Vehicle integrated control device

Abstract

In the present invention, when driving control is performed by controlling the attitude of a vehicle so that motion sickness is not susceptible to onset, the driving control is prevented from becoming insufficient for an occupant who is highly sensitive to motion sickness even if the acquisition state of the head attitude for predicting the head swing of the occupant is poor. This integrated vehicle control device (2) is provided with: a target value generation unit (21) that generates a target value (22) for motion control of a vehicle; a head posture acquisition unit (23) that acquires the head posture of the occupant; an acquisition state determination unit (25) that determines the acquisition state of the head posture; a vehicle movement acquisition unit (26) that acquires vehicle movement information indicating the state of movement of the vehicle; a head swing prediction unit (27) that predicts and outputs the head swing of the occupant on the basis of the head posture and at least one of the target value (22) and the vehicle motion information; and a target value correction unit (28) that corrects and outputs a target value (22) so as to reduce the head oscillation on the basis of the prediction result of the predicted head oscillation. The head swing prediction unit (27) or the target value correction unit (28) corrects the output on the basis of the acquired state.

Description

Technical Field

[0001] The present invention relates to a vehicle integrated control device. Background Art

[0002] As a motion sickness inference system and vehicle that can reduce the possibility of motion sickness (motion sickness), the technology described in Patent Document 1 is known. For example, the abstract of Patent Document 1 states the following: "A motion sickness inference system (10) includes an inference unit (312) and an output unit (313). The inference unit (312) is configured to perform an inference process, wherein the inference process is to infer whether the person is in a condition that causes motion sickness based on person information indicating the condition of the person riding in the vehicle. The output unit (313) is configured to output the result of the inference process in the inference unit (312)." In addition, Claim 14 of Patent Document 1 states the following: "The travel control device is configured to control the travel of the vehicle in a manner that reduces the burden on the person during the preventive process." Prior art literature Patent Literature

[0003] Patent Document 1: International Publication No. 2019 / 177002 Summary of the Invention Problems to be solved by the invention

[0004] However, the motion sickness inference method of Patent Document 1 is based on the premise that "personal information indicating the state of a person riding in a vehicle" can be properly obtained (for example, obtaining a good state, such as obtaining with high accuracy). Therefore, there is a problem that, when the person information cannot be properly obtained (for example, obtaining a poor state, such as obtaining with low accuracy), the inference of "whether the person is in a condition prone to motion sickness" may be erroneous.

[0005] For example, when acquiring and utilizing head posture to predict a passenger's head sway, if head posture acquisition is poor, the magnitude of the passenger's head sway will be underestimated compared to when acquisition is good. This can lead to passengers who are highly susceptible to motion sickness being misidentified as being less susceptible. Consequently, when driving control is performed by controlling the vehicle's posture to minimize the onset of motion sickness, driving control becomes insufficient, leading to the onset of motion sickness.

[0006] The problem to be solved by the present invention is to provide a vehicle integrated control device that, when controlling the vehicle's posture to perform driving control in a manner that makes motion sickness less likely to occur, can prevent driving control from becoming insufficient for an occupant who is highly sensitive to motion sickness even if the acquisition state of the head posture used to predict the occupant's head swing is poor. Technical means to solve the problem

[0007] In order to solve the above-mentioned problems, the vehicle integrated control device of the present invention is characterized in that it comprises: a target value generating unit, which generates a target value for the motion control of the vehicle; a head posture acquiring unit, which acquires the head posture of the occupant; an acquisition state determining unit, which determines the acquisition state of the head posture; a vehicle motion acquiring unit, which acquires vehicle motion information representing the state of the motion of the vehicle; a head swing prediction unit, which predicts and outputs the head swing of the occupant based on the target value and at least one of the vehicle motion information and the head posture; and a target value correction unit, which corrects and outputs the target value in a manner to reduce the head swing based on the predicted result of the predicted head swing; the head swing prediction unit or the target value correction unit corrects the output according to the acquisition state. Effects of the Invention

[0008] According to the vehicle integrated control device of the present invention, when driving control is performed by controlling the vehicle's posture in a manner that makes it less likely for motion sickness to occur, even if the acquisition state of the head posture used to predict the occupant's head swing is poor, driving control can be prevented from becoming insufficient for occupants who are highly sensitive to motion sickness.

[0009] Furthermore, other problems, structures, and effects than those described above will become clearer through the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 It is a plan view showing the overall structure of the vehicle of Example 1. Figure 2 This is a schematic diagram illustrating input and output signals of the vehicle integrated control device according to the first embodiment. Figure 3 This is a functional block diagram of the vehicle integrated control device of Example 1. Figure 4A This is an explanatory diagram showing an example of a camera mounted on a vehicle according to the first embodiment. Figure 4B This is an explanatory diagram showing an example of a camera mounted on a vehicle according to the first embodiment. Figure 5A It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment. Figure 5B It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment. Figure 5C It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment. Figure 6A It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment. Figure 6B It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment. Figure 7A It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment. Figure 7B It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment. Figure 8A It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment. Figure 8B It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment. Figure 8C It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment. Figure 9 It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment. Figure 10A This is an explanatory diagram showing an example of a head swing prediction method in the head swing prediction unit of the vehicle integrated control device according to the first embodiment. Figure 10B This is an explanatory diagram showing an example of a head swing prediction method in the head swing prediction unit of the vehicle integrated control device according to the first embodiment. Figure 11 This is an explanatory diagram showing an example of driving data for learning a coefficient for head swing prediction in the head swing prediction unit of the vehicle integrated control device according to the first embodiment. Figure 12 This is an explanatory diagram showing an example of a coefficient learning result for head swing prediction in the head swing prediction unit of the vehicle integrated control device according to the first embodiment. Figure 13 This is an explanatory diagram showing an example of a driving scenario for explaining the behavior of the target value correction unit of the vehicle integrated control device according to the first embodiment. Figure 14 This is an explanatory diagram showing an example of the behavior of the target value correction unit of the vehicle integrated control device according to the first embodiment. Figure 15 This is an explanatory diagram showing an example of the behavior of the target value correction unit of the vehicle integrated control device according to the first embodiment. Figure 16 This is an explanatory diagram showing an example of the behavior of the target value correction unit of the vehicle integrated control device according to the first embodiment. Figure 17This is an explanatory diagram showing an example of the behavior of an actuator for realizing the behavior of the target value correction unit of the vehicle integrated control device of the first embodiment. Figure 18 It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the second embodiment. Figure 19 This is a functional block diagram of the vehicle integrated control device of Example 3. DETAILED DESCRIPTION

[0011] In the following, embodiments of the present invention will be described using the accompanying drawings. In each figure and each embodiment, the same or similar components are denoted by the same reference numerals and repeated descriptions are omitted. Example 1

[0012] Figure 1 It is a plan view showing the overall structure of the vehicle of Example 1.

