effect of the Invention

[0007] According to the present disclosure, the estimation results of signs of aggressive driving of the host vehicle can be utilized to prevent aggressive driving of the host vehicle. [Brief description of the drawings]

[0008] [Figure 1] 1 is a diagram showing an example of a host vehicle 1 to which a road rage symptom estimation device 20 according to a first embodiment is applied. [Diagram 2] 2 is a diagram showing an example of a specific configuration of the road rage symptom estimation device 20 shown in FIG. 1. FIG. [Diagram 3] FIG. 2 is a diagram showing an example of a specific configuration of the sedation treatment device 30 shown in FIG. [Figure 4] 4 is a flowchart for explaining an example of processing executed in a host vehicle 1 to which the aggressive driving sign estimation device 20 of the first embodiment is applied. [Diagram 5] 5 is a flowchart for explaining an example of a process performed by the estimation unit 234 to estimate the facial expression and emotion De of the driver in step S2D of FIG. 4. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, with reference to the drawings, embodiments of the road rage sign estimating device, road rage sign estimating method, and program of the present disclosure will be described.

[0010] First Embodiment Fig. 1 is a diagram showing an example of a vehicle 1 to which a road rage sign estimation device 20 according to a first embodiment is applied. Fig. 2 is a diagram showing an example of a specific configuration of the road rage sign estimation device 20 shown in Fig. 1. Fig. 3 is a diagram showing an example of a specific configuration of the tranquilization processing device 30 shown in Fig. 1. 1 to 3, a vehicle 1 includes an external sensor 2, a vehicle control device 10, and an aggressive driving symptom estimation device 20. The external sensor 2 includes, for example, a front camera, a surrounding camera, a radar, a LiDAR (Light Detection And Ranging), a sonar, a light sensor, a temperature sensor, a humidity sensor, and the like. 1 to 3, the vehicle 1 includes a front camera, a surrounding camera, a radar, a LiDAR, a sonar, a light sensor, a microphone, a temperature sensor, and a humidity sensor as the external sensor 2. In other examples, the vehicle 1 may not include all of these sensors as the external sensor 2, or the vehicle 1 may include a sensor other than these sensors as the external sensor 2.

[0011] In the example shown in FIG. 1 to FIG. 3, the front camera captures an image including surrounding vehicles in front of the vehicle 1 to generate image data, and transmits the image data to the vehicle control device 10 and the tailgating sign estimation device 20 via the in-vehicle network 14. The surrounding camera captures an image including surrounding vehicles located around the vehicle 1 to generate image data, and transmits the image data to the vehicle control device 10 and the tailgating sign estimation device 20 via the in-vehicle network 14. The radar is, for example, a millimeter wave radar, a 24 GHz narrow band radar, etc., and detects the relative position and relative speed of the surrounding vehicles and the surrounding road structure with respect to the vehicle 1, and transmits the detection result to the vehicle control device 10 and the tailgating sign estimation device 20 via the in-vehicle network 14. The LiDAR detects the relative position and relative speed of the surrounding vehicles and the surrounding road structure with respect to the vehicle 1, and transmits the detection result to the vehicle control device 10 and the tailgating sign estimation device 20 via the in-vehicle network 14. The sonar detects the distance between the vehicle 1 and surrounding vehicles, etc., and transmits the detection result to the vehicle control device 10 and the tailgating sign estimation device 20 via the in-vehicle network 14. The optical sensor detects the light of surrounding vehicles (e.g., oncoming vehicles, following vehicles, etc.) located around the vehicle 1 (e.g., in front, behind, etc.), and transmits the detection result to the tailgating sign estimation device 20 via the in-vehicle network 14. The microphone detects sounds outside the vehicle 1, and transmits the detection result to the tailgating sign estimation device 20 via the in-vehicle network 14. The temperature sensor detects the temperature outside or inside the vehicle 1, and transmits the detection result to the tailgating sign estimation device 20 via the in-vehicle network 14. The humidity sensor detects the humidity outside or inside the vehicle 1, and transmits the detection result to the tailgating sign estimation device 20 via the in-vehicle network 14.

[0012] In the example shown in FIG. 1 to FIG. 3, the vehicle 1 is equipped with a driver monitor camera 3. The driver monitor camera 3 captures an image including the face of the driver of the vehicle 1 to generate image data, and transmits the image data to the vehicle control device 10 and the tailgating sign estimation device 20 via the in-vehicle network 14. The vehicle 1 is also equipped with a communication device 4. The communication device 4 communicates with a communication device (not shown) outside the vehicle 1. The communication device 4 acquires traffic light (color) information from, for example, a roadside device (not shown) on the road side, and transmits the traffic light information to the tailgating sign estimation device 20 via the in-vehicle network 14. The vehicle 1 is also equipped with an HMI (Human Machine Interface) 5. The HMI 5 accepts input of information (for example, gender, etc.) related to the driver's license of the driver of the vehicle 1, and transmits the information related to the driver's license of the driver of the vehicle 1 to the tailgating sign estimation device 20 via the in-vehicle network 14. The HMI 5 also accepts input of the age of the driver of the vehicle 1, and transmits information indicating the age of the driver of the vehicle 1 to the tailgating sign estimation device 20 via the in-vehicle network 14. The HMI 5 also accepts input of information regarding the illness of the driver of the vehicle 1, and transmits the information regarding the illness of the driver of the vehicle 1 to the tailgating sign estimation device 20 via the in-vehicle network 14. When the HMI 5 accepts input that the driver of the vehicle 1 has psychopathic characteristics, it transmits information indicating that the driver of the vehicle 1 has psychopathic characteristics to the tailgating sign estimation device 20. When the HMI 5 accepts input that the driver of the vehicle 1 has a mental illness, it transmits information indicating that the driver of the vehicle 1 has a mental illness to the tailgating sign estimation device 20. When the HMI 5 accepts input that the driver of the vehicle 1 has a neurological disorder, it transmits information indicating that the driver of the vehicle 1 has a neurological disorder to the tailgating sign estimation device 20. In addition, the microphone of the HMI 5 detects sounds made by the driver of the vehicle 1 and transmits the detection result to the aggressive driving sign estimation device 20 via the in-vehicle network 14. In addition, the HMI 5 accepts input of information related to the income of the driver of the vehicle 1 and transmits the information related to the income of the driver of the vehicle 1 to the aggressive driving sign estimation device 20 via the in-vehicle network 14.The HMI 5 also accepts input of information regarding the driving skills of the driver of the vehicle 1, and transmits the information regarding the driving skills of the driver of the vehicle 1 to the tailgating sign estimation device 20 via the in-vehicle network 14. The HMI 5 also accepts input of information regarding the color of the driver's license of the driver of the vehicle 1, and transmits the information regarding the color of the driver's license of the driver of the vehicle 1 to the tailgating sign estimation device 20 via the in-vehicle network 14. The HMI 5 also accepts input of information regarding the occupation of the driver of the vehicle 1, and transmits information regarding the occupation of the driver of the vehicle 1 (for example, information that the driver of the vehicle 1 is a professional driver, etc.) to the tailgating sign estimation device 20 via the in-vehicle network 14.

[0013] In the example shown in FIGS. 1 to 3, the host vehicle 1 is equipped with a GPS (Global Positioning System) unit 6, a map information unit 7, a steering actuator 11, a braking actuator 12, and a drive actuator 13. In the example shown in Figures 1 to 3, the vehicle 1 is an autonomous vehicle and has an autonomous driving control function, but in other examples, the vehicle 1 may be a manually driven vehicle and may not have an autonomous driving control function. 1 to 3, a tailgating sign estimation device 20 estimates tailgating signs of a vehicle 1. The tailgating sign estimation device 20 is configured by a microcomputer including a communication interface (I / F) 21, a memory 22, and a processor 23. The communication interface 21, the memory 22, and the processor 23 are connected via a signal line 24. 1 to 3, the road rage sign estimation device 20 includes one processor 23, but in other examples, the road rage sign estimation device 20 may include multiple processors. Also, in the examples shown in Fig. 1 to 3, the road rage sign estimation device 20 is configured with one ECU (Electronic Control Unit), but in other examples, the road rage sign estimation device 20 may be configured with multiple ECUs.

[0014] In the example shown in FIGS. 1 to 3, the processor 23 includes an acquisition unit 231, a detection unit 232, a recognition unit 233, an estimation unit 234, a road rage symptom degree calculation unit 235, and a road rage symptom determination unit 236. The acquisition unit 231 acquires image data including the face of the driver of the vehicle 1 captured by the driver monitor camera 3.

[0015] The acquisition unit 231 also acquires driver attribute information Dc (one of the factors that indicate aggressive driving of the vehicle 1) as information about the driver of the vehicle 1. The driver attribute information Dc acquired by the acquisition unit 231 includes the gender, age (generation), income, driving skills, color of the driver's license, occupation, etc. of the driver of the vehicle 1. The acquisition unit 231 estimates the gender of the driver of the vehicle 1 based on an image including the face of the driver of the vehicle 1 captured by the driver monitor camera 3, for example, by using a face recognition AI (artificial intelligence) or the like, and acquires the estimation result as the driver attribute information Dc. When the HMI 5 accepts an input of the gender of the driver of the vehicle 1, the acquisition unit 231 acquires the input gender of the driver of the vehicle 1 as the driver attribute information Dc. When the acquisition unit 231 acquires, as the driver attribute information Dc, that the gender of the driver of the vehicle 1 is male, for example, the acquisition unit 231 sets a score Dc1 and a weight wDc1 corresponding to the gender of the driver of the vehicle 1 being male. When the acquisition unit 231 acquires, as the driver attribute information Dc, that the gender of the driver of the vehicle 1 is female, for example, the acquisition unit 231 sets a score Dc2 and a weight wDc2 corresponding to the gender of the driver of the vehicle 1 being female (0 <Dc2<Dc1)。

[0016] The acquisition unit 231 estimates the age (generation) of the driver of the vehicle 1 based on an image including the face of the driver of the vehicle 1 captured by the driver monitor camera 3, for example, by using face recognition AI (artificial intelligence), and acquires the estimation result as the driver attribute information Dc. When the HMI 5 accepts an input of the age (generation) of the driver of the vehicle 1, the acquisition unit 231 acquires the input age (generation) of the driver of the vehicle 1 as the driver attribute information Dc. When the acquisition unit 231 acquires the generation (teenagers) of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc3 and a weight wDc3 corresponding to the generation (teenagers) of the driver of the vehicle 1. When the acquisition unit 231 acquires the generation (twenties) of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc4 and a weight wDc4 corresponding to the generation (twenties) of the driver of the vehicle 1. When the acquisition unit 231 acquires the age (30s) of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc5 and a weight wDc5 corresponding to the age (30s) of the driver of the vehicle 1. When the acquisition unit 231 acquires the age (40s) of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc6 and a weight wDc6 corresponding to the age (40s) of the driver of the vehicle 1. When the acquisition unit 231 acquires the age (50s) of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc7 and a weight wDc7 corresponding to the age (50s) of the driver of the vehicle 1. When the acquisition unit 231 acquires the age (60s) of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc8 and a weight wDc8 corresponding to the age (60s) of the driver of the vehicle 1. When the acquisition unit 231 acquires the age (70s) of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc9 and a weight wDc9 corresponding to the age (70s) of the driver of the vehicle 1.When the acquisition unit 231 acquires the age group (80s) of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc10 and a weight wDc10 corresponding to the age group (80s) of the driver of the vehicle 1 (0 <Dc3、Dc7、Dc8、Dc9、Dc10<Dc4、Dc5、Dc6)。

[0017] The acquisition unit 231 acquires information related to the income of the driver of the vehicle 1, the input of which is accepted by the HMI 5, as the driver attribute information Dc. When the acquisition unit 231 acquires the income (low) of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc11 and a weight wDc11 corresponding to the income (low) of the driver of the vehicle 1. When the acquisition unit 231 acquires the income (medium) of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc12 and a weight wDc12 corresponding to the income (medium) of the driver of the vehicle 1. When the acquisition unit 231 acquires the income (high) of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc13 and a weight wDc13 corresponding to the income (high) of the driver of the vehicle 1 (0 <Dc11<Dc12<Dc13)。 The acquisition unit 231 acquires information related to the driving skill of the driver of the vehicle 1, the input of which has been accepted by the HMI 5, as the driver attribute information Dc. When the acquisition unit 231 acquires the driving skill (low) of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc14 and a weight wDc14 corresponding to the driving skill (low) of the driver of the vehicle 1. When the acquisition unit 231 acquires the driving skill (medium) of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc15 and a weight wDc15 corresponding to the driving skill (medium) of the driver of the vehicle 1. When the acquisition unit 231 acquires the driving skill (high) of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc16 and a weight wDc16 corresponding to the driving skill (high) of the driver of the vehicle 1 (0 <Dc14<Dc15<Dc16)。 The acquisition unit 231 acquires information relating to the color of the driver's license of the driver of the vehicle 1, the input of which has been accepted by the HMI 5, as the driver attribute information Dc. When the acquisition unit 231 acquires the color (non-gold) of the driver's license of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc17 and a weight wDc17 corresponding to the color (non-gold) of the driver's license of the driver of the vehicle 1. When the acquisition unit 231 acquires the color (gold) of the driver's license of the driver of the vehicle 1 as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc18 and a weight wDc18 corresponding to the color (gold) of the driver's license of the driver of the vehicle 1 (0 <Dc18<Dc17)。 The acquisition unit 231 acquires information related to the occupation of the driver of the vehicle 1, the input of which has been accepted by the HMI 5, as the driver attribute information Dc. When the acquisition unit 231 acquires the occupation of the driver of the vehicle 1 (occupation other than professional driver) as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc19 and a weight wDc19 corresponding to the occupation of the driver of the vehicle 1 (occupation other than professional driver). When the acquisition unit 231 acquires the occupation of the driver of the vehicle 1 (professional driver) as the driver attribute information Dc, for example, the acquisition unit 231 sets a score Dc20 and a weight wDc20 corresponding to the occupation of the driver of the vehicle 1 (professional driver) (0 <Dc20<Dc19)。