[0013] The vehicle 1 includes a vehicle integrated control device 2, an external control device 3, a combination sensor 4, a wheel 11, a motor 12, a braking mechanism 13 (a wheel cylinder 13a and a braking control device 13b), a steering mechanism 14 (a steering control device 14a and a steering motor 14b), a suspension 15, an accelerator pedal 16, a stroke sensor 16a, an acceleration control device 16b, a brake pedal 17, a steering wheel 18, a steering torque detection device 18a, a steering angle detection device 18b, and an external sensor 19. Regarding the subscripts in the figure, FL is a symbol indicating that the structure corresponds to the left front, FR is a symbol indicating that the structure corresponds to the right front, RL is a symbol indicating that the structure corresponds to the left rear, RR is a symbol indicating that the structure corresponds to the right rear, SL is a symbol indicating that the structure corresponds to the left side, and SR is a symbol indicating that the structure corresponds to the right side. Taking the wheel 11 as an example, 11 FL 、11 FR 、11 RL 、11 RR In addition, F is a symbol indicating that the structure corresponds to the front side, R is a symbol indicating that the structure corresponds to the rear side, and S is a symbol indicating that the structure corresponds to the side.

[0014] In the following, the front-to-back direction of the vehicle 1 is defined as the x-axis (the front direction is defined as positive), the left-to-right direction is defined as the y-axis (the left direction is defined as positive), and the up-down direction is defined as the z-axis (the up direction is defined as positive), and the details of each structure are explained in sequence on this basis.

[0015] The vehicle integrated control device 2 is a control device that performs integrated control of actuators such as the motor 12, the braking mechanism 13, the steering mechanism 14, and the suspension 15 based on the driver's operation, external instructions from the external control device 3, and detection signals from the combination sensor 4 (detection signals related to the control axes of the six degrees of freedom, including the accelerations in the three translation directions (front and back, left and right, and up and down) and the angular velocities in the three rotation directions (roll, pitch, and yaw)). The vehicle integrated control device 2 is specifically an ECU (Electronic Control Unit) equipped with hardware such as a CPU, a main storage device such as a semiconductor memory, an auxiliary storage device, and a communication device. Therefore, the various functions described below are realized by the execution of a program loaded from the auxiliary storage device to the main storage device by the computing device, but such known technologies will be omitted as appropriate for description below.

[0016] The external control device 3 is a high-level controller that performs driving assistance control or automatic driving control via the vehicle integrated control device 2. It calculates the speed command value or acceleration command value for adaptive cruise control (ACC) to follow the preceding vehicle, or the yaw command value for lane keeping control (LKC) to maintain lane keeping, based on the external information acquired by the external sensor 19 (camera, radar, LiDAR, etc.), and outputs these command values ​​as external commands to the vehicle integrated control device 2. Figure 1 In the example, the vehicle integrated control device 2 and the external control device 3 are provided as separate entities, but both can be realized by a single ECU.

[0017] As the external sensor 19, for example, a fisheye camera with a 180° viewing angle can be installed on the front surface, left and right sides, and rear surface of the vehicle 1 (19 F 、19 SL 、19 SR 、19 R ), thereby detecting the relative distance and relative speed to objects such as vehicles, bicycles, pedestrians, and obstacles around vehicle 1. Furthermore, while this embodiment illustrates a combination of the aforementioned sensors as an example of a sensor configuration, this is not limiting. Combinations with ultrasonic sensors, stereo cameras, infrared cameras, laser radars, and the like are also possible. Alternatively, a laser radar capable of sensing a 360-degree angle of the surrounding area may be installed on the roof of vehicle 1. The aforementioned sensor signals (output signals from the aforementioned sensors) are input to the vehicle integrated control unit 2 or the external control unit 3.

[0018] Here, the drive system of the vehicle 1 is described. The vehicle 1 is equipped with a torque generating device that applies driving force to each wheel 11 as a key part of the drive system. An example of the torque generating device is an engine or a motor that transmits driving force to a pair of left and right wheels 11 via a differential and a drive shaft. Another example of the torque generating device is a hub motor 12 that drives each wheel 11 independently. Figure 1 The details of this embodiment are described based on the vehicle structure of FIG.

[0019] To move vehicle 1 forward (or reverse), the driver sets the shift lever to the desired gear position and then operates accelerator pedal 16. A stroke sensor 16a detects the amount of accelerator pedal 16 depressed, and acceleration control device 16b converts the depression amount into an acceleration command, which is output to vehicle integrated control device 2. Vehicle integrated control device 2 supplies power corresponding to the input acceleration command from a battery (not shown) to each wheel motor 12, controlling the torque of each motor. As a result, vehicle 1 can be accelerated or decelerated based on the operation of accelerator pedal 16.

[0020] Furthermore, when driving assistance or automated driving is implemented based on external commands from external control device 3, vehicle integrated control device 2 supplies the desired power to each wheel motor 12 in accordance with the input external commands, thereby controlling the torque of each motor. As a result, vehicle 1 is accelerated or decelerated, enabling the desired driving assistance or automated driving to be performed.

[0021] Next, the braking system of vehicle 1 will be described. Vehicle 1 is equipped with a wheel cylinder 13a, a key component of the braking system, that applies braking force to each wheel 11. The wheel cylinder 13a is composed of, for example, a cylinder body, a piston, brake pads, and a brake disc. In the wheel cylinder 13a, hydraulic fluid supplied from a master cylinder pushes the piston forward, pressing the brake pad connected to the piston against the brake disc, which rotates with the wheel 11. The braking torque acting on the brake disc then becomes the braking force acting between the wheel 11 and the road surface.