[0018] The acquisition unit 231 also acquires, as driver disease information Dd, information on the illness of the driver of the vehicle 1, the input of which has been accepted by the HMI 5, as information on the driver of the vehicle 1. The driver disease information Dd acquired by the acquisition unit 231 includes information indicating that the driver of the vehicle 1 has psychopathic characteristics, a mental illness, a neurological disorder, or the like. When the acquisition unit 231 acquires, as the driver disease information Dd, that the driver of the vehicle 1 does not have psychopathic traits, for example, the acquisition unit 231 sets a score Dd1 and a weight wDd1 corresponding to the driver of the vehicle 1 not having psychopathic traits. When the acquisition unit 231 acquires, as the driver disease information Dd, that the driver of the vehicle 1 has psychopathic traits, for example, the acquisition unit 231 sets a score Dd2 and a weight wDd2 corresponding to the driver of the vehicle 1 having psychopathic traits (0 <Dd1<Dd2)。 When the acquisition unit 231 acquires, as the driver illness information Dd, that the driver of the vehicle 1 does not have a mental illness, for example, the acquisition unit 231 sets a score Dd3 and a weight wDd3 corresponding to the driver of the vehicle 1 not having a mental illness. When the acquisition unit 231 acquires, as the driver illness information Dd, that the driver of the vehicle 1 has a mental illness, for example, the acquisition unit 231 sets a score Dd4 and a weight wDd4 corresponding to the driver of the vehicle 1 having a mental illness (0 <Dd3<Dd4)。 When the acquisition unit 231 acquires, as the driver disease information Dd, that the driver of the vehicle 1 does not have a neurological disorder, for example, the acquisition unit 231 sets a score Dd5 and a weight wDd5 corresponding to the driver of the vehicle 1 not having a neurological disorder. When the acquisition unit 231 acquires, as the driver disease information Dd, that the driver of the vehicle 1 has a neurological disorder, for example, the acquisition unit 231 sets a score Dd6 and a weight wDd6 corresponding to the driver of the vehicle 1 having a neurological disorder (0 <Dd5<Dd6)。

[0019] The detection unit 232 detects the position of the face of the driver of the vehicle 1 based on an image including the face of the driver of the vehicle 1 captured by the driver monitor camera 3. The detection unit 232 also detects facial parts (e.g., eyes, mouth, etc.) of the driver of the vehicle 1, clothing worn by the driver of the vehicle 1 (e.g., glasses, sunglasses, mask, etc.) and facial feature amounts of the driver of the vehicle 1 based on an image including the face of the driver of the vehicle 1 captured by the driver monitor camera 3.

[0020] The recognition unit 233 recognizes the behavior Sb of the surrounding vehicle based on the detection result of the external sensor 2. The recognition unit 233 has a function of recognizing that the surrounding vehicle has cut in front of the vehicle 1 based on the detection result of the front camera, radar, and LiDAR as the external sensor 2. Specifically, the recognition unit 233 estimates whether the surrounding vehicle detected by the external sensor 2 has cut in front of the vehicle 1 based on the detection result of the surrounding vehicle by the external sensor 2 by using a model obtained by learning using teacher data, which is a data set of the detection result of the surrounding vehicle by the external sensor and a label indicating whether the surrounding vehicle detected by the external sensor has cut in, and recognizes the behavior Sb of the surrounding vehicle. The following may be used as a definition of "cutting in." For example, if the host vehicle 1 is a vehicle equipped with ACC (adaptive cruise control), a surrounding vehicle is considered to have cut in when it comes closer than 1.2 times the ACC following distance set by the driver (30 m for a short setting at a vehicle speed of 80 km / h, 56 m for a long setting, etc.). In other words, the range in which the driver of the host vehicle 1 may feel that there is a near miss, even if the automatic brakes of the host vehicle 1 are not activated, is defined as "cutting in." In this example, the above is taken into consideration when creating training data. In the example shown in FIG. 1 to FIG. 3, when the recognition unit 233 recognizes that the surrounding vehicle has not cut in front of the host vehicle 1 as the behavior Sb of the surrounding vehicle, for example, the recognition unit 233 sets a score Sb1 and a weight wSb1 corresponding to the surrounding vehicle not cutting in front of the host vehicle 1. When the recognition unit 233 recognizes that the surrounding vehicle has cut in front of the host vehicle 1 as the behavior Sb of the surrounding vehicle, for example, the recognition unit 233 sets a score Sb2 and a weight wSb2 corresponding to the surrounding vehicle cutting in front of the host vehicle 1 (0 <Sb1<Sb2)。 The recognition unit 233 also has a function of recognizing that the speed of a surrounding vehicle (preceding vehicle) in front of the vehicle 1 detected by the external sensor 2 is slow based on the detection results of the front camera, radar, and LiDAR as the external sensor 2. Specifically, the recognition unit 233 estimates whether the speed of the surrounding vehicle (preceding vehicle) in front of the vehicle 1 detected by the external sensor 2 is slow or not based on the detection result of the surrounding vehicle (preceding vehicle) in front of the vehicle 1 by the external sensor 2 by using a model obtained by performing learning using teacher data, which is a data set of the detection result of the surrounding vehicle (preceding vehicle) by the external sensor and a label indicating whether the speed of the surrounding vehicle (preceding vehicle) detected by the external sensor is slow or not, and recognizes the behavior Sb of the surrounding vehicle (preceding vehicle). When the recognition unit 233 recognizes that the speed of the surrounding vehicle (preceding vehicle) is not slow as the behavior Sb of the surrounding vehicle, for example, the recognition unit 233 sets a score Sb3 and a weight wSb3 corresponding to the speed of the surrounding vehicle (preceding vehicle) not being slow. When the recognition unit 233 recognizes that the speed of the surrounding vehicle (preceding vehicle) is slow as the behavior Sb of the surrounding vehicle, for example, the recognition unit 233 sets a score Sb4 and a weight wSb4 corresponding to the slow speed of the surrounding vehicle (preceding vehicle) (0 <Sb3<Sb4)。

[0021] The recognition unit 233 has a function of recognizing that the host vehicle 1 has given way to a surrounding vehicle based on an image of the front of the host vehicle 1 captured by a forward camera as the external sensor 2. Specifically, the recognition unit 233 estimates whether the host vehicle 1 has given way to a surrounding vehicle included in an image based on an image in front of the host vehicle 1 captured by the forward camera by using a model obtained by performing learning using teacher data, which is a data set of an image including a surrounding vehicle and a label indicating whether the surrounding vehicle has changed lanes, merged, or the like because the surrounding vehicle has given way to the surrounding vehicle included in the image. The recognition unit 233 has a function of recognizing that the surrounding vehicle did not signal when the host vehicle 1 gave way to the surrounding vehicle based on an image including the surrounding vehicle in front of the host vehicle 1 captured by the forward camera as the external sensor 2. Specifically, the recognition unit 233 uses a model obtained by learning using teacher data, which is a data set of images including surrounding vehicles and labels included in the images indicating whether the surrounding vehicles made a signal when given the right of way, to estimate whether the surrounding vehicles made a signal when the host vehicle 1 gave way to the surrounding vehicle based on an image including the surrounding vehicles in front of the host vehicle 1 captured by the forward camera, and recognizes the behavior Sb of the surrounding vehicles. When the recognition unit 233 recognizes that the surrounding vehicle made a signal as the behavior Sb of the surrounding vehicle, for example, the recognition unit 233 sets a score Sb5 and a weight wSb5 corresponding to the surrounding vehicle making a signal. When the recognition unit 233 recognizes that the surrounding vehicle did not make a signal as the behavior Sb of the surrounding vehicle, for example, the recognition unit 233 sets a score Sb6 and a weight wSb6 corresponding to the surrounding vehicle not making a signal (0 <Sb5<Sb6)。

[0022] The recognition unit 233 has a function of recognizing that a surrounding vehicle located around the vehicle 1 has pulled over to the vehicle 1 based on the detection results of the surrounding camera, radar, LiDAR, and sonar as the external sensor 2. Specifically, the recognition unit 233 estimates whether a surrounding vehicle located around the vehicle 1 has pulled over to the vehicle 1 based on the detection results of the surrounding vehicle located around the vehicle 1 by the external sensor 2 by using a model obtained by performing learning using teacher data, which is a data set of the detection results of the surrounding vehicle by the external sensor and a label indicating whether the surrounding vehicle detected by the external sensor has pulled over to the vehicle 1, and recognizes the behavior Sb of the surrounding vehicle. When the recognition unit 233 recognizes that the surrounding vehicle did not pull over to the vehicle 1 as the behavior Sb of the surrounding vehicle, for example, the recognition unit 233 sets a score Sb7 and a weight wSb7 corresponding to the surrounding vehicle not pulling over to the vehicle 1. When the recognition unit 233 recognizes that a surrounding vehicle has cut in close to the host vehicle 1 as the behavior Sb of the surrounding vehicle, for example, the recognition unit 233 sets a score Sb8 and a weight wSb8 corresponding to the surrounding vehicle cutting in close to the host vehicle 1 (0 <Sb7<Sb8)。 The recognition unit 233 also has a function of recognizing that a surrounding vehicle located around the vehicle 1 has run alongside the vehicle 1 based on the detection results of the surrounding camera, radar, LiDAR, and sonar as the external sensor 2. Specifically, the recognition unit 233 estimates whether a surrounding vehicle located around the vehicle 1 has run alongside the vehicle 1 based on the detection results of the surrounding vehicles located around the vehicle 1 by the external sensor 2 by using a model obtained by performing learning using teacher data, which is a data set of the detection results of the surrounding vehicles by the external sensor and labels indicating whether the surrounding vehicles detected by the external sensor have run alongside the vehicle 1, and recognizes the behavior Sb of the surrounding vehicles. When the recognition unit 233 recognizes that the surrounding vehicles did not run alongside the vehicle 1 as the behavior Sb of the surrounding vehicles, for example, the recognition unit 233 sets a score Sb9 and a weight wSb9 corresponding to the surrounding vehicles not running alongside the vehicle 1. When the recognition unit 233 recognizes that a nearby vehicle has run alongside the host vehicle 1 as the behavior Sb of the nearby vehicle, for example, the recognition unit 233 sets a score Sb10 and a weight wSb10 corresponding to the nearby vehicle running alongside the host vehicle 1 (0 <Sb9<Sb10)。

[0023] The recognition unit 233 has a function of recognizing that a surrounding vehicle has illuminated a high beam based on the light of the surrounding vehicle located around the vehicle 1 detected by an optical sensor serving as the external sensor 2. Specifically, the recognition unit 233 estimates whether or not a surrounding vehicle located around the vehicle 1 has illuminated a high beam based on the detection result of the optical sensor by using a model obtained by performing learning using teacher data, which is a data set of detection results of the optical sensor and labels indicating whether or not the light detected by the optical sensor is a high beam illuminated by the surrounding vehicle, and recognizes the behavior Sb of the surrounding vehicle. When the recognition unit 233 recognizes that the surrounding vehicle did not illuminate a high beam as the behavior Sb of the surrounding vehicle, for example, the recognition unit 233 sets a score Sb11 and a weight wSb11 corresponding to the surrounding vehicle not illuminating a high beam. When the recognition unit 233 recognizes that the surrounding vehicle illuminated a high beam as the behavior Sb of the surrounding vehicle, for example, the recognition unit 233 sets a score Sb12 and a weight wSb12 corresponding to the surrounding vehicle illuminating a high beam (0 <Sb11<Sb12)。 The recognition unit 233 has a function of recognizing that a vehicle surrounding the vehicle 1 has honked its horn based on a sound outside the vehicle 1 detected by a microphone as the external sensor 2. Specifically, the recognition unit 233 estimates whether a vehicle surrounding the vehicle 1 has honked its horn based on the detection result of the microphone by using a model obtained by performing learning using teacher data, which is a data set of a detection result of the microphone and a label indicating whether the sound detected by the microphone is a horn honked by the surrounding vehicle, and recognizes the behavior Sb of the surrounding vehicle. When the recognition unit 233 recognizes that the surrounding vehicle did not honk its horn as the behavior Sb of the surrounding vehicle, for example, the recognition unit 233 sets a score Sb13 and a weight wSb13 corresponding to the surrounding vehicle not honking its horn. When the recognition unit 233 recognizes that the surrounding vehicle honked its horn as the action Sb of the surrounding vehicle, for example, the recognition unit 233 sets a score Sb14 and a weight wSb14 corresponding to the surrounding vehicle honking its horn (0 <Sb13<Sb14)。 The recognition unit 233 also has a function of recognizing that a surrounding vehicle (following vehicle) located behind the vehicle 1 has closed the distance between the vehicle 1 and the vehicle 1 based on the detection results of the peripheral camera, radar, LiDAR, and sonar as the external sensor 2. Specifically, the recognition unit 233 estimates whether the surrounding vehicle (following vehicle) located behind the vehicle 1 has closed the distance between the vehicle 1 and the vehicle 1 based on the detection results of the surrounding vehicle (following vehicle) located behind the vehicle 1 by the external sensor 2, by using a model obtained by learning using teacher data, which is a data set of the detection results of the surrounding vehicle (following vehicle) by the external sensor and a label indicating whether the surrounding vehicle (following vehicle) detected by the external sensor has closed the distance between the vehicles, and recognizes the behavior Sb of the surrounding vehicle. When the recognition unit 233 recognizes that the surrounding vehicle (following vehicle) did not close the distance between itself and its own vehicle 1 as the behavior Sb of the surrounding vehicle, for example, the recognition unit 233 sets a score Sb15 and a weight wSb15 corresponding to the surrounding vehicle (following vehicle) not closing the distance between itself and its own vehicle 1. When the recognition unit 233 recognizes that the surrounding vehicle (following vehicle) closed the distance between itself and its own vehicle 1 as the behavior Sb of the surrounding vehicle, for example, the recognition unit 233 sets a score Sb16 and a weight wSb16 corresponding to the surrounding vehicle (following vehicle) closing the distance between itself and its own vehicle 1 (0 <Sb15<Sb16)。 The recognition unit 233 has a function of recognizing that the driver of the surrounding vehicle (parallel running vehicle) has seen the vehicle 1 based on an image including the surrounding vehicle (parallel running vehicle) located to the side of the vehicle 1 captured by a surrounding camera as the external sensor 2, and estimating that the eyes of the driver of the vehicle 1 and the driver of the parallel running vehicle may have met. Specifically, the recognition unit 233 estimates whether the driver of the surrounding vehicle (parallel running vehicle) has seen the vehicle 1 (whether the eyes of the driver of the vehicle 1 and the driver of the parallel running vehicle may have met) based on an image including the surrounding vehicle (parallel running vehicle) located to the side of the vehicle 1 captured by the surrounding camera, by using a model obtained by performing learning using teacher data that is a data set of images including the surrounding vehicle (parallel running vehicle) and labels indicating whether the driver of the surrounding vehicle (parallel running vehicle) has seen the vehicle 1 (whether the eyes of the driver of the vehicle 1 and the driver of the parallel running vehicle may have met), and recognizes the behavior Sb of the surrounding vehicle. When the recognition unit 233 recognizes that the driver of the nearby vehicle (vehicle running parallel to the vehicle) did not see the host vehicle 1 (there was no possibility that the eyes of the driver of the host vehicle 1 and the driver of the vehicle running parallel to the vehicle) as the behavior Sb of the nearby vehicle, for example, the recognition unit 233 sets a score Sb17 and a weight wSb17 corresponding to the fact that the driver of the nearby vehicle (vehicle running parallel to the vehicle) did not see the host vehicle 1 (there was no possibility that the eyes of the driver of the host vehicle 1 and the driver of the vehicle running parallel to the vehicle) as the behavior Sb of the nearby vehicle. When the recognition unit 233 recognizes that the driver of the nearby vehicle (vehicle running parallel to the vehicle) saw the host vehicle 1 (there was a possibility that the eyes of the driver of the host vehicle 1 and the driver of the vehicle running parallel to the vehicle) as the behavior Sb of the nearby vehicle, for example, the recognition unit 233 sets a score Sb18 and a weight wSb18 corresponding to the fact that the driver of the nearby vehicle (vehicle running parallel to the vehicle) saw the host vehicle 1 (there was a possibility that the eyes of the driver of the host vehicle 1 and the driver of the vehicle running parallel to the vehicle) as the behavior Sb of the nearby vehicle. <Sb17<Sb18)。

[0024] The recognition unit 233 recognizes the feature Sf of the surrounding vehicle based on the detection result of the external sensor 2. The recognition unit 233 has a function of recognizing that the seat position of the surrounding vehicle included in the image is high based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera as the external sensor 2. Specifically, the recognition unit 233 estimates whether the seat position of the surrounding vehicle included in the image is high or not based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera by using a model obtained by performing learning using teacher data that is a data set of an image including the surrounding vehicle and a label indicating whether the seat position of the surrounding vehicle included in the image is high or not, and recognizes the feature Sf of the surrounding vehicle. When the recognition unit 233 recognizes that the seat position of the surrounding vehicle in front of the vehicle 1 is not high as the feature Sf of the surrounding vehicle, for example, the recognition unit 233 sets a score Sf1 and a weight wSf1 corresponding to the seat position of the surrounding vehicle in front of the vehicle 1 is not high. When the recognition unit 233 recognizes that the seat position of the surrounding vehicle in front of the vehicle 1 is high as a feature Sf of the surrounding vehicle, for example, the recognition unit 233 sets a score Sf2 and a weight wSf2 corresponding to the seat position of the surrounding vehicle in front of the vehicle 1 being high (0 <Sf1<Sf2)。 The recognition unit 233 has a function of recognizing that the surrounding vehicle included in the image is large based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera as the external sensor 2. Specifically, the recognition unit 233 estimates whether the surrounding vehicle included in the image is large or not based on the image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera by using a model obtained by performing learning using teacher data, which is a data set of images including the surrounding vehicle and labels indicating whether the surrounding vehicle included in the image is large or not, and recognizes the characteristics Sf of the surrounding vehicle. When the recognition unit 233 recognizes that the surrounding vehicle in front of the vehicle 1 is not large as the characteristics Sf of the surrounding vehicle, the recognition unit 233 sets, for example, a score Sf3 and a weight wSf3 corresponding to the surrounding vehicle in front of the vehicle 1 being not large. When the recognition unit 233 recognizes that the surrounding vehicle in front of the vehicle 1 is large as the characteristics Sf of the surrounding vehicle, the recognition unit 233 sets, for example, a score Sf4 and a weight wSf4 corresponding to the surrounding vehicle in front of the vehicle 1 being large (0 <Sf3<Sf4)。 The recognition unit 233 has a function of recognizing that a large number of stickers are attached to the surrounding vehicles included in the image based on an image including the surrounding vehicles in front of the vehicle 1 captured by the forward camera as the external sensor 2. Specifically, the recognition unit 233 estimates whether or not a large number of stickers are attached to the surrounding vehicles included in the image based on an image including the surrounding vehicles in front of the vehicle 1 captured by the forward camera by using a model obtained by performing learning using teacher data, which is a data set of images including the surrounding vehicles and labels indicating whether or not a large number of stickers are attached to the surrounding vehicles included in the image, and recognizes the characteristics Sf of the surrounding vehicles. When the recognition unit 233 recognizes that a large number of stickers are not attached to the surrounding vehicles in front of the vehicle 1 as the characteristics Sf of the surrounding vehicles, for example, the recognition unit 233 sets a score Sf5 and a weight wSf5 corresponding to the fact that a large number of stickers are not attached to the surrounding vehicles in front of the vehicle 1. When the recognition unit 233 recognizes that a large number of stickers are attached to the surrounding vehicle in front of the vehicle 1 as the feature Sf of the surrounding vehicle, for example, the recognition unit 233 sets a score Sf6 and a weight wSf6 corresponding to the fact that a large number of stickers are attached to the surrounding vehicle in front of the vehicle 1 (0 <Sf5<Sf6)。

[0025] The recognition unit 233 has a function of recognizing that the model of the surrounding vehicle included in the image is a luxury car based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera as the external sensor 2. Specifically, the recognition unit 233 estimates whether the model of the surrounding vehicle included in the image is a luxury car based on the image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera by using a model obtained by performing learning using teacher data, which is a data set of an image including the surrounding vehicle and a label indicating whether the model of the surrounding vehicle included in the image is a luxury car or not, and recognizes the characteristics Sf of the surrounding vehicle. When the recognition unit 233 recognizes that the model of the surrounding vehicle in front of the vehicle 1 is not a luxury car as the feature Sf of the surrounding vehicle, for example, the recognition unit 233 sets a score Sf7 and a weight wSf7 corresponding to the model of the surrounding vehicle in front of the vehicle 1 not being a luxury car. When the recognition unit 233 recognizes that the model of the surrounding vehicle in front of the vehicle 1 is a luxury car as the feature Sf of the surrounding vehicle, for example, the recognition unit 233 sets a score Sf8 and a weight wSf8 corresponding to the model of the surrounding vehicle in front of the vehicle 1 being a luxury car (0 <Sf7<Sf8)。 The recognition unit 233 has a function of recognizing that the model of the surrounding vehicle included in the image is a light vehicle based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera as the external sensor 2. Specifically, the recognition unit 233 estimates whether the model of the surrounding vehicle included in the image is a light vehicle based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera by using a model obtained by performing learning using teacher data, which is a data set of an image including the surrounding vehicle and a label indicating whether the model of the surrounding vehicle included in the image is a light vehicle or not, and recognizes the characteristics Sf of the surrounding vehicle. When the recognition unit 233 recognizes that the model of the surrounding vehicle in front of the vehicle 1 is not a light vehicle as the feature Sf of the surrounding vehicle, for example, the recognition unit 233 sets a score Sf9 and a weight wSf9 corresponding to the model of the surrounding vehicle in front of the vehicle 1 being not a light vehicle. When the recognition unit 233 recognizes that the vehicle type of the surrounding vehicle in front of the vehicle 1 is a light vehicle as the surrounding vehicle feature Sf, for example, the recognition unit 233 sets a score Sf10 and a weight wSf10 corresponding to the vehicle type of the surrounding vehicle in front of the vehicle 1 being a light vehicle (0 <Sf9<Sf10)。

[0026] The recognition unit 233 has a function of recognizing that a green leaf mark is attached to a surrounding vehicle included in an image captured by a front camera as the external sensor 2, based on the image including the surrounding vehicle in front of the vehicle 1. Specifically, the recognition unit 233 estimates whether or not a green leaf mark is attached to a surrounding vehicle included in the image based on an image including the surrounding vehicle in front of the vehicle 1 captured by the front camera, by using a model obtained by performing learning using teacher data, which is a data set of an image including the surrounding vehicle and a label indicating whether or not the surrounding vehicle included in the image has a green leaf mark, and recognizes the characteristics Sf of the surrounding vehicle. When the recognition unit 233 recognizes that the surrounding vehicle in front of the vehicle 1 does not have a green leaf mark as the characteristics Sf of the surrounding vehicle, for example, the recognition unit 233 sets a score Sf11 and a weight wSf11 corresponding to the fact that the surrounding vehicle in front of the vehicle 1 does not have a green leaf mark. When the recognition unit 233 recognizes that the surrounding vehicle in front of the host vehicle 1 has a young leaf mark as a feature Sf of the surrounding vehicle, for example, the recognition unit 233 sets a score Sf12 and a weight wSf12 corresponding to the surrounding vehicle in front of the host vehicle 1 having a young leaf mark (0 <Sf11<Sf12)。 The recognition unit 233 has a function of recognizing that a silver mark is attached to a surrounding vehicle included in an image captured by a front camera as the external sensor 2, based on the image including the surrounding vehicle in front of the vehicle 1. Specifically, the recognition unit 233 estimates whether or not a silver mark is attached to a surrounding vehicle included in the image based on an image including the surrounding vehicle in front of the vehicle 1 captured by the front camera, by using a model obtained by performing learning using teacher data, which is a data set of an image including the surrounding vehicle and a label indicating whether or not the surrounding vehicle included in the image has a silver mark, and recognizes the characteristics Sf of the surrounding vehicle. When the recognition unit 233 recognizes that the surrounding vehicle in front of the vehicle 1 does not have a silver mark as the characteristics Sf of the surrounding vehicle, for example, the recognition unit 233 sets a score Sf13 and a weight wSf13 corresponding to the fact that the surrounding vehicle in front of the vehicle 1 does not have a silver mark. When the recognition unit 233 recognizes that the surrounding vehicle in front of the host vehicle 1 has a silver mark as a feature Sf of the surrounding vehicle, for example, the recognition unit 233 sets a score Sf14 and a weight wSf14 corresponding to the surrounding vehicle in front of the host vehicle 1 having a silver mark (0 <Sf13<Sf14)。

[0027] Further, the recognition unit 233 recognizes the characteristics Sd of the driver of the surrounding vehicle based on the detection result of the external sensor 2. The recognition unit 233 has a function of recognizing that the gender of the driver of the surrounding vehicle included in the image is male based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera as the external sensor 2. Specifically, the recognition unit 233 uses a model obtained by performing learning using teacher data, which is a data set of an image including the surrounding vehicle and a label indicating whether the gender of the driver of the surrounding vehicle included in the image is male or not, based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera, and recognizes the characteristics Sd of the driver of the surrounding vehicle. When the recognition unit 233 recognizes that the gender of the driver of the surrounding vehicle is male, for example, the recognition unit 233 sets a score Sd1 and a weight wSd1 corresponding to the gender of the driver of the surrounding vehicle being male. The recognition unit 233 has a function of recognizing that the gender of the driver of the surrounding vehicle included in the image is female, based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera as the external sensor 2. Specifically, the recognition unit 233 uses a model obtained by performing learning using teacher data, which is a data set of images including the surrounding vehicles and labels indicating whether the gender of the driver of the surrounding vehicle included in the image is female, based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera, and recognizes the characteristics Sd of the driver of the surrounding vehicle. When the recognition unit 233 recognizes that the gender of the driver of the surrounding vehicle is female, for example, the recognition unit 233 sets a score Sd2 and a weight wSd2 corresponding to the gender of the driver of the surrounding vehicle being female (0 <Sd2<Sd1)。 The recognition unit 233 has a function of recognizing that the driver of a surrounding vehicle included in an image is young, based on an image including the surrounding vehicle in front of the vehicle 1 captured by a forward camera as the external sensor 2. Specifically, the recognition unit 233 estimates whether the driver of the surrounding vehicle included in the image is young or not, based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera, by using a model obtained by learning using teacher data, which is a data set of images including the surrounding vehicle and labels indicating whether the driver of the surrounding vehicle included in the image is young or not, and recognizes the characteristics Sd of the driver of the surrounding vehicle. When the recognition unit 233 recognizes that the driver of the surrounding vehicle is young, for example, the recognition unit 233 sets a score Sd3 and a weight wSd3 corresponding to the driver of the surrounding vehicle being young (0 <Sd3)。 The recognition unit 233 has a function of recognizing that the driver of a surrounding vehicle included in an image is elderly based on an image including the surrounding vehicle in front of the vehicle 1 captured by a forward camera as the external sensor 2. Specifically, the recognition unit 233 estimates whether or not the driver of the surrounding vehicle included in the image is elderly based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera, by using a model obtained by learning using teacher data, which is a data set of images including the surrounding vehicles and labels indicating whether or not the driver of the surrounding vehicle included in the image is elderly, and recognizes the characteristics Sd of the driver of the surrounding vehicle. When the recognition unit 233 recognizes that the driver of the surrounding vehicle is elderly, for example, the recognition unit 233 sets a score Sd4 and a weight wSd4 corresponding to the driver of the surrounding vehicle being elderly (0 <Sd4)。