[0022] When the driver wishes to brake the vehicle 1, he or she operates the brake pedal 17. At this time, the driver's pedal force on the brake pedal 17 is increased by a brake booster (not shown), and a hydraulic pressure approximately proportional to the pedal force is generated by the master cylinder. The generated hydraulic pressure is supplied to the wheel cylinder 13a of each wheel via the brake mechanism 13. FL 、13a FR 、13a RL 、13a RRTherefore, in response to the driver's brake pedal operation, the piston of each wheel cylinder 13a is pressed against the brake disc, generating braking force at each wheel. Furthermore, in a vehicle 1 equipped with the vehicle integrated control device 2, the brake booster and master cylinder may be omitted. In this case, the following mechanism may be used: the brake pedal 17 is directly connected to the brake mechanism 13, and the brake mechanism 13 is directly activated when the driver steps on the brake pedal 17.

[0023] Furthermore, when driving assistance or automated driving is implemented based on external commands from the external control device 3, the vehicle integrated control device 2 controls the brake mechanism 13 and the wheel cylinders 13a of each wheel via the brake control device 13b in accordance with the input external commands. As a result, the vehicle 1 is braked, and the desired driving assistance or automated driving is performed. Furthermore, the brake control device 13b also has the function of converting the driver's operation on the brake pedal 17 into a braking command and outputting it as an external command to the vehicle integrated control device 2.

[0024] Next, a description will be given of a steering system of the vehicle 1. The vehicle 1 is equipped with a steering mechanism 14 for applying a steering force to each wheel 11 as a main part of the steering system. Figure 1 The front wheel 11 is shown in F (Left front wheel 11 FL 、Right front wheel 11 FR ) Steering mechanism 14 on the front side F and for the rear wheel 11 R (Left rear wheel 11 RL 、Right rear wheel 11 RR ) The rear steering mechanism 14 for steering R , but it is not necessary to equip the steering mechanism 14 on both the front and rear sides. For example, the steering mechanism 14 on the rear side can be omitted. R .

[0025] When the driver wishes to steer the vehicle 1, the driver operates the steering wheel 18. At this time, the "steering torque" and "steering angle" input by the driver via the steering wheel 18 are detected by the steering torque detection device 18a and the steering angle detection device 18b. F The front steering motor 14b is controlled based on the detected steering torque and steering angle. F The front wheel 11 F Similarly, the rear steering control device 14a R The rear steering motor 14b is controlled based on the detected steering torque and steering angle. R The rear wheel 11 R Auxiliary torque for steering.

[0026] Furthermore, when driving assistance or automated driving is implemented based on external commands from the external control device 3, the vehicle integrated control device 2 controls the steering torque of the steering motor 14b via the steering control device 14a. As a result, the vehicle 1 is steered, and the desired driving assistance or automated driving is performed. In this case, the steering wheel 18 can be omitted.

[0027] Next, the suspension system of vehicle 1 will be described. Vehicle 1 is equipped with a suspension 15, a key component of the suspension system, that absorbs vibrations and shocks generated by each wheel 11, thereby improving vehicle stability and ride comfort. This suspension 15 can be, for example, a semi-active suspension combining a variable viscosity damper and coil springs, or a fully active suspension combining an adjustable-length actuator, a damper, and coil springs to arbitrarily adjust the relative distance between the vehicle body and wheels 11.

[0028] The vehicle integrated control device 2 controls the viscosity of the semi-active suspension and the length of the fully active suspension, thereby improving ride comfort and the like and appropriately controlling the posture of the vehicle 1 according to the environment.

[0029] Figure 2 This is a schematic diagram illustrating input and output signals of the vehicle integrated control device of Example 1. Figure 2 The input and output of the vehicle integrated control device 2 will be described.

[0030] Acceleration commands, braking commands, steering torque, steering angle, and other external commands generated by the driver's operation of the accelerator pedal 16, brake pedal 17, and steering wheel 18 are input to the vehicle integrated control device 2 as external commands. External information acquired by external sensors 19 may also be input to the vehicle integrated control device 2. Furthermore, the vehicle integrated control device 2 generates vehicle motion command values ​​generated by the external control device 3 during driving assistance control or automatic driving control. This illustration uses the example of inputting external commands for up to six degrees of freedom, including forward and backward command values, left and right command values, up and down command values, roll command values, pitch command values, and yaw command values. These command values ​​include, for example, velocity, acceleration, jerk, angle, angular velocity, angular acceleration, and steering angle. Furthermore, the vehicle integrated control device 2 receives detection values ​​of forward and backward, left and right, and up and down accelerations, as well as angular velocities of roll, pitch, and yaw, from the combined sensor 4.

[0031] Then, the vehicle integrated control device 2 appropriately allocates the motor 12 (12 FL ~12 RR )、Brake mechanism 13 (wheel cylinder 13a FL ~13a RR )、Steering mechanism 14 (steering motor 14b F 、14b R)、Suspension 15(15 FL ~15 RR ) (hereinafter sometimes referred to as actuators) respectively and perform driving, braking, steering, and suspension control, thereby achieving the desired vehicle control including posture control. Figure 1 Vehicle 1 supports manual driving, so Figure 2 While external commands from the driver are also illustrated, the present invention can also be applied to vehicles 1 that only support fully automated driving or remote operation. In this case, the configuration can be such that external commands from the driver are omitted. During automated driving, external commands with up to six degrees of freedom can be input from the external control device 3, or target values ​​for automated driving can be generated using external sensors 19 and map information stored within the vehicle integrated control device 2. This embodiment is described assuming that target values ​​for automated driving are generated within the vehicle integrated control device 2.

[0032] Furthermore, information from the actuators (12-15) is also input into the vehicle integrated control device 2. For example, this information includes the position of the motor 12. Furthermore, if jerk information is required, the acceleration is differentiated with respect to time to obtain the jerk. Similarly, if angular acceleration information is required, the angular velocity is differentiated with respect to time to obtain the angular acceleration. Furthermore, if velocity and angle information is required, these are estimated through calculations based on various information.

[0033] The occupant's head posture 24 is also input to the vehicle integrated control unit 2. A camera is installed in the interior of the vehicle 1 to measure the occupant's head movement and estimate their susceptibility to motion sickness, as described in detail later. Alternatively, a mechanism may be provided to obtain information related to the occupant's susceptibility to motion sickness from a mobile device held by the occupant while in the vehicle.

[0034] Figure 3 This is a functional block diagram of the vehicle integrated control device of Example 1.

[0035] Figure 2 The example shown is a vehicle integrated control device 2 that inputs three external instructions from the driver and also inputs a maximum of six external instructions from the external control device 3. As mentioned above, in this embodiment, the details of the vehicle integrated control device 2 of this embodiment are explained by taking the structure of generating target values ​​for automatic driving within the vehicle integrated control device 2 as an example.

[0036] The vehicle integrated control device 2 of this embodiment includes a target value generating unit 21 , a head posture acquiring unit 23 , an acquired state determining unit 25 , a vehicle motion acquiring unit 26 , a head swing predicting unit 27 , and a target value correcting unit 28 .

[0037] The target value generation unit 21 generates a target value 22 for motion control of the vehicle 1. Specifically, the target value generation unit 21 outputs the target value 22, which is a vehicle motion goal for achieving a specific driving task in autonomous driving (such as following a route or driving at the same speed as a preceding vehicle), to the target value correction unit 28. The target value 22 typically consists of three types: a longitudinal acceleration command value, a lateral acceleration command value, and a yaw command value. Additionally, a roll angle command value, a pitch angle command value, and a vertical direction command value are added, resulting in a maximum of six command values. Furthermore, when the three external commands from the driver (acceleration command, braking command, steering torque, and steering angle) are input, the target value generation unit 21 converts these external commands into longitudinal acceleration command values, lateral acceleration command values, and yaw command values, and outputs them as the target value 22.

[0038] The head posture acquisition unit 23 acquires the occupant's head posture. Specifically, the head posture acquisition unit 23 acquires the occupant's head posture 24 using sensors such as cameras or input devices installed in the vehicle 1. The occupant's head posture 24 is sensing information obtained by sensing the occupant's head, such as a camera image.

[0039] Figure 4A and Figure 4B This is an explanatory diagram showing an example of a camera mounted on a vehicle according to the first embodiment.

[0040] Figure 4A The illustration is based on the example of a vehicle 1 in the shape of a bus capable of self-driving. In this example, a camera 100 with a 360° field of view mounted on the ceiling detects the postures and head movements of two passengers 51 as the passenger head posture 24. The detected passenger head posture 24 is output to the head posture acquisition unit 23. The camera 100 is not limited to such a shape and position. Multiple cameras 100 may be installed in the vehicle interior, and the field of view may not be 360°. In addition, the camera 100 may be used to acquire the passenger 51's line of sight and the task they perform while riding (reading, sleeping, etc.) and include this information in the passenger head posture 24. Alternatively, the head posture acquisition unit 23 may analyze and acquire the passenger 51's line of sight and the task they perform while riding (reading, sleeping, etc.) based on the passenger head posture 24 from the camera 100, for example, by performing image analysis.

[0041] In the image Figure 4B When vehicle 1 is in the shape of a car like that, camera 100 also can be arranged near the connecting point of windshield and roof (usually the part where interior mirror is housed). In this case, the passenger head posture 24 of the passenger 51 sitting on the rear seat is detected.

[0042] This description assumes that the sensor is a visible light camera, but any sensor capable of capturing an image of the occupant's head will suffice. For example, a camera utilizing not only visible light but also infrared light can be used, and 3D shape information measured by LiDAR or other sensors can also be used.

[0043] The acquisition status determination unit 25 determines the acquisition status of the head posture. Specifically, the acquisition status determination unit 25 determines the acquisition status, specifically the sensing status and acquisition accuracy, based on the sensing information acquired by the head posture acquisition unit 23, namely, the occupant's head posture 24. Furthermore, the acquisition status determination unit 25 can also determine the occupant's 51 status, specifically, their riding position, posture, orientation, and attributes (gender, age group, etc.).

[0044] Next, use Figures 5A to 9 An example of the behavior of the acquisition state determination unit 25 is described. Furthermore, it is assumed here that the head posture acquisition unit 23 acquires the image of the camera as information on the occupant's head posture 24. As a method of inferring the head posture using the image of the camera, the following well-known technology (for example, open source software such as OpenPose) is used as an example to illustrate: the facial parts (eyes, nose, ears, etc.) of the occupant are identified to detect the 2D posture of the head. In addition, among the acquisition states, the description is made with particular attention to the point of view of the acquisition accuracy (sensing accuracy) of the head posture. In addition, in all examples, occupant 51g represents an example of a case where the acquisition accuracy is high (acquisition state is good), and occupant 51n represents an example of a case where the acquisition accuracy is low (acquisition state is poor). As described later, the lower the acquisition accuracy, the more likely the occupant is to suffer from motion sickness, and the vehicle control is changed.

[0045] Figures 5A to 5C It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment.

[0046] Here is an example of a decrease in the accuracy of acquiring the head posture due to the passenger 51n wearing a mask 110. Figure 5A The facial part of the passenger 51g is recognized as shown in the recognition image diagram to estimate the head posture. Figure 5B As shown, the mask 110 covers part of the face of the occupant 51n, so the head posture inferred from it is likely to become inaccurate. Figure 5B In FIG, since the occupant 51n wears the mask 110, although the head is approximately horizontal, the eye recognition is tilted. Figure 5C An example of a time waveform of the head roll angle 74g of the occupant 51g and the head roll angle 74n of the occupant 51n when the occupant travels on the same travel path is shown. Figure 5CIn the figure, the horizontal axis represents time [s], and the vertical axis represents the head roll angle [deg]. The head roll angle 74g, shown by the solid line, is approximately 0° between 10 and 20 seconds. Meanwhile, the head roll angle 74n, shown by the dashed line, deviates in the negative direction at the same time. Furthermore, after 30 seconds, the head roll angle 74n undergoes a "jump" in value.