[0028] Further, the recognition unit 233 recognizes the characteristics Sp of the passenger of the surrounding vehicle based on the detection result of the external sensor 2. The recognition unit 233 has a function of recognizing that the gender of the passenger of the surrounding vehicle included in the image is male based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera as the external sensor 2. Specifically, the recognition unit 233 estimates whether the gender of the passenger of the surrounding vehicle included in the image is male based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera by using a model obtained by performing learning using teacher data, which is a data set of an image including the surrounding vehicle and a label indicating whether the gender of the passenger of the surrounding vehicle included in the image is male or not, and recognizes the characteristics Sp of the passenger of the surrounding vehicle. When the recognition unit 233 recognizes that the gender of the passenger of the surrounding vehicle is male, for example, the recognition unit 233 sets a score Sp1 and a weight wSp1 corresponding to the gender of the passenger of the surrounding vehicle being male (0 <Sp1)。 The recognition unit 233 has a function of recognizing that the gender of the passenger of the surrounding vehicle included in the image is female, based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera as the external sensor 2. Specifically, the recognition unit 233 uses a model obtained by performing learning using teacher data, which is a data set of images including the surrounding vehicles and labels indicating whether the gender of the passenger of the surrounding vehicle included in the image is female, based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera, and recognizes the characteristics Sp of the passenger of the surrounding vehicle. When the recognition unit 233 recognizes that the gender of the passenger of the surrounding vehicle is female, for example, the recognition unit 233 sets a score Sp2 and a weight wSp2 corresponding to the gender of the passenger of the surrounding vehicle being female (0 <Sp2)。 The recognition unit 233 has a function of recognizing that a passenger of a surrounding vehicle included in an image captured by a forward camera serving as the external sensor 2 is young, based on the image including the surrounding vehicle in front of the vehicle 1. Specifically, the recognition unit 233 estimates whether or not the passenger of the surrounding vehicle included in the image is young, based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera, by using a model obtained by performing learning using teacher data, which is a data set of images including the surrounding vehicle and labels indicating whether the passenger of the surrounding vehicle included in the image is young, and recognizes the characteristics Sp of the passenger of the surrounding vehicle. When the recognition unit 233 recognizes that the passenger of the surrounding vehicle is young, for example, the recognition unit 233 sets a score Sp3 and a weight wSp3 corresponding to the passenger of the surrounding vehicle being young (0 <Sp3)。 The recognition unit 233 has a function of recognizing that a passenger of a surrounding vehicle included in an image, which is captured by a forward camera serving as the external sensor 2, is an elderly person, based on an image including the surrounding vehicle in front of the vehicle 1. Specifically, the recognition unit 233 estimates whether or not a passenger of the surrounding vehicle included in the image is an elderly person based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera, by using a model obtained by performing learning using teacher data, which is a data set of images including the surrounding vehicles and labels indicating whether or not the passenger of the surrounding vehicle included in the image is an elderly person, and recognizes the characteristics Sp of the passenger of the surrounding vehicle. When the recognition unit 233 recognizes that the passenger of the surrounding vehicle is an elderly person, for example, the recognition unit 233 sets a score Sp4 and a weight wSp4 corresponding to the passenger of the surrounding vehicle being an elderly person (0 <Sp4)。 The recognition unit 233 has a function of recognizing that the driver and passenger of the surrounding vehicle included in the image are a couple based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera as the external sensor 2. Specifically, the recognition unit 233 uses a model obtained by performing learning using teacher data, which is a data set of an image including the surrounding vehicle and a label indicating whether the driver and passenger of the surrounding vehicle included in the image are a couple, based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera, and recognizes the characteristics Sp of the passenger of the surrounding vehicle. When the recognition unit 233 recognizes that the driver and passenger of the surrounding vehicle are not a couple, for example, the recognition unit 233 sets a score Sp5 and a weight wSp5 corresponding to the driver and passenger of the surrounding vehicle not being a couple. When the recognition unit 233 recognizes that the driver and the passenger of the nearby vehicle are a couple, for example, the recognition unit 233 sets a score Sp6 and a weight wSp6 corresponding to the driver and the passenger of the nearby vehicle being a couple (0 <Sp5<Sp6)。

[0029] Furthermore, the recognition unit 233 recognizes the state Eb of the host vehicle 1 based on the communication result of the communication device 4 (information acquired by the communication device 4). The recognition unit 233 has a function of recognizing that the host vehicle 1 is waiting a long time for a traffic light based on information (traffic light information) acquired by the communication device 4. Specifically, the recognition unit 233 determines that the host vehicle 1 is waiting a long time for a traffic light when the waiting time for a traffic light is longer than a predetermined threshold, for example, and recognizes that the host vehicle 1 is waiting a long time for a traffic light as the state Eb of the host vehicle 1. When the recognition unit 233 recognizes that the host vehicle 1 is not waiting a long time for a traffic light, for example, the recognition unit 233 sets a score Eb1 and a weight wEb1 corresponding to the host vehicle 1 not waiting a long time for a traffic light. When the recognition unit 233 recognizes that the host vehicle 1 is waiting a long time for a traffic light, for example, the recognition unit 233 sets a score Eb2 and a weight wEb2 corresponding to the host vehicle 1 being waiting a long time for a traffic light (0 <Eb1<Eb2)。 The recognition unit 233 has a function of recognizing that the road on which the host vehicle 1 is traveling is congested based on information (traffic congestion information) acquired by the communication device 4, etc. Specifically, the recognition unit 233 determines whether the road on which the host vehicle 1 is traveling is congested based on information (traffic congestion information) acquired by the communication device 4, etc., and when it determines that the road on which the host vehicle 1 is traveling is congested, recognizes that the road on which the host vehicle 1 is traveling is congested as the state Eb of the host vehicle 1. When the recognition unit 233 recognizes that the road on which the host vehicle 1 is traveling is not congested, for example, the recognition unit 233 sets a score Eb3 and a weight wEb3 corresponding to the road on which the host vehicle 1 is traveling is not congested. When the recognition unit 233 recognizes that the road on which the host vehicle 1 is traveling is congested, for example, the recognition unit 233 sets a score Eb4 and a weight wEb4 corresponding to the road on which the host vehicle 1 is traveling being congested (0 <Eb3<Eb4)。

[0030] The estimation unit 234 estimates the driver's state Ds based on an image including the face of the driver of the vehicle 1 captured by the driver monitor camera 3. The estimation unit 234 has a function of estimating whether or not the driver of the vehicle 1 included in the image is muttering (driver's state Ds) based on an image including the face of the driver of the vehicle 1 captured by the driver monitor camera 3 or a detection result of the microphone of the HMI 5. Specifically, the estimation unit 234 estimates whether or not the driver of the vehicle 1 included in the image is muttering (driver's state Ds) based on an image including the face of the driver of the vehicle 1 captured by the driver monitor camera 3 by using a model obtained by performing learning using teacher data, which is a data set of an image including the driver's face and a label indicating whether or not the driver included in the image is muttering. Alternatively, the estimation unit 234 estimates whether or not the driver of the host vehicle 1, who is emitting a sound included in the detection result, is muttering (driver state Ds), based on the detection result of the microphone of the HMI 5, by using a model obtained by performing learning using teacher data, which is a data set of detection results of sounds emitted by the driver and labels indicating whether or not the driver emitting the sound included in the detection result is muttering. When the estimation unit 234 estimates that the driver of the host vehicle 1 is not muttering, for example, the estimation unit 234 sets a score Ds1 and a weight wDs1 corresponding to the driver of the host vehicle 1 not muttering. When the estimation unit 234 estimates that the driver of the host vehicle 1 is muttering, for example, the estimation unit 234 sets a score Ds2 and a weight wDs2 corresponding to the driver of the host vehicle 1 muttering (0 <Ds1<Ds2)。 The estimation unit 234 has a function of estimating whether or not the driver of the vehicle 1 included in the image captured by the driver monitor camera 3 is glaring (driver state Ds) based on the image including the face of the driver of the vehicle 1. Specifically, the estimation unit 234 estimates whether or not the driver of the vehicle 1 included in the image captured by the driver monitor camera 3 is glaring (driver state Ds) based on the image including the face of the driver of the vehicle 1, which is captured by the driver monitor camera 3, by using a model obtained by performing learning using teacher data, which is a data set of an image including the driver's face and a label indicating whether or not the driver included in the image is glaring. When the estimation unit 234 estimates that the driver of the vehicle 1 is not glaring, for example, the estimation unit 234 sets a score Ds3 and a weight wDs3 corresponding to the driver of the vehicle 1 not glaring. When the estimation unit 234 estimates that the driver of the vehicle 1 is glaring, for example, the estimation unit 234 sets a score Ds4 and a weight wDs4 corresponding to the driver of the vehicle 1 glaring (0 <Ds3<Ds4)。

[0031] The estimation unit 234 has a function of estimating whether or not the driver of the vehicle 1 included in the image is yelling (driver state Ds) based on an image including the face of the driver of the vehicle 1 captured by the driver monitor camera 3 or a detection result of the microphone of the HMI 5. Specifically, the estimation unit 234 estimates whether or not the driver of the vehicle 1 included in the image is yelling (driver state Ds) based on an image including the face of the vehicle 1 captured by the driver monitor camera 3, by using a model obtained by learning using teacher data that is a data set of an image including the face of the driver of the vehicle 1 and a label indicating whether or not the driver included in the image is yelling. Alternatively, the estimation unit 234 estimates whether or not the driver of the vehicle 1 emitting the sound included in the detection result is yelling (driver state Ds) based on the detection result of the microphone of the HMI 5, by using a model obtained by learning using teacher data that is a data set of a detection result of a sound emitted by the driver and a label indicating whether or not the driver emitting the sound included in the detection result is yelling. When the estimation unit 234 estimates that the driver of the vehicle 1 is not yelling, for example, the estimation unit 234 sets a score Ds5 and a weight wDs5 corresponding to the driver of the vehicle 1 not yelling. When the estimation unit 234 estimates that the driver of the vehicle 1 is yelling, for example, the estimation unit 234 sets a score Ds6 and a weight wDs6 corresponding to the driver of the vehicle 1 yelling (0 <Ds5<Ds6)。 The estimation unit 234 has a function of estimating whether or not the driver of the vehicle 1 included in the image captured by the driver monitor camera 3 is in a hurry (driver state Ds) based on the image including the face of the driver of the vehicle 1. Specifically, the estimation unit 234 estimates whether or not the driver of the vehicle 1 included in the image captured by the driver monitor camera 3 is in a hurry (driver state Ds) based on the image including the face of the driver of the vehicle 1, which is captured by the driver monitor camera 3, by using a model obtained by performing learning using teacher data, which is a data set of an image including the face of the driver of the vehicle 1 and a label indicating whether or not the driver included in the image is in a hurry. When the estimation unit 234 estimates that the driver of the vehicle 1 is not in a hurry, for example, the estimation unit 234 sets a score Ds7 and a weight wDs7 corresponding to the driver of the vehicle 1 not in a hurry. When the estimation unit 234 estimates that the driver of the vehicle 1 is in a hurry, for example, the estimation unit 234 sets a score Ds8 and a weight wDs8 corresponding to the driver of the vehicle 1 being in a hurry (0 <Ds7<Ds8)。 The estimation unit 234 has a function of estimating whether or not the gaze of the driver of the vehicle 1 included in the image captured by the driver monitor camera 3 is stationary (driver state Ds) based on the image including the face of the driver of the vehicle 1. Specifically, the estimation unit 234 estimates whether or not the gaze of the driver of the vehicle 1 included in the image is stationary (driver state Ds) based on the image including the face of the driver of the vehicle 1 captured by the driver monitor camera 3 by using a model obtained by performing learning using teacher data, which is a data set of an image including the driver's face and a label indicating whether or not the gaze of the driver included in the image is stationary. When the estimation unit 234 estimates that the gaze of the driver of the vehicle 1 is not stationary, for example, the estimation unit 234 sets a score Ds9 and a weight wDs9 corresponding to the gaze of the driver of the vehicle 1 not stationary. When the estimation unit 234 estimates that the line of sight of the driver of the vehicle 1 is stationary, for example, the estimation unit 234 sets a score Ds10 and a weight wDs10 corresponding to the line of sight of the driver of the vehicle 1 being stationary (0 <Ds9<Ds10)。

[0032] The estimation unit 234 has a function of estimating, based on an image including a peripheral vehicle in front of the vehicle 1 captured by a forward camera as the external sensor 2, whether the state of the driver of the vehicle 1 is a state in which the driver of the vehicle can visually recognize the driver of the peripheral vehicle included in the image or a state in which the driver cannot visually recognize the driver of the peripheral vehicle included in the image. Specifically, the estimation unit 234 estimates, for example, based on the position of the face of the driver of the vehicle 1 included in the image captured by the driver monitor camera 3, the position of the forward camera in the vehicle 1, and the image including the peripheral vehicle in front of the vehicle 1 captured by the forward camera, whether the state of the driver of the vehicle 1 is a state in which the driver of the vehicle can visually recognize the driver of the peripheral vehicle included in the image (driver state Ds). When the estimation unit 234 estimates that the state of the driver of the vehicle 1 is a state in which the driver of the peripheral vehicle cannot visually recognize the driver of the peripheral vehicle, for example, the estimation unit 234 sets a score Ds11 and a weight wDs11 corresponding to the state in which the driver of the vehicle 1 is a state in which the driver of the vehicle 1 cannot visually recognize the driver of the peripheral vehicle. When the estimation unit 234 estimates that the state of the driver of the vehicle 1 is a state in which the driver of the surrounding vehicle can be seen, for example, the estimation unit 234 sets a score Ds12 and a weight wDs12 corresponding to the state in which the driver of the vehicle 1 is a state in which the driver of the surrounding vehicle can be seen (0 <Ds11<Ds12)。 The estimation unit 234 has a function of estimating whether the state of the driver of the host vehicle 1 is a state where the driver feels hot, based on the temperature outside or inside the host vehicle 1 detected by a temperature sensor as the external sensor 2. Specifically, for example, when the temperature outside or inside the host vehicle 1 detected by the temperature sensor is equal to or lower than a predetermined threshold, the estimation unit 234 estimates that the state of the driver of the host vehicle 1 is a state where the driver does not feel hot (driver state Ds). For example, when the temperature outside or inside the host vehicle 1 detected by the temperature sensor is higher than a predetermined threshold, the estimation unit 234 estimates that the state of the driver of the host vehicle 1 is a state where the driver feels hot (driver state Ds). When the estimation unit 234 estimates that the state of the driver of the host vehicle 1 is a state where the driver does not feel hot, for example, the estimation unit 234 sets a score Ds13 and a weight wDs13 corresponding to the state where the driver of the host vehicle 1 does not feel hot. When the estimation unit 234 estimates that the state of the driver of the vehicle 1 is a state in which the driver feels hot, for example, the estimation unit 234 sets a score Ds14 and a weight wDs14 corresponding to the state in which the driver of the vehicle 1 feels hot (0 <Ds13<Ds14)。

[0033] The estimation unit 234 estimates the facial expression and emotion De of the driver based on an image including the face of the driver of the vehicle 1 captured by the driver monitor camera 3. The estimation unit 234 has a function of estimating whether or not the driver of the vehicle 1 included in the image including the face of the driver of the vehicle 1 captured by the driver monitor camera 3 is angry (the driver's facial expression and emotion De). When the estimation unit 234 estimates that the driver of the vehicle 1 is not angry, for example, the estimation unit 234 sets a score De1 and a weight wDe1 corresponding to the driver of the vehicle 1 not being angry. When the estimation unit 234 estimates that the driver of the vehicle 1 is angry, for example, the estimation unit 234 sets a score De2 and a weight wDe2 corresponding to the driver of the vehicle 1 being angry (0 <De1<De2)。 The estimation unit 234 has a function of estimating whether or not the driver of the vehicle 1 dislikes something (the driver's facial expression and emotion De) contained in an image captured by the driver monitor camera 3, based on the image including the face of the driver of the vehicle 1. When the estimation unit 234 estimates that the driver of the vehicle 1 does not dislike something, for example, the estimation unit 234 sets a score De3 and a weight wDe3 corresponding to the driver of the vehicle 1 not disliking something. When the estimation unit 234 estimates that the driver of the vehicle 1 dislikes something, for example, the estimation unit 234 sets a score De4 and a weight wDe4 corresponding to the driver of the vehicle 1 disliking something (0 <De3<De4)。 The estimation unit 234 has a function of estimating whether or not the driver of the vehicle 1 is sad (the driver's facial expression and emotion De) based on an image including the face of the driver of the vehicle 1 captured by the driver monitor camera 3. When the estimation unit 234 estimates that the driver of the vehicle 1 is not sad, for example, the estimation unit 234 sets a score De5 and a weight wDe5 corresponding to the driver of the vehicle 1 not being sad. When the estimation unit 234 estimates that the driver of the vehicle 1 is sad, for example, the estimation unit 234 sets a score De6 and a weight wDe6 corresponding to the driver of the vehicle 1 being sad (0 <De5<De6)。 The estimation unit 234 has a function of estimating whether or not the driver of the host vehicle 1 has feelings of contempt or contempt (the driver's facial expression and emotion De) based on an image including the face of the driver of the host vehicle 1 captured by the driver monitor camera 3. When the estimation unit 234 estimates that the driver of the host vehicle 1 does not have feelings of contempt or contempt, for example, the estimation unit 234 sets a score De7 and a weight wDe7 corresponding to the driver of the host vehicle 1 not having feelings of contempt or contempt. When the estimation unit 234 estimates that the driver of the host vehicle 1 has feelings of contempt or contempt, for example, the estimation unit 234 sets a score De8 and a weight wDe8 corresponding to the driver of the host vehicle 1 having feelings of contempt or contempt (0 <De7<De8)。 The estimation unit 234 has a function of estimating whether or not the driver of the host vehicle 1 is surprised (the driver's facial expression and emotion De) based on an image including the face of the driver of the host vehicle 1 captured by the driver monitor camera 3. When the estimation unit 234 estimates that the driver of the host vehicle 1 is not surprised, for example, the estimation unit 234 sets a score De9 and a weight wDe9 corresponding to the driver of the host vehicle 1 not being surprised. When the estimation unit 234 estimates that the driver of the host vehicle 1 is surprised, for example, the estimation unit 234 sets a score De10 and a weight wDe10 corresponding to the driver of the host vehicle 1 being surprised (0 <De9<De10)。

[0034] The estimation unit 234 has a function of estimating whether or not the driver of the host vehicle 1 felt fear (the driver's facial expression and emotion De) based on an image including the face of the driver of the host vehicle 1 captured by the driver monitor camera 3. When the estimation unit 234 estimates that the driver of the host vehicle 1 does not feel fear, for example, the estimation unit 234 sets a score De11 and a weight wDe11 corresponding to the driver of the host vehicle 1 not feeling fear. When the estimation unit 234 estimates that the driver of the host vehicle 1 felt fear, for example, the estimation unit 234 sets a score De12 and a weight wDe12 corresponding to the driver of the host vehicle 1 feeling fear (0 <De11<De12)。 The estimation unit 234 has a function of estimating whether or not the driver of the vehicle 1 included in the image captured by the driver monitor camera 3 has panicked (the driver's facial expression and emotion De) based on the image including the face of the driver of the vehicle 1. Specifically, the estimation unit 234 estimates whether or not the driver of the vehicle 1 included in the image has panicked (the driver's facial expression and emotion De) based on the image including the face of the driver of the vehicle 1 captured by the driver monitor camera 3 by using a model obtained by performing learning using teacher data, which is a data set of an image including the driver's face and a label indicating whether or not the driver included in the image has panicked. When the estimation unit 234 estimates that the driver of the vehicle 1 is not in a panic, for example, the estimation unit 234 sets a score De13 and a weight wDe13 corresponding to the driver of the vehicle 1 not being in a panic. When the estimation unit 234 estimates that the driver of the vehicle 1 has entered a panicked state, for example, the estimation unit 234 sets a score De14 and a weight wDe14 corresponding to the driver of the vehicle 1 having entered a panicked state (0 <De13<De14)。 The estimation unit 234 has a function of estimating whether or not the driver of the vehicle 1 included in the image captured by the driver monitor camera 3 is in an irritated state (the driver's facial expression and emotion De) based on an image including the face of the driver of the vehicle 1 captured by the driver monitor camera 3. Specifically, the estimation unit 234 estimates whether or not the driver of the vehicle 1 included in the image is in an irritated state (the driver's facial expression and emotion De) based on an image including the face of the vehicle 1 captured by the driver monitor camera 3, by using a model obtained by performing learning using teacher data, which is a data set of an image including the driver's face and a label indicating whether or not the driver included in the image is in an irritated state. When the estimation unit 234 estimates that the driver of the vehicle 1 is not in an irritated state, for example, the estimation unit 234 sets a score De15 and a weight wDe15 corresponding to the driver of the vehicle 1 not being in an irritated state. When the estimation unit 234 estimates that the driver of the vehicle 1 is in an irritated state, for example, the estimation unit 234 sets a score De16 and a weight wDe16 corresponding to the driver of the vehicle 1 being in an irritated state (0 <De15<De16)。

[0035] In the example shown in FIG. 1 to FIG. 3, the tailgating symptom degree calculation unit 235 calculates the tailgating symptom degree Id of the host vehicle 1. In detail, the tailgating symptom degree calculation unit 235 calculates the tailgating symptom degree Id of the host vehicle 1 by adding the product of the scores and weights of a plurality of factors of the tailgating symptom of the host vehicle 1 (driver attribute information Dc, driver disease information Dd, driver state Ds, driver facial expression and emotion De, behavior Sb of the surrounding vehicle, characteristics Sf of the surrounding vehicle, characteristics Sd of the driver of the surrounding vehicle, characteristics Sp of the passenger of the surrounding vehicle, and state Eb of the host vehicle 1). Specifically, the tailgating symptom degree calculation unit 235 calculates the tailgating symptom degree Id of the host vehicle 1 based on the following formula (1). Id=Σ(Sbn×wSbn(n is 1-18))+Σ(Sfn×wSfn(n is 1-14))+Σ(Sdn×wSdn(n is 1-4))+Σ(Spn×wSpn(n is 1-6))+Σ(Ebn×wEbn(n is 1-4))+Σ(Dcn×wDcn(n is 1-20))+Σ(Ddn×wDdn(n is 1-6))+Σ(Dsn×wDsn(n is 1-14))+Σ(Den×wDen(n is 1-16)) (1)

[0036] The tailgating symptom determination unit 236 determines that the tailgating symptom of the host vehicle 1 exists when the tailgating symptom level Id of the host vehicle 1 calculated by the tailgating symptom level calculation unit 235 is greater than the threshold value Th (>0). That is, the tailgating symptom determination unit 236 determines whether or not the host vehicle 1 exists as a tailgating symptom based on a plurality of factors of the tailgating symptom of the host vehicle 1, which are information on the driver of the host vehicle 1 (driver attribute information Dc, driver illness information Dd, driver state Ds, driver facial expression and emotion De), information on the surrounding vehicles (behavior Sb of the surrounding vehicles, characteristics Sf of the surrounding vehicles, characteristics Sd of the driver of the surrounding vehicles, characteristics Sp of the passenger of the surrounding vehicles), and information on the host vehicle 1 (state Eb of the host vehicle). In the example shown in Figures 1 to 3, the aggressive driving sign determination unit 236 determines whether or not there are signs of aggressive driving in the vehicle 1 based on all of the information about the driver of the vehicle 1, the information about the surrounding vehicles, and the information about the vehicle 1, but in other examples, the aggressive driving sign determination unit 236 may determine whether or not there are signs of aggressive driving in the vehicle 1 based on at least any of the information about the driver of the vehicle 1, the information about the surrounding vehicles, and the information about the vehicle 1.

[0037] In the example shown in Figures 1 to 3, the information about the driver of the vehicle 1 includes the driver attribute information Dc, the driver disease information Dd, the driver state Ds, and the driver's facial expression and emotion De, but in other examples, the information about the driver of the vehicle 1 does not need to include all of the driver attribute information Dc, the driver disease information Dd, the driver state Ds, and the driver's facial expression and emotion De, and in still another example, the information about the driver of the vehicle 1 may further include information other than the driver attribute information Dc, the driver disease information Dd, the driver state Ds, and the driver's facial expression and emotion De. In still another example, the information about the driver of the vehicle 1 may include information other than the driver attribute information Dc, the driver disease information Dd, the driver state Ds, and the driver's facial expression and emotion De instead of any of the driver attribute information Dc, the driver disease information Dd, the driver state Ds, and the driver's facial expression and emotion De. In the example shown in FIG. 1 to FIG. 3, the information on the surrounding vehicle includes the behavior Sb of the surrounding vehicle, the characteristics Sf of the surrounding vehicle, the characteristics Sd of the driver of the surrounding vehicle, and the characteristics Sp of the passenger of the surrounding vehicle, but in another example, the information on the surrounding vehicle does not need to include all of the behavior Sb of the surrounding vehicle, the characteristics Sf of the surrounding vehicle, the characteristics Sd of the driver of the surrounding vehicle, and the characteristics Sp of the passenger of the surrounding vehicle, and in yet another example, the information on the surrounding vehicle may further include information other than the behavior Sb of the surrounding vehicle, the characteristics Sf of the surrounding vehicle, the characteristics Sd of the driver of the surrounding vehicle, and the characteristics Sp of the passenger of the surrounding vehicle. In yet another example, the information on the surrounding vehicle may include information other than the behavior Sb of the surrounding vehicle, the characteristics Sf of the surrounding vehicle, the characteristics Sd of the driver of the surrounding vehicle, and the characteristics Sp of the passenger of the surrounding vehicle, instead of any of the behavior Sb of the surrounding vehicle, the characteristics Sf of the surrounding vehicle, the characteristics Sd of the driver of the surrounding vehicle, and the characteristics Sp of the passenger of the surrounding vehicle. 1 to 3, the information about the host vehicle 1 includes the state Eb of the host vehicle 1, but in other examples, the information about the host vehicle 1 may further include information other than the state Eb of the host vehicle 1. In still other examples, the information about the host vehicle 1 may include information other than the state Eb of the host vehicle 1 instead of the state Eb of the host vehicle 1.

[0038] 1 to 3, when the aggressive driving symptom determination unit 236 determines that there is a symptom of aggressive driving of the vehicle 1, the tranquilization processing device 30 executes a process for tranquilizing the driver of the vehicle 1. The tranquilization processing device 30 is configured by a microcomputer including a communication interface 31, a memory 32, and a processor 33. The communication interface 31, the memory 32, and the processor 33 are connected via a signal line 34. 1 to 3, the sedation processing device 30 includes one processor 33, but in other examples, the sedation processing device 30 may include multiple processors. Also, in the examples shown in Fig. 1 to 3, the sedation processing device 30 is configured by one ECU, but in other examples, the sedation processing device 30 may be configured by multiple ECUs. In still other examples, the road rage symptom estimation device 20 and the sedation processing device 30 may be configured by one ECU.

[0039] In the example shown in FIG. 1 to FIG. 3, the processor 33 has a cause prediction unit 331 and a calming processing unit 332. When the tailgating sign determination unit 236 determines that the host vehicle 1 has a sign of tailgating, the cause prediction unit 331 predicts the cause of the sign of tailgating of the host vehicle 1. When the tailgating sign determination unit 236 determines that the host vehicle 1 has a sign of tailgating, the calming processing unit 332 executes a process to calm the driver of the host vehicle 1. In detail, when the tailgating sign determination unit 236 determines that the host vehicle 1 has a sign of tailgating, the calming processing unit 332 executes a process to output the cause of the sign of tailgating of the host vehicle 1 predicted by the cause prediction unit 331 to the HMI 5.