[0047] In this way, the fewer the number of facial parts (feature points) recognized, the lower the accuracy of head posture acquisition. In addition, for images, it is significantly more difficult to recognize changes in head posture in the vertical direction (in this case, the longitudinal tilt direction), so the occupant's sensitivity to motion sickness due to the front-to-back movement of the vehicle 1 is likely to be evaluated as low. In this case, it is not clear whether the occupant 51n wearing the mask 110 is actually prone to motion sickness, and for occupants who are prone to motion sickness, the change in vehicle control may become insufficient. Therefore, in this case, the acquisition accuracy is judged to be low, and the vehicle control is changed based on the occupant being deemed to be prone to motion sickness. Furthermore, this phenomenon occurs not only when a mask is worn or not, but also when it is difficult to see the face, for example, due to the setting sun.

[0048] Figure 6A and Figure 6B It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment.

[0049] Here, an example is shown in which the passenger 51n hides his face by reading a book 111, etc., thereby partially hiding his face and causing the head posture to be acquired with reduced accuracy. Figure 4B In that case, in a car-shaped vehicle 1, when the camera 100 is shooting the rear from the front, the passenger 51 sitting in the rear seat will experience motion sickness. Generally speaking, a passenger 51n who is reading a book 111 or the like is more likely to experience motion sickness than a passenger 51g who is not reading a book. If the accuracy of acquiring the head posture is reduced, resulting in a low evaluation of the sensitivity of the passenger 51n to motion sickness, the vehicle control change may become insufficient for the passenger 51n who is prone to motion sickness. Therefore, in this case, the acquisition accuracy is determined to be low, and the vehicle control is changed as if the passenger is prone to motion sickness. Furthermore, Figure 6B The illustration is based on the book 111 as an example, and the situation where the face cannot be seen because it is blocked by the front seat passenger 51 or the headrest rather than the book 111 also meets this example.

[0050] Figure 7A and Figure 7B It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment.

[0051] Shown here because like Figure 7AThe passenger 51g shown does not look forward, Figure 7B In this example, occupant 51n is looking to the side (e.g., the outside scenery visible through window 112), which results in partial loss of facial recognition and reduced head posture acquisition accuracy. In this example, it's unclear whether occupant 51n is actually susceptible to motion sickness. However, a person who is susceptible to motion sickness or already experiencing motion sickness may be observing the outside scenery to prevent or reduce motion sickness. If the reduced head posture acquisition accuracy results in a low assessment of occupant 51n's motion sickness susceptibility, vehicle control changes may be insufficient for occupant 51n, who is likely susceptible to motion sickness. Therefore, in this case, the acquisition accuracy is determined to be low, and vehicle control changes are made based on the assumption that the occupant is susceptible to motion sickness.

[0052] picture Figures 5A to 7B As described above, the fewer facial parts (feature points) that are recognized, the lower the head posture acquisition accuracy. Therefore, the acquisition status determination unit 25 may determine the acquisition status based on the number of feature points on the head of the occupant 51 in the sensed information, i.e., the occupant head posture 24. For example, if the number of feature points on the head of the occupant 51 is less than a predetermined threshold, the acquisition status may be determined to be poor (low acquisition accuracy).

[0053] Next, use Figures 8A to 9 An example of determining acquisition accuracy based on the relative positional relationship between camera 100 and occupant 51 will be described. The acquisition status determination unit 25 determines the acquisition status based on the relative positional relationship between the camera 100 or other sensor that acquires the sensing information and the occupant's 51 head. More specifically, the lower the resolution or the more distorted the head image captured by camera 100, the lower the acquisition accuracy. Lower resolution means lower resolution of the head's posture, making it more difficult to detect subtle movements. This can result in a lower assessment of the occupant's motion sickness susceptibility. Therefore, for occupants determined to have low head posture acquisition accuracy, vehicle control is altered based on the occupant's perceived susceptibility to motion sickness.

[0054] Figures 8A to 8C It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment.

[0055] Figure 8A Here, the camera 100 is used as a so-called driver monitor camera and the camera 100 is directed toward the driver to mainly photograph the driver's head. Figure 8A The example shown in the figure is that the driver's seat is on the right side. Figure 8BAs shown in FIG. 1 , the passenger 51g in the rear right seat is close to the optical axis 101 of the camera 100 and is located in the center of the image, so the accuracy of obtaining the passenger 51g is high. Figure 8C As shown, the passenger 51n in the rear left seat is far from the optical axis 101 and is located in the corner of the image, so the acquisition accuracy is judged to be low. The basis for this judgment is that the image provided by the camera 100 usually has a higher resolution as it is closer to the center and a lower resolution as it is closer to the corner, so the head posture acquisition accuracy is more likely to decrease as it is closer to the corner. If it is a wide-angle lens, the image will be more distorted as it is closer to the corner, so the head posture acquisition accuracy is more likely to decrease as it is closer to the corner. According to this example, for example, the passenger seat is close to the camera 100 but is located in the corner in the image, so the acquisition accuracy is considered to be in the middle. Similarly, the rear right seat image is Figure 8B That is considered to be a high-precision acquisition, the left seat in the back row is like Figure 8C This is considered to be low acquisition accuracy, and the rear center seat has an intermediate acquisition accuracy.

[0056] Therefore, the acquisition status determination unit 25 can simply determine the acquisition status based on the position of the occupant 51's head within the image. For example, the further away from the center (the closer to the corner) the occupant 51's head is within the image, the worse the acquisition status is determined to be (lower acquisition accuracy). Alternatively, the acquisition status determination unit 25 can determine the acquisition status based on the vertical distance from the optical axis 101 of a sensor such as the camera 100. For example, the greater the vertical distance from the optical axis 101, the worse the acquisition status is determined to be (lower acquisition accuracy).

[0057] Figure 9 It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the first embodiment.

[0058] Figure 9 In this example, as in Figure 8 , camera 100 captures the image from the front and the rear. In this example, the farther away occupant 51 is from camera 100, the lower the acquisition accuracy. Specifically, the head pose of occupant 51g in the front row is captured with high accuracy, while the head pose of occupant 51n in the back row is captured with low accuracy. This takes into account the fact that for subjects of the same size, the closer they are, the higher the resolution.

[0059] Therefore, the acquisition status determination unit 25 can determine the acquisition status based on the proportion of the occupant 51's head within the image. For example, the smaller the proportion of the occupant 51's head within the image, the poorer the acquisition status (lower the acquisition accuracy). Alternatively, the acquisition status determination unit 25 can determine the acquisition status based on the distance from the sensor, such as the camera 100, to the occupant 51's head. For example, the greater the distance from the sensor, the poorer the acquisition status (lower the acquisition accuracy).

[0060] Furthermore, the previous description is based on the image Figure 4B The following description is given of the case where the camera 100 captures the rear view from the front in a car-shaped vehicle 1. Figure 4A When a 360° camera or multiple cameras mounted on the ceiling are filming in all directions inside the vehicle, the head posture can be acquired with high accuracy even for rear-facing passengers 51. Generally, rear-facing passengers 51 are more susceptible to motion sickness than non-rear-facing passengers 51. In such cases, information such as "rear-facing" can be separately acquired as "acquisition status" to determine that the passenger is susceptible to motion sickness and to modify vehicle control.

[0061] While several examples of the behavior of the acquisition status determination unit 25 have been shown above, the acquisition status is actually calculated by integrating this information. For example, the acquisition accuracy of the head posture is expressed as y, which is a quantitative value representing the acquisition accuracy.

[0062] [Formula 1]

[0063] Here, c is the maximum value of the acquired accuracy when there are no precision-reducing factors, Σ is the sum of the series, x is a numerical value indicating the presence or absence of each factor that causes precision reduction, A is a coefficient indicating the degree of contribution of each factor to precision reduction, and n indicates that the number of factors that cause precision reduction is n. For example, x1 is set to "the number of feature points reduced in head posture measurement", x2 is set to "the distance from the camera to the object (m)", and x3 is set to "the lateral distance from the optical axis of the camera to the object (m)", thereby setting c = 5, A1 = 0.5, A2 = 0.1, A3 = 0.05, etc. At this time, for example, for an image Figure 8B For the passenger 51g sitting in the rear right seat looking forward, x1 = 0, x2 = 3m, x3 = 1m, and y = 4.65. Figure 8C Like in the back left seat Figure 5B For passenger 51n wearing a mask, the mask hides their nose and mouth, so x1 = 2, x2 = 3m, and x3 = 3m, resulting in y = 3.55. In this case, the accuracy of the acquisition for passenger 51n decreases by approximately three times compared to passenger 51g, so the target value change range for vehicle motion, described later, increases by approximately three times.

[0064] Furthermore, the method shown here uses a formula to obtain a quantitative value y, but the structure of the formula is not limited to this. It can also be expressed using a high-order function or a nonlinear function, and the quantitative value can be obtained using a lookup table corresponding to the value of each element.

[0065] Furthermore, occupants may change their head posture regardless of vehicle motion. In this case, if the occupant's behavior can be identified using image recognition technology, for example, from the images acquired by the head posture acquisition unit 23, the acquisition state determination unit 25 can use this information to isolate head posture changes caused by vehicle motion. Furthermore, given that head posture may change regardless of vehicle motion, it is preferable to determine acquisition accuracy only when vehicle motion, particularly translational acceleration, reaches a certain value or above.

[0066] The vehicle motion acquisition unit 26 acquires vehicle motion information indicating the state of motion of the vehicle. Specifically, the vehicle motion acquisition unit 26 acquires vehicle information such as the speed, acceleration, and angular velocity in the translational and rotational directions of the vehicle. Figure 2 As also shown in [ ], the front-back, left-right, and up-down accelerations and the angular velocities of roll, pitch, and yaw can be acquired from the combined sensor 4. These can also be calculated (estimated) based on the operation amounts of the various actuators, such as the motor 12, brake mechanism 13, steering mechanism 14, and suspension 15. Furthermore, the roll and pitch angles can be acquired based on information from the stroke sensors of the suspension 15 mounted on each wheel.

[0067] The head swing prediction unit 27 predicts and outputs the occupant's head swing based on at least one of the target value 22 generated by the target value generation unit 21 and the vehicle motion information acquired by the vehicle motion acquisition unit 26, as well as the occupant's head posture 24 acquired by the head posture acquisition unit 23. Specifically, the head swing prediction unit 27 learns and predicts the occupant's head swing tendency based on at least one of the target value 22 and the vehicle motion information, as well as the occupant's head posture 24, thereby predicting the occupant's susceptibility to motion sickness (motion sickness sensitivity) and calculating a motion sickness sensitivity index 76.

[0068] In this case, in this embodiment, the head swing prediction unit 27 modifies the output based on the acquisition status acquired by the acquisition status determination unit 25. Specifically, the head swing prediction unit 27 modifies the head swing prediction result based on the acquisition status. More specifically, the acquisition status determination unit 25 determines the acquisition accuracy of the head posture as the acquisition status. If the acquisition accuracy is low, the head swing prediction unit 27 modifies the prediction result to produce a more wobbly prediction result compared to a prediction result that does not consider the acquisition accuracy. As described in detail below, the head swing prediction unit 27 of this embodiment changes the predicted head swing tendency based on the acquisition status acquired by the acquisition status determination unit 25.

[0069] A known example of a motion sickness sensitivity index 76 for evaluating a passenger's susceptibility to motion sickness is the motion sickness incidence (MSI), which represents the incidence of motion sickness. The motion sickness incidence (MSI) can be calculated based on three inputs: "three-axis head acceleration + gravitational acceleration," "three-axis head angular velocity," and "three-axis head acceleration." Here, "head acceleration" and "head angular velocity" refer to the acceleration and angular velocity experienced by the passenger's head while riding in vehicle 1. The motion sickness incidence (MSI) is an indicator of the vehicle motion with which smaller movements are considered less prone to motion sickness. Therefore, it is desirable to generate a target value for vehicle motion that reduces the motion sickness incidence (MSI). Based on the principle of MSI, the occupant's head movement in response to a given vehicle motion is correlated with the incidence of motion sickness. In this embodiment, the motion sickness sensitivity index 76 is described based on the MSI. As will be described later, the target value correction unit 28 corrects and outputs the target value based on the predicted head swing result so as to reduce the occupant's head swing.