[0040] In the first specific example, the recognition unit 233 has a function of recognizing that the number of lanes on which the vehicle 1 can travel is reduced due to the presence of road construction, parked vehicles, etc., based on an image in front of the vehicle 1 captured by a front camera as the external sensor 2. For example, the recognition unit 233 estimates whether the number of lanes on which the vehicle 1 can travel, included in the image, is reduced, based on an image in front of the vehicle 1 captured by the front camera, by using a model obtained by performing learning using teacher data, which is a data set of an image including lanes on which the vehicle 1 can travel, and a label included in the image indicating whether the number of lanes on which the vehicle 1 can travel is reduced. In addition, as described above, the recognition unit 233 has a function of recognizing that a peripheral vehicle included in the image has cut in front of the vehicle 1 (behavior Sb of the peripheral vehicle) based on an image including peripheral vehicles in front of the vehicle 1 captured by the front camera as the external sensor 2. Furthermore, as described above, the recognition unit 233 has a function of recognizing that the seat position of the surrounding vehicle included in the image, which includes the surrounding vehicle in front of the vehicle 1 and is captured by the forward camera as the external sensor 2, is high (characteristic Sf of the surrounding vehicle). The cause prediction unit 331 has a function of recognizing that the reason why the surrounding vehicle included in the image cuts in front of the vehicle 1 is, for example, due to the presence of road construction, parked vehicles, etc., based on an image including surrounding vehicles in front of the vehicle 1 captured by the forward camera as the external sensor 2, is that the number of lanes on which the vehicle 1 can travel is reduced due to the presence of road construction, parked vehicles, etc. Specifically, the cause prediction unit 331 uses a model obtained by performing learning using teacher data, which is a data set of, for example, images including drivable lanes and surrounding vehicles cutting in front, and labels indicating whether the reason why the surrounding vehicles cut in front is, for example, due to the presence of road construction, parked vehicles, etc., based on an image in front of the vehicle 1 captured by the forward camera (an image including surrounding vehicles cutting in front of the vehicle 1), estimates whether the reason why the surrounding vehicle included in the image cuts in front of the vehicle 1 is, for example, due to the presence of road construction, parked vehicles, etc. Furthermore, the cause prediction unit 331 estimates that at the time when the driver of a nearby vehicle in front of the vehicle 1, which has a high seat position, notices the presence of, for example, road construction, parked vehicles on the road, etc., the driver of the vehicle 1 may not have noticed the presence of, for example, road construction, parked vehicles on the road, etc. The calming processing unit 332 causes the HMI 5 to output (e.g., display) a message such as "The number of lanes is decreasing ahead," thereby conveying to the driver of the host vehicle 1 that the cutting in of the surrounding vehicle in front of the host vehicle 1 is not intended to harass the driver of the host vehicle 1, and urging the driver of the host vehicle 1 to accept the behavior of the surrounding vehicle in front of the host vehicle 1.

[0041] In the second specific example, the recognition unit 233 has a function of recognizing that the speed of the surrounding vehicle included in the image is slow (behavior of the surrounding vehicle Sb) based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera as the external sensor 2, as described above. Also, the recognition unit 233 has a function of recognizing that the vehicle type of the surrounding vehicle included in the image is a light vehicle (characteristic Sf of the surrounding vehicle) based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera as the external sensor 2, as described above. Furthermore, the recognition unit 233 has a function of recognizing that the surrounding vehicle included in the image has a silver mark (characteristic Sf of the surrounding vehicle) based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera as the external sensor 2, as described above. Also, the recognition unit 233 has a function of recognizing that the driver of the surrounding vehicle included in the image is an elderly person (characteristic Sd of the driver of the surrounding vehicle) based on an image including the surrounding vehicle in front of the vehicle 1 captured by the forward camera as the external sensor 2, as described above. Based on an image including surrounding vehicles in front of the vehicle 1 captured by the forward camera as the external sensor 2, when the recognition unit 233 recognizes that the speed of the surrounding vehicle included in the image is slow (behavior of the surrounding vehicle Sb), that the model of the surrounding vehicle included in the image is a light vehicle (characteristics of the surrounding vehicle Sf), that the surrounding vehicle included in the image has a silver mark (characteristics of the surrounding vehicle Sf), and that the driver of the surrounding vehicle included in the image is elderly (driver of the surrounding vehicle Sd), the cause prediction unit 331 estimates that the driver of the surrounding vehicle in front of the vehicle 1 is elderly and therefore has poor cognitive ability, judgment ability, and operating ability, and as a result, the surrounding vehicle in front of the vehicle 1 is traveling at a slow speed. Specifically, the cause prediction unit 331 uses a model obtained by learning using teacher data, which is a dataset of images including a light vehicle marked with a silver mark and traveling at a slow speed, where the image indicates that the driver is elderly, and a label indicating whether the light vehicle is traveling at a slow speed because the driver of the light vehicle is elderly and has poor cognitive abilities, judgment abilities, and operating abilities, and estimates that the driver of the light vehicle in front of the vehicle 1 is elderly and has poor cognitive abilities, judgment abilities, and operating abilities, and therefore the light vehicle is traveling at a slow speed, based on an image of the front of the vehicle 1 captured by the forward camera (an image including a light vehicle in front of the vehicle 1 traveling at a slow speed). The calming processing unit 332 causes the HMI 5 to output (for example, display) a message saying, "The vehicle ahead is driven by an elderly person. Please keep a safe distance and drive with consideration."

[0042] In the third specific example, as described above, the recognition unit 233 has a function of recognizing that a surrounding vehicle (oncoming vehicle) has illuminated a high beam (behavior of surrounding vehicle Sb) based on the light of the surrounding vehicle (oncoming vehicle) located around the vehicle 1 detected by an optical sensor serving as the external sensor 2. Also, as described above, the recognition unit 233 has a function of recognizing, based on an image including the surrounding vehicle in front of the vehicle 1 captured by a forward camera serving as the external sensor 2, that the model of the surrounding vehicle included in the image is a luxury car (characteristics Sf of surrounding vehicle). The cause prediction unit 331 has a function of recognizing that the host vehicle 1 does not behave in a manner that would cause a surrounding vehicle (oncoming vehicle) in front of the host vehicle 1 to irradiate a high beam to the host vehicle 1, based on an image in front of the host vehicle 1 captured by a forward camera as the external sensor 2. Specifically, the cause prediction unit 331 estimates whether the host vehicle 1 has behaved in a manner that would cause a surrounding vehicle (oncoming vehicle) in front of the host vehicle 1 to irradiate a high beam to the host vehicle 1, based on an image in front of the host vehicle 1 captured by the forward camera, by using a model obtained by performing learning using teacher data, which is a data set of an image in front of the learning host vehicle captured by the forward camera and a label included in the image indicating whether or not the learning host vehicle has behaved in a manner that would cause an oncoming vehicle to irradiate a high beam. Furthermore, when the recognition unit 233 recognizes that a nearby vehicle (oncoming vehicle) located around the vehicle 1 has illuminated its high beam (behavior of the nearby vehicle Sb), that the nearby vehicle (oncoming vehicle) is a luxury vehicle (characteristics Sf of the nearby vehicle), and that the vehicle 1 is not behaving in a manner that would cause the nearby vehicle (oncoming vehicle) to illuminate its high beam at the vehicle 1, the cause prediction unit 331 estimates that the cause of the nearby vehicle (oncoming vehicle)'s high beam may be a malfunction of the AHB (Auto High Beam) of the nearby vehicle (oncoming vehicle). The sedation processing unit 332 causes the HMI 5 to output (for example, display) a message saying "The AHB of the oncoming vehicle may be malfunctioning."

[0043] FIG. 4 is a flowchart for explaining an example of a process executed in the host vehicle 1 to which the aggressive driving sign estimation device 20 of the first embodiment is applied. In the example shown in FIG. 4, in step S1, the recognition unit 233 executes recognition of the host vehicle 1 and surrounding vehicles located around the host vehicle 1. In detail, in step S1A, the recognition unit 233 recognizes the behavior Sb of the surrounding vehicle. In addition, the recognition unit 233 sets a score and a weight for the behavior Sb of the surrounding vehicle. In step S1B, the recognition unit 233 recognizes the features Sf of the surrounding vehicles. In addition, the recognition unit 233 sets a score and a weight for the features Sf of the surrounding vehicles. In step S1C, the recognition unit 233 recognizes the characteristics Sd of the driver of the surrounding vehicle. In addition, the recognition unit 233 sets a score and a weight for the characteristics Sd of the driver of the surrounding vehicle. In step S1D, the recognition unit 233 recognizes the characteristics Sp of the passenger of the surrounding vehicle. In addition, the recognition unit 233 sets a score and a weight for the characteristics Sp of the passenger of the surrounding vehicle. In step S1E, the recognition unit 233 recognizes the state Eb of the host vehicle 1. In addition, the recognition unit 233 sets a score and a weight for the state Eb of the host vehicle 1. In step S2, the acquisition unit 231 acquires information about the driver of the host vehicle 1, and the estimation unit 234 executes estimation regarding the driver of the host vehicle 1. In detail, in step S2A, the acquisition unit 231 acquires the driver attribute information Dc. In addition, the acquisition unit 231 sets a score and a weight for the driver attribute information Dc. In step S2B, the acquisition unit 231 acquires the driver disease information Dd. In addition, the acquisition unit 231 sets a score and a weight for the driver disease information Dd. In step S2C, the estimation unit 234 estimates the driver's state Ds. In addition, the estimation unit 234 sets a score and a weight for the driver's state Ds. In step S2D, the estimation unit 234 estimates the driver's facial expression and emotion De. In addition, the estimation unit 234 sets scores and weights for the driver's facial expression and emotion De. In the example shown in FIG. 4, steps S1 and S2 are executed in parallel, but in other examples, steps S1 and S2 may be executed in series (in time series). In the example shown in Figure 4, in step S3, the aggressive driving symptom degree calculation unit 235 calculates the aggressive driving symptom degree Id of the vehicle 1 by adding up the product of the scores and weights of each of the multiple factors of the aggressive driving symptoms of the vehicle 1 (driver attribute information Dc, driver illness information Dd, driver condition Ds, driver's facial expression and emotions De, behavior Sb of surrounding vehicles, characteristics Sf of surrounding vehicles, characteristics Sd of the drivers of surrounding vehicles, characteristics Sp of passengers of surrounding vehicles, and state Eb of the vehicle 1). In step S4, the aggressive driving symptom determination unit 236 determines whether or not the aggressive driving symptom level Id of the host vehicle 1 calculated in step S3 is greater than a threshold value Th. When the symptom level Id of aggressive driving of the host vehicle 1 calculated in step S3 is greater than the threshold value Th, the aggressive driving symptom determination unit 236 determines that there is a symptom of aggressive driving of the host vehicle 1, and proceeds to step S5. If the symptom level Id of aggressive driving of the host vehicle 1 calculated in step S3 is equal to or less than the threshold value Th, the aggressive driving symptom determination unit 236 determines that there is no symptom of aggressive driving of the host vehicle 1, and returns to steps S1 and S2. In step S5, the cause prediction unit 331 predicts the cause of the symptoms of aggressive driving of the host vehicle 1, and the pacification processing unit 332 executes processing to calm the driver of the host vehicle 1.

[0044] FIG. 5 is a flowchart for explaining an example of a process performed by the estimation unit 234 to estimate the facial expression and emotion De of the driver in step S2D of FIG. In the example shown in FIG. 5, the acquisition unit 231 acquires image data including the face of the driver of the vehicle 1 captured by the driver monitor camera 3 in step S10. In step S11, the detection unit 232 detects the position of the face of the driver of the vehicle 1 based on the image data including the face of the driver of the vehicle 1 acquired in step S10. In step S12, the detection unit 232 detects facial parts (e.g., eyes, mouth, etc.) of the driver of the vehicle 1, items worn by the driver of the vehicle 1 (e.g., glasses, sunglasses, mask, etc.) and facial features of the driver of the vehicle 1 based on image data including the face of the driver of the vehicle 1 acquired in step S10. In step S2D, the estimation unit 234 estimates the driver's facial expression and emotions De (irritation, panic, etc.) based on the image data including the face of the driver of the vehicle 1 acquired in step S10 and the detection results of steps S11 and S12.

[0045] <Second embodiment> The vehicle 1 to which the tailgating sign estimation device 20 of the second embodiment is applied is configured in the same manner as the vehicle 1 to which the tailgating sign estimation device 20 of the first embodiment is applied, as described above, except for the points described below. As described above, in the host vehicle 1 to which the first embodiment of the tailgating sign estimation device 20 is applied, the tailgating sign determination unit 236 determines that the host vehicle 1 has signs of tailgating when the tailgating sign level Id of the host vehicle 1 calculated by the tailgating sign level calculation unit 235 is greater than the threshold value Th. On the other hand, in the host vehicle 1 to which the road rage sign estimation device 20 of the second embodiment is applied, the road rage sign determination unit 236 has a function of determining that the host vehicle 1 has a sign of road rage when the road rage sign degree Id of the host vehicle 1 calculated by the road rage sign degree calculation unit 235 is greater than the threshold value Th in the same manner as the host vehicle 1 to which the road rage sign estimation device 20 of the first embodiment is applied. In other words, the road rage sign determination unit 236 has a function of executing a determination (first determination) of the presence or absence of a sign of road rage of the host vehicle 1 based on the road rage sign degree Id of the host vehicle 1 calculated by the road rage sign degree calculation unit 235. In addition, in the host vehicle 1 to which the tailgating sign estimation device 20 of the second embodiment is applied, the tailgating sign determination unit 236 uses a model obtained by performing learning using teacher data, which is a data set of factors of tailgating signs (driver attribute information Dc, driver's disease information Dd, driver state Ds, driver's facial expression and emotion De, behavior Sb of surrounding vehicles, characteristics Sf of surrounding vehicles, characteristics Sd of drivers of surrounding vehicles, characteristics Sp of passengers of surrounding vehicles, and state Eb of the host vehicle 1) obtained by the acquisition unit 231, recognition unit 233, and estimation unit 234, and labels indicating whether or not there are tailgating signs. The tailgating sign determination unit 236 has a function of performing a determination (second determination) of the presence or absence of tailgating signs of the host vehicle 1 based on multiple factors of the tailgating signs of the host vehicle 1 obtained by the acquisition unit 231, recognition unit 233, and estimation unit 234 by using the model. Furthermore, in the host vehicle 1 to which the aggressive driving sign estimation device 20 of the second embodiment is applied, the aggressive driving sign determination unit 236 has a function of switching between the first determination and the second determination.