[0070] Another example of the motion sickness sensitivity index 76 is the MSDV (Motion Sickness Dose Value). This value is obtained by selecting specific frequency components of acceleration generated by the human body, which are believed to be particularly prone to motion sickness. Generally speaking, the higher the value, the more prone to motion sickness. Therefore, focusing on this sensitivity index, target values ​​are generated for controlling vehicle movement such as forward, backward, left, right, and up and down acceleration to avoid the generation of these specific frequency components.

[0071] Furthermore, the motion sickness sensitivity index 76 is not limited to this. For example, the current degree of motion sickness can be calculated based on the occupant's biological signals (sweating, pulse, etc.), or it can be calculated based on input related to the occupant's susceptibility to motion sickness or based on the occupant's past history of motion sickness.

[0072] use Figures 10A to 12 An example of the behavior of the head movement prediction unit 27 will be described.

[0073] Figure 10A and Figure 10B This is an explanatory diagram showing an example of a head swing prediction method in the head swing prediction unit of the vehicle integrated control device according to the first embodiment.

[0074] Figure 10A The diagram shows a situation in which a head roll angle 74 is generated due to a lateral acceleration 72 applied to the occupant 51 . Figure 10B An example of using a common spring-mass-damper mechanical model as a head motion model is shown. Figure 10A and Figure 10BThe roll direction is used as an example for explanation, but the pitch direction can also be expressed with the same model.

[0075] First, if Figure 10A As shown in FIG, when a lateral acceleration 72 is applied to the occupant 51, the head generates an inertial acceleration, resulting in a head roll angle 74. At this point, we know that the neck of the occupant 51 (the connection between the shoulders and the head) generally has the characteristics of a spring (generating a reaction force proportional to the displacement) and a damper (generating a reaction force proportional to the time variation of the displacement). If this structure is simplified into a head motion model, then Figure 10B As shown, it can be represented as a structure in which the inertia 93 is connected to the fixed end via the spring 91 and the damper 92.

[0076] The dynamic input to this model is the inertial acceleration generated at the center of gravity of the inertial mass 93, and the resulting displacement 94 is equivalent to the head roll angle 74. By assuming this model, it is possible to infer the temporal changes in the occupant's head roll angle 74 caused by the temporal changes in the lateral acceleration 72 set at the target value 22.

[0077] Figure 10B The coefficients of the spring 91 and damper 92 shown in FIG are considered to vary from person to person (there are individual differences), and the head swing prediction unit 27 learns the coefficients of the spring 91 and damper 92. Generally speaking, the larger the coefficients of the spring 91 and damper 92, the smaller the head swing.

[0078] Figure 11 This is an explanatory diagram showing an example of driving data for learning coefficients of head swing prediction in the head swing prediction unit of the vehicle integrated control device of the first embodiment. Figure 11 An example of learning the coefficients of the spring 91 and the damper 92 will be described. Figure 11 In the middle graph, the vertical axis represents lateral acceleration, and the horizontal axis represents time. Figure 11 In the lower graph, the vertical axis represents the head roll angle, and the horizontal axis represents time.

[0079] Figure 11 As shown in the upper diagram, a scenario is presented in which vehicle 1 changes lanes from left to right on a two-lane road. Lateral acceleration 72 is generated, as shown in the middle diagram. Specifically, steering is first performed to the right, generating negative lateral acceleration 72. Subsequently, steering is performed to the left, generating positive lateral acceleration 72. At this point, the head roll angles 74 corresponding to the same lateral acceleration 72 differ for occupant 51a, whose head is less likely to move, and occupant 51b, whose head is more likely to move, as 74a and 74b, respectively. Therefore, these are shown as the original head roll angles 74a and 74b.

[0080] Furthermore, even for the same occupant 51b, if the head posture acquisition accuracy is low, the acquired head roll angle 74 is smaller than the original head roll angle 74b acquired with higher accuracy. Here, the acquired head roll angle 74 for an occupant with low head posture acquisition accuracy is assumed to be the same magnitude as 74a and is shown as head roll angle 74b'.

[0081] Figure 12 This is an explanatory diagram showing an example of a coefficient learning result for head swing prediction in the head swing prediction unit of the vehicle integrated control device according to the first embodiment. Figure 12 The vertical axis of the upper graph is the head posture acquisition accuracy 95, and the horizontal axis is the time. Figure 12 The vertical axis of the lower graph represents the learned value of the head firmness coefficient, and the horizontal axis represents time.

[0082] Figure 12 Shown in Figure 11 In the illustrated driving scene, the acquisition state determination unit 25 acquires the aforementioned head posture acquisition accuracy 95 as an acquisition state and learns the coefficients of the spring 91 and the damper 92 in consideration of the acquisition state.

[0083] First, in the upper graph, head posture acquisition accuracy 95a and 95b illustrate examples of time-series changes in head posture acquisition accuracy 95 calculated for occupant 51a and occupant 51b, respectively, according to equation (1). The acquisition accuracy for both occupants remains generally high. On the other hand, head posture acquisition accuracy 95b' illustrates an example where head posture acquisition accuracy 95 for the same occupant 51b remains consistently low.

[0084] The lower graph shows the behavior change of the head swing prediction unit 27 caused by the difference of 95% in the head posture acquisition accuracy mentioned above. Figure 11 The coefficients of the spring 91 and the damper 92 are learned based on the time series relationship between the lateral acceleration 72 and the head roll angle 74 shown in FIG.