[0046] <Third embodiment> The vehicle 1 to which the tailgating sign estimation device 20 of the third embodiment is applied is configured in the same manner as the vehicle 1 to which the tailgating sign estimation device 20 of the first embodiment is applied, as described above, except for the points described below. In the host vehicle 1 to which the road rage sign estimation device 20 of the third embodiment is applied, similarly to the host vehicle 1 to which the road rage sign estimation device 20 of the first embodiment is applied, the road rage sign determination unit 236 has a function of determining that the host vehicle 1 has a sign of road rage when the sign level Id of road rage of the host vehicle 1 calculated by the road rage sign degree calculation unit 235 is greater than the threshold value Th. In other words, the road rage sign determination unit 236 has a function of executing a determination (first determination) of the presence or absence of a sign of road rage of the host vehicle 1 based on the sign level Id of road rage of the host vehicle 1 calculated by the road rage sign degree calculation unit 235. In addition, in the host vehicle 1 to which the third embodiment of the tailgating sign estimation device 20 is applied, the tailgating sign determination unit 236 has a function of determining that the host vehicle 1 has a sign of tailgating (executing a third determination) when multiple factors of the tailgating sign of the host vehicle 1 (driver attribute information Dc, driver's disease information Dd, driver state Ds, driver's facial expression and emotion De, behavior Sb of the surrounding vehicles, characteristics Sf of the surrounding vehicles, characteristics Sd of the drivers of the surrounding vehicles, characteristics Sp of the passengers of the surrounding vehicles, and the state Eb of the host vehicle) correspond to a predetermined combination pattern. Furthermore, in the host vehicle 1 to which the third embodiment of the tailgating sign estimation device 20 is applied, the tailgating sign determination unit 236 has a function of switching between the first determination and the third determination.

[0047] In the first example of the above-mentioned predetermined combination pattern, the tailgating symptom determination unit 236 judges whether or not the driver of the host vehicle 1 is muttering to himself / herself (driver state Ds) and whether or not at least one of the following conditions is met: a nearby vehicle has cut in front of the host vehicle 1; the speed of the nearby vehicle (preceding vehicle) in front of the host vehicle 1 is slow; the nearby vehicle has not signaled; the nearby vehicle has cut in front of the host vehicle 1; the nearby vehicle has run alongside the host vehicle 1; the nearby vehicle has illuminated its high beam; the nearby vehicle has honked its horn; the nearby vehicle (following vehicle) behind the host vehicle 1 has closed the distance between the host vehicle 1 and the driver of the nearby vehicle (running alongside the host vehicle) has seen the host vehicle 1 (there is a possibility that the eyes of the driver of the host vehicle 1 and the driver of the running alongside the vehicle) (behavior of the nearby vehicles Sb), and the seat position of the nearby vehicle in front of the host vehicle 1 is high; the nearby vehicle in front of the host vehicle 1 is large; When at least one of the following applies: the model of the surrounding vehicle in front of the vehicle 1 is a luxury car, the model of the surrounding vehicle in front of the vehicle 1 is a light car, the surrounding vehicle in front of the vehicle 1 has a young leaf mark, and the surrounding vehicle in front of the vehicle 1 has a silver mark (characteristics Sf of surrounding vehicles), the gender of the driver of the vehicle 1 is male (driver attribute information Dc), the age of the driver of the vehicle 1 is in his / her teens, twenties, or thirties (driver attribute information Dc), and when any of the following applies: the driver of the vehicle 1 is angry, the driver of the vehicle 1 is disgusted, the driver of the vehicle 1 has feelings of contempt or insult, the driver of the vehicle 1 is surprised, the expression of the driver of the vehicle 1 is in a state of panic, and the expression of the driver of the vehicle 1 is in a state of irritation (driver's expressions and emotions De).

[0048] In the second example of the above-mentioned predetermined combination pattern, the tailgating symptom determination unit 236 judges whether or not the driver of the host vehicle 1 is glaring (driver state Ds) and whether or not at least one of the following conditions is met: a nearby vehicle has cut in front of the host vehicle 1; the speed of the nearby vehicle (preceding vehicle) in front of the host vehicle 1 is slow; the nearby vehicle has not given a signal; the nearby vehicle has cut in front of the host vehicle 1; the nearby vehicle has run alongside the host vehicle 1; the nearby vehicle has irradiated its high beam; the nearby vehicle has honked its horn; the nearby vehicle (following vehicle) behind the host vehicle 1 has closed the distance between the host vehicle 1 and the driver of the nearby vehicle (running alongside the host vehicle) has seen the host vehicle 1 (the eyes of the driver of the host vehicle 1 may have met the eyes of the driver of the nearby vehicle) (behavior of nearby vehicles Sb), and If at least one of the following applies: the model of the surrounding vehicle in front of both vehicles 1 is a luxury car, the model of the surrounding vehicle in front of vehicle 1 is a light car, the surrounding vehicle in front of vehicle 1 has a young leaf mark, and the surrounding vehicle in front of vehicle 1 has a silver mark (characteristics Sf of surrounding vehicles), the gender of the driver of vehicle 1 is male (driver attribute information Dc), the age of the driver of vehicle 1 is in his teens, twenties, or thirties (driver attribute information Dc), and if any of the following applies: the driver of vehicle 1 is angry, the driver of vehicle 1 is disgusted, the driver of vehicle 1 has feelings of contempt or insult, the driver of vehicle 1 is surprised, the driver of vehicle 1 has a panicked expression, and the driver of vehicle 1 has an irritated expression (driver's expression and emotions De).

[0049] In the third example of the above-mentioned predetermined combination pattern, the tailgating symptom determination unit 236 judges that the driver of the host vehicle 1 is yelling (driver state Ds) and that at least one of the following conditions is met: a nearby vehicle has cut in front of the host vehicle 1; the speed of the nearby vehicle (preceding vehicle) in front of the host vehicle 1 is slow; the nearby vehicle has not signaled; the nearby vehicle has cut in front of the host vehicle 1; the nearby vehicle has run alongside the host vehicle 1; the nearby vehicle has irradiated its high beam; the nearby vehicle has honked its horn; a nearby vehicle (following vehicle) behind the host vehicle 1 has closed the distance between the host vehicle 1 and the driver of the nearby vehicle (running alongside the host vehicle) has seen the host vehicle 1 (the eyes of the driver of the host vehicle 1 may have met the eyes of the driver of the nearby vehicle) (behavior of nearby vehicles Sb), and If at least one of the following applies: the model of the surrounding vehicle in front of both vehicles 1 is a luxury car, the model of the surrounding vehicle in front of vehicle 1 is a light car, the surrounding vehicle in front of vehicle 1 has a young leaf mark, and the surrounding vehicle in front of vehicle 1 has a silver mark (characteristics Sf of surrounding vehicles), the gender of the driver of vehicle 1 is male (driver attribute information Dc), the age of the driver of vehicle 1 is in his teens, twenties, or thirties (driver attribute information Dc), and if any of the following applies: the driver of vehicle 1 is angry, the driver of vehicle 1 is disgusted, the driver of vehicle 1 has feelings of contempt or insult, the driver of vehicle 1 is surprised, the driver of vehicle 1 has a panicked expression, and the driver of vehicle 1 has an irritated expression (driver's expression and emotions De).

[0050] In the fourth example of the above-mentioned predetermined combination pattern, the tailgating symptom determination unit 236 judges that the driver of the host vehicle 1 is in a hurry (driver state Ds) and that at least one of the following conditions applies: a nearby vehicle has cut in front of the host vehicle 1; the speed of the nearby vehicle (preceding vehicle) in front of the host vehicle 1 is slow; the nearby vehicle has not signaled; the nearby vehicle has cut in front of the host vehicle 1; the nearby vehicle has run alongside the host vehicle 1; the nearby vehicle has irradiated its high beam; the nearby vehicle has honked its horn; a nearby vehicle (following vehicle) behind the host vehicle 1 has closed the distance between the host vehicle 1 and the driver of the nearby vehicle (running alongside the host vehicle) has seen the host vehicle 1 (the eyes of the driver of the host vehicle 1 may have met the eyes of the driver of the nearby vehicle) (behavior of the nearby vehicles Sb), and If at least one of the following applies: the model of the surrounding vehicle in front of both vehicles 1 is a luxury car, the model of the surrounding vehicle in front of vehicle 1 is a light car, the surrounding vehicle in front of vehicle 1 has a young leaf mark, and the surrounding vehicle in front of vehicle 1 has a silver mark (characteristics Sf of surrounding vehicles), the gender of the driver of vehicle 1 is male (driver attribute information Dc), the age of the driver of vehicle 1 is in his teens, twenties, or thirties (driver attribute information Dc), and if any of the following applies: the driver of vehicle 1 is angry, the driver of vehicle 1 is disgusted, the driver of vehicle 1 has feelings of contempt or insult, the driver of vehicle 1 is surprised, the driver of vehicle 1 has a panicked expression, and the driver of vehicle 1 has an irritated expression (driver's expression and emotions De).

[0051] In the fifth example of the above-mentioned predetermined combination pattern, the tailgating symptom determination unit 236 judges that the driver of the host vehicle 1 is stationary (driver state Ds) and that at least one of the following conditions is met: a surrounding vehicle has cut in front of the host vehicle 1; the speed of the surrounding vehicle (preceding vehicle) in front of the host vehicle 1 is slow; the surrounding vehicle has not signaled; the surrounding vehicle has cut in front of the host vehicle 1; the surrounding vehicle has run alongside the host vehicle 1; the surrounding vehicle has illuminated its high beam; the surrounding vehicle has honked its horn; the surrounding vehicle (following vehicle) behind the host vehicle 1 has closed the distance between the host vehicle 1 and the driver of the surrounding vehicle (running alongside the host vehicle) has seen the host vehicle 1 (the eyes of the driver of the host vehicle 1 may have met the eyes of the driver of the running alongside the host vehicle) (behavior of the surrounding vehicles Sb), and the seat position of the surrounding vehicle in front of the host vehicle 1 is high; the surrounding vehicle in front of the host vehicle 1 is large; When at least one of the following applies: the model of the surrounding vehicle in front of the vehicle 1 is a luxury car, the model of the surrounding vehicle in front of the vehicle 1 is a light car, the surrounding vehicle in front of the vehicle 1 has a young leaf mark, and the surrounding vehicle in front of the vehicle 1 has a silver mark (characteristics Sf of surrounding vehicles), the gender of the driver of the vehicle 1 is male (driver attribute information Dc), the age of the driver of the vehicle 1 is in his / her teens, twenties, or thirties (driver attribute information Dc), and when any of the following applies: the driver of the vehicle 1 is angry, the driver of the vehicle 1 is disgusted, the driver of the vehicle 1 has feelings of contempt or insult, the driver of the vehicle 1 is surprised, the driver of the vehicle 1 has a panicked expression, and the driver of the vehicle 1 has an irritated expression (driver's expression and emotions De).

[0052] That is, in the host vehicle 1 to which the tailgating sign estimation device 20 of the third embodiment is applied, the tailgating sign determination unit 236 has the function of switching between determining whether or not the host vehicle 1 has signs of tailgating based on the tailgating sign level Id of the host vehicle 1 calculated by the tailgating sign level calculation unit 235 (first determination), and determining that the host vehicle 1 has signs of tailgating when multiple factors of the tailgating signs of the host vehicle 1 (driver attribute information Dc, driver's disease information Dd, driver's state Ds, driver's facial expression and emotions De, behavior Sb of surrounding vehicles, characteristics Sf of surrounding vehicles, characteristics Sd of the drivers of surrounding vehicles, characteristics Sp of passengers of surrounding vehicles, and state Eb of the host vehicle) correspond to any of the patterns of the first to fifth examples described above.

[0053] <Fourth embodiment> The vehicle 1 to which the tailgating sign estimation device 20 of the fourth embodiment is applied is configured in the same manner as the vehicle 1 to which the tailgating sign estimation device 20 of the first embodiment is applied, as described above, except for the points described below. In one example of the host vehicle 1 to which the road rage symptom estimation device 20 of the first embodiment is applied, the host vehicle 1 does not need to be equipped with the scent provision device 40. On the other hand, in an example of a host vehicle 1 to which the tailgating symptom estimation device 20 of the fourth embodiment is applied, as shown in FIG. 1, the host vehicle 1 is equipped with a scent providing device 40. The scent providing device 40 utilizes, for example, the principle of an air cannon. An air cannon is a device in which air inside a housing is discharged from an opening provided in the housing when an impact is applied to a part of the housing. The inside of the housing is connected to the outside of the housing through the opening. The air cannon is provided with a tank in which a scent component is sealed. The tank is fixed to the housing through a head equipped with an ejection mechanism. The head ejects the scent component in the tank. In addition, when the scent component is a gas, the head is an opening / closing valve, and the scent component can be ejected by opening the opening / closing valve. A plunger capable of reciprocating motion is connected to the housing. The plunger applies an impact (pressure) to the housing, which causes elastic deformation of the housing. When the head ejects the scent component into the housing and the plunger is driven, the air containing the scent component in the housing is ejected through the opening, and the scent is provided to the driver of the host vehicle 1. As described above, in one example of a vehicle 1 to which the first embodiment of the aggressive driving sign estimation device 20 is applied, when the aggressive driving sign determination unit 236 determines that there are signs of aggressive driving in the vehicle 1, the calming processing unit 332 executes a process to output the cause of the signs of aggressive driving in the vehicle 1 predicted by the cause prediction unit 331 to the HMI 5. On the other hand, in an example of a host vehicle 1 to which the fourth embodiment of the aggressive driving sign estimation device 20 is applied, when the aggressive driving sign determination unit 236 determines that there are signs of aggressive driving in the host vehicle 1, the calming processing unit 332 executes a process to output to the HMI 5 the cause of the signs of aggressive driving in the host vehicle 1 predicted by the cause prediction unit 331, and also executes a process to cause the fragrance providing device 40 to emit an aromatic scent. In another example, when the aggressive driving sign determination unit 236 determines that there are signs of aggressive driving of the vehicle 1, the calming processing unit 332 may execute a process of causing the fragrance providing device 40 to emit an aromatic scent, without executing a process of outputting the cause of the signs of aggressive driving of the vehicle 1 predicted by the cause prediction unit 331 to the HMI 5.

[0054] <Fifth embodiment> The vehicle 1 to which the tailgating sign estimation device 20 of the fifth embodiment is applied is configured in the same manner as the vehicle 1 to which the tailgating sign estimation device 20 of the first embodiment is applied, as described above, except for the points described below. In one example of the host vehicle 1 to which the aggressive driving sign estimation device 20 of the first embodiment is applied, the audio head unit 60 does not need to be used to calm the driver of the host vehicle 1. On the other hand, in an example of the host vehicle 1 to which the tailgating sign estimation device 20 of the fifth embodiment is applied, the host vehicle 1 is equipped with an audio head unit 60, which is used to calm the driver of the host vehicle 1. The audio head unit 60 has an audio ECU mainly composed of a computer, and a speaker connected to the audio ECU. The audio ECU receives on / off information (ACC information) of the accessory power supply, which is supplied or stopped by the accessory operation by the driver of the host vehicle 1. The audio ECU drives the speaker to output predetermined music or sound based on the input state of various switches, etc. The speaker is provided in the vehicle interior of the host vehicle 1, and outputs music and sound according to a command from the audio ECU. As described above, in an example of a host vehicle 1 to which the first embodiment of the aggressive driving sign estimation device 20 is applied, when the aggressive driving sign determination unit 236 determines that there are signs of aggressive driving in the host vehicle 1, the calming processing unit 332 executes a process of outputting the cause of the signs of aggressive driving in the host vehicle 1 predicted by the cause prediction unit 331 to the HMI 5. On the other hand, in an example of a vehicle 1 to which the fifth embodiment of the aggressive driving sign estimation device 20 is applied, when the aggressive driving sign determination unit 236 determines that there are signs of aggressive driving in the vehicle 1, the calming processing unit 332 executes a process to output the cause of the signs of aggressive driving in the vehicle 1 predicted by the cause prediction unit 331 to the HMI 5, and also executes a process to output, for example, music that can relax the driver of the vehicle 1 to the audio head unit 60. In another example, when the aggressive driving sign determination unit 236 determines that there are signs of aggressive driving of the vehicle 1, the calming processing unit 332 may execute a process of outputting to the audio head unit 60, for example, music that can relax the driver of the vehicle 1, without executing a process of outputting to the HMI 5 the cause of the signs of aggressive driving of the vehicle 1 predicted by the cause prediction unit 331.

[0055] Sixth embodiment The vehicle 1 to which the tailgating sign estimation device 20 of the sixth embodiment is applied is configured in the same manner as the vehicle 1 to which the tailgating sign estimation device 20 of the first embodiment is applied, as described above, except for the points described below. In one example of the host vehicle 1 to which the aggressive driving symptom estimation device 20 of the first embodiment is applied, the air conditioner 50 is not used to calm the driver of the host vehicle 1. On the other hand, in an example of the host vehicle 1 to which the tailgating symptom estimation device 20 of the sixth embodiment is applied, the host vehicle 1 is equipped with an air conditioner 50, and the air conditioner 50 is used to calm the driver of the host vehicle 1. The air conditioner 50 includes a heat pump, an interior air conditioning unit for blowing out temperature-controlled air adjusted by the heat pump into the vehicle interior of the host vehicle 1, and an air conditioning control device for controlling the operation of various electrically-driven components of the air conditioner 50. The heat pump is configured to be switchable between a refrigerant circuit in a cooling mode that cools the air sent into the vehicle interior of the host vehicle 1 to cool the vehicle interior of the host vehicle 1, a refrigerant circuit in a heating mode that heats the blown air to heat the vehicle interior, and a refrigerant circuit in a dehumidification heating mode that reheats the blown air dehumidified by cooling to dehumidify and heat the vehicle interior. In addition, the heat pump is equipped with a compressor that compresses and discharges the refrigerant, an indoor condenser for heating the blown air, an indoor evaporator for cooling the blown air, a fixed throttle for heating and a fixed throttle for cooling that reduce the pressure and expand the refrigerant, an outdoor heat exchanger, an on-off valve for switching the refrigerant circuit, a three-way valve, etc. As described above, in an example of a host vehicle 1 to which the first embodiment of the aggressive driving sign estimation device 20 is applied, when the aggressive driving sign determination unit 236 determines that there are signs of aggressive driving in the host vehicle 1, the calming processing unit 332 executes a process of outputting the cause of the signs of aggressive driving in the host vehicle 1 predicted by the cause prediction unit 331 to the HMI 5. On the other hand, in an example of a host vehicle 1 to which the sixth embodiment of the aggressive driving sign estimation device 20 is applied, when the aggressive driving sign determination unit 236 determines that there are signs of aggressive driving in the host vehicle 1, the calming processing unit 332 executes a process to output to the HMI 5 the cause of the signs of aggressive driving in the host vehicle 1 predicted by the cause prediction unit 331, and also executes a process to cause the air conditioning unit 50 to perform cooling operation. In another example, when the aggressive driving sign determination unit 236 determines that there are signs of aggressive driving of the vehicle 1, the calming processing unit 332 may execute a process to cause the air conditioner 50 to perform cooling operation without executing a process to output to the HMI 5 the cause of the signs of aggressive driving of the vehicle 1 predicted by the cause prediction unit 331.

[0056] Seventh embodiment The vehicle 1 to which the tailgating sign estimation device 20 of the seventh embodiment is applied is configured in the same manner as the vehicle 1 to which the tailgating sign estimation device 20 of the first embodiment is applied, as described above, except for the points described below. In one example of the host vehicle 1 to which the aggressive driving symptom estimation device 20 of the first embodiment is applied, the driver's seat 70 in which the driver of the host vehicle 1 sits is not used to calm the driver of the host vehicle 1. On the other hand, in an example of the host vehicle 1 to which the tailgating symptom estimation device 20 of the seventh embodiment is applied, the host vehicle 1 includes a driver's seat 70 on which the driver of the host vehicle 1 sits, and the driver's seat 70 is used to calm the driver of the host vehicle 1. The driver's seat 70 includes a seat cushion that supports the buttocks and thighs of the seated driver, a seat back that is provided to extend upward from the rear end of the seat cushion and supports the driver's waist and back, an air bag drive device, and the like. A plurality of air bags for massage are provided inside the seat cushion and the seat back of the driver's seat 70. Each air bag is configured to massage the driver's waist, back, thighs, etc. by inflating and pressing against a part of the driver's body, thereby relaxing the driver. The air bag drive device includes an air supply and exhaust device that supplies and exhausts air to and individually inflates and contracts each air bag, and a control device that controls the air supply and exhaust device. The air intake and exhaust device includes an intake port, an exhaust port, a pump as an air pressure delivery means, and a plurality of solenoid valves as control valves. The control device is electrically connected to the air intake and exhaust device, and controls the operation of the pump and the opening and closing of each port of the solenoid valve (common port, intake port, exhaust port). When the massage operation of the driver's seat 70 is started, the pump is operated under the control of the control device, and the air bag repeatedly expands and contracts at a constant cycle, providing vibrations to the waist, back, thighs, etc. of the driver seated in the driver's seat 70. As described above, in an example of a host vehicle 1 to which the first embodiment of the aggressive driving sign estimation device 20 is applied, when the aggressive driving sign determination unit 236 determines that there are signs of aggressive driving in the host vehicle 1, the calming processing unit 332 executes a process of outputting the cause of the signs of aggressive driving in the host vehicle 1 predicted by the cause prediction unit 331 to the HMI 5. On the other hand, in an example of a vehicle 1 to which the seventh embodiment of the aggressive driving sign estimation device 20 is applied, when the aggressive driving sign determination unit 236 determines that there are signs of aggressive driving in the vehicle 1, the calming processing unit 332 executes a process to output to the HMI 5 the cause of the signs of aggressive driving in the vehicle 1 predicted by the cause prediction unit 331, and also executes a process to cause the driver's seat 70 to vibrate the driver of the vehicle 1. In another example, when the aggressive driving sign determination unit 236 determines that there are signs of aggressive driving of the vehicle 1, the calming processing unit 332 may execute a process of causing the driver's seat 70 to vibrate the driver of the vehicle 1, without executing a process of outputting the cause of the aggressive driving signs of the vehicle 1 predicted by the cause prediction unit 331 to the HMI 5.

[0057] Eighth embodiment The vehicle 1 to which the tailgating sign estimation device 20 of the eighth embodiment is applied is configured in the same manner as the vehicle 1 to which the tailgating sign estimation device 20 of the first embodiment is applied, as described above, except for the points described below. In one example of the host vehicle 1 to which the aggressive driving sign estimation device 20 of the first embodiment is applied, the lighting device 80 in the cabin of the host vehicle 1 is not used to calm the driver of the host vehicle 1. On the other hand, in an example of the host vehicle 1 to which the tailgating symptom estimation device 20 of the seventh embodiment is applied, the host vehicle 1 is provided with a lighting device 80 arranged in the vehicle cabin of the host vehicle 1, and the lighting device 80 is used to calm the driver of the host vehicle 1. The lighting device 80 includes, for example, a meter panel provided on an instrument panel, various switches, foot lamps that illuminate the feet of the driver of the host vehicle 1, room lamps provided on the roof of the host vehicle 1, overhead lamps, door lamps provided on the doors of the host vehicle 1, and the like. The lighting device 80 includes, for example, a red light source, a green light source, a blue light source, a control circuit that controls the drive current of these light sources, and the like, and is configured to be able to change the color of the light emitted. As described above, in an example of a host vehicle 1 to which the first embodiment of the aggressive driving sign estimation device 20 is applied, when the aggressive driving sign determination unit 236 determines that there are signs of aggressive driving in the host vehicle 1, the calming processing unit 332 executes a process of outputting the cause of the signs of aggressive driving in the host vehicle 1 predicted by the cause prediction unit 331 to the HMI 5. On the other hand, in an example of a host vehicle 1 to which the seventh embodiment of the aggressive driving sign estimation device 20 is applied, when the aggressive driving sign determination unit 236 determines that there are signs of aggressive driving in the host vehicle 1, the calming processing unit 332 executes a process to output to the HMI 5 the cause of the signs of aggressive driving in the host vehicle 1 predicted by the cause prediction unit 331, and also executes a process to display illumination (e.g., blue illumination) on the lighting device 80 to calm the driver of the host vehicle 1. In another example, when the tailgating sign determination unit 236 determines that there are signs of aggressive driving of the vehicle 1, the calming processing unit 332 may execute a process of causing the lighting device 80 to display illumination (e.g., blue illumination) to calm the driver of the vehicle 1, without executing a process of outputting the cause of the signs of aggressive driving of the vehicle 1 predicted by the cause prediction unit 331 to the HMI 5.

[0058] As described above, the embodiments of the road rage sign estimating device, road rage sign estimating method, and program of the present disclosure have been described with reference to the drawings, but the road rage sign estimating device, vehicle, road rage sign estimating method, and program of the present disclosure are not limited to the above-mentioned embodiments, and can be appropriately modified without departing from the spirit of the present disclosure. The configurations of the examples of the above-mentioned embodiments can be appropriately combined. [Explanation of symbols]

[0059] 1... hostile driving sign estimation device, 23... processor, 231... acquisition unit, 232... detection unit, 233... recognition unit, 234... estimation unit, 235... tailgating sign degree calculation unit, 236... tailgating sign determination unit, 30... calming down processing device

Claims

1. A tailgating symptom determination unit that determines whether or not the host vehicle has tailgating symptoms based on a plurality of factors of the tailgating symptoms of the host vehicle, which are at least one of information about the driver of the host vehicle and information about the host vehicle; a tailgating symptom degree calculation unit that calculates a tailgating symptom degree of the vehicle by adding up the products of the scores of each of a plurality of factors of the tailgating symptom of the vehicle, The tailgating sign determination unit determines that the tailgating sign of the vehicle is present when the tailgating sign degree calculated by the tailgating sign degree calculation unit is greater than a threshold value, thereby estimating the tailgating sign.

2. The tailgating symptom determination unit is Determining whether or not there is a sign of tailgating of the vehicle based on the sign level of tailgating of the vehicle calculated by the tailgating sign level calculation unit; The tailgating signs estimation device of claim 1 has the function of switching between determining whether or not the vehicle has signs of tailgating based on multiple factors of the signs of tailgating by using a model obtained by learning from teacher data, which is a data set of factors of signs of tailgating and labels indicating whether or not there are signs of tailgating.

3. The tailgating symptom determination unit is Determining whether or not there is a sign of tailgating of the vehicle based on the sign level of tailgating of the vehicle calculated by the tailgating sign level calculation unit; The tailgating sign estimation device of claim 1 has the function of switching between determining that the vehicle has signs of tailgating when multiple factors of the signs of tailgating of the vehicle correspond to a predetermined combination pattern and determining that the vehicle has signs of tailgating.

4. A tailgating sign determination step in which the tailgating sign estimation device determines whether or not the tailgating sign of the host vehicle is present based on a plurality of factors of the tailgating sign of the host vehicle, which are at least one of information about the driver of the host vehicle and information about the host vehicle; The road rage sign estimation device includes a road rage sign degree calculation step for calculating a road rage sign degree of the vehicle by adding up the products of the scores of each of a plurality of factors of road rage signs of the vehicle, In the tailgating sign determination step, it is determined that the vehicle has signs of tailgating if the degree of tailgating signs of the vehicle calculated in the tailgating sign calculation step is greater than a threshold value.

5. To the processor a tailgating symptom determination step of determining whether or not the host vehicle has a tailgating symptom based on a plurality of factors of the tailgating symptom of the host vehicle, which are at least one of information about the driver of the host vehicle and information about the host vehicle; A program for executing a step of calculating a symptom degree of tailgating of the vehicle by adding up the products of the scores of each of a plurality of factors of the symptoms of tailgating of the vehicle, A program in which, in the aggressive driving sign determination step, it is determined that the vehicle has signs of aggressive driving if the sign level of aggressive driving of the vehicle calculated in the aggressive driving sign level calculation step is greater than a threshold value.

Images