[0085] When the head posture acquisition accuracy is 95%, for example Figure 11In the figure, the amplitude of the head roll angle 74b is about twice the size of the head roll angle 74a, so it is assumed that the spring coefficient 91b of the occupant 51b in the identified spring coefficient will be approximately half the size of the spring coefficient 91a of the occupant 51a. It is assumed that the damper coefficients 92a and 92b also show the same tendency. Shown here is an example in which the learning of the coefficients of the spring 91 and the damper 92 at the initial moment is not completed and the learning is completed after the initial value and the single-point dashed line time point. After the learning is completed, the learned value of the coefficient for the occupant 51b (spring coefficient 91b, damper coefficient 92b) is approximately half of the learned value of the coefficient for the occupant 51a (spring coefficient 91a, damper coefficient 92a). Furthermore, Figure 12 The lower graph shows that the higher the direction, the more stable it is and therefore less likely to shake.

[0086] On the other hand, even for the same occupant 51b, if the acquisition accuracy is low, as in the case of head posture acquisition accuracy 95b', the amplitude of the acquired head roll angle 74b' will be smaller than the original head roll angle 74b. Therefore, if learning is performed using the head roll angle 74b' without considering acquisition accuracy, the learned coefficient values ​​for occupant 51b (spring coefficient 91b', damper coefficient 92b') will be roughly equivalent to the learned coefficient values ​​for occupant 51a (spring coefficient 91a, damper coefficient 92a). As a result, using such conventional prediction results that disregard acquisition accuracy for driving control will result in inadequate driving control for occupant 51b who is highly susceptible to motion sickness.

[0087] Therefore, when the acquisition accuracy is low, the head swing prediction unit 27 of this embodiment corrects the prediction result so that it is more likely to swing compared to the prediction result without considering the acquisition accuracy. Specifically, when the head posture acquisition accuracy 95b' is low, the prediction result (output) is corrected so that it is smaller than the learned values ​​of the coefficients (spring coefficient 91b', damper coefficient 92b') that are the usual prediction results without considering the acquisition accuracy, that is, the learned values ​​of the coefficients that are more likely to swing (spring coefficient 91b", damper coefficient 92b") are smaller. Furthermore, here, correction is made so that it is less likely to swing compared to the original learned values ​​of the coefficients (spring coefficient 91b, damper coefficient 92b) to avoid over-correction. Thus, when the acquisition accuracy is low, the occupant's susceptibility to motion sickness is increased and the coefficients are lowered, that is, the prediction is corrected in a direction that makes the head more likely to swing.

[0088] Furthermore, as mentioned above, during learning, considering that the occupant's head posture may change regardless of vehicle motion, it is ideal to perform learning only when the vehicle motion, especially the acceleration in the translational direction, reaches a certain value or above.

[0089] The target value correction unit 28 corrects the target value 22 to reduce head motion based on the head motion prediction result (learned coefficient value) predicted by the head motion prediction unit 27, and outputs the corrected target value 22 as a final target value 29. Specifically, the target value correction unit 28 corrects the input target value 22, generates target values ​​for other types not input, and outputs the final target value 29. The target value correction unit 28 functions to generate a vehicle motion target that takes into account ride comfort and motion sickness reduction. It uses the prediction result (learned coefficient value) predicted by the head motion prediction unit 27 to generate the final target value 29 that optimizes the motion sickness sensitivity index 76. The target value correction unit 28 corrects at least one of the vehicle 1's travel speed, front-to-back, left-to-right, and vertical accelerations, jerk, roll, pitch, and yaw angles, angular velocity, and angular acceleration within the target value 22. Jerk, the time derivative of acceleration, is highly correlated with ride comfort and is therefore a candidate for correction.

[0090] In this embodiment, the head swing prediction unit 27 corrects the prediction result in consideration of the acquisition accuracy. The target value correction unit 28 then uses the corrected prediction result to perform driving control, thereby preventing driving control from becoming insufficient for the occupant 51b who is highly susceptible to motion sickness.

[0091] use Figures 13 to 17 , an example of the behavior of the target value correction unit 28 is described.

[0092] Figure 13 This is an explanatory diagram showing an example of a driving scenario for explaining the behavior of the target value correction unit of the vehicle integrated control device according to the first embodiment.

[0093] like Figure 13 As shown, the road shape illustrated here is a left curve, and vehicle 1 is entering this left-curve road. The driving maneuver performed here is a left turn. The road shown here is divided into a first section (A) with a curvature of zero (straight line), a second section (A-B) with gradually increasing curvature (monotonically increasing curvature: increasing lateral acceleration), a third section (B-C) with a constant curvature (constant curve), a fourth section (C-D) with gradually decreasing curvature (monotonically decreasing curvature: decreasing lateral acceleration), and a fifth section (D-D) with a curvature of zero (straight line).

[0094] Figure 14 This is an explanatory diagram showing an example of the behavior of the target value correction unit of the vehicle integrated control device according to the first embodiment. Figure 14The vertical axis of the first graph is lateral acceleration 72, the vertical axis of the second graph is head roll angle 74, and the vertical axis of the third graph is motion sickness sensitivity index 76. The horizontal axes of the first to third graphs represent positions. The positions are the distances moved from the position immediately before the curve. The dot-dash lines correspond to Figure 13 The positions of the intervals A to D are shown.

[0095] Here, it is assumed that the vehicle is traveling through the curve section A to D while the speed 71 is kept constant. It is assumed that there are two passengers 51a and 51b in the vehicle 1. Generally speaking, even with the same vehicle behavior (lateral acceleration, roll angle), the head swing will be different for each passenger, or even for the same passenger, the head swing will be different depending on the task during the ride. There are individual differences, that is, for a certain lateral acceleration 72, the head swing of the passenger 51a is small, and for the same lateral acceleration 72, the head swing of the passenger 51b is large. In this case, if Figure 14 As shown in the first graph, lateral accelerations 72a and 72b generated by passengers 51a and 51b are identical. Lateral accelerations 72a and 72b increase as the curvature gradually increases in the second interval (A-B), remain constant in the third interval (B-C) where the curvature is constant (constant turn), and gradually decrease in the fourth interval (C-D) where the curvature gradually decreases.

[0096] At this point, the second graph compares the head roll angles 74 of the two individuals. First, as shown by head roll angle 74a, the head roll of passenger 51a evolves in a manner roughly similar to lateral acceleration 72a. Furthermore, lateral acceleration 72 is defined as positive when the vehicle 1 is moving to the left relative to the direction of travel, that is, when the vehicle 1 turns left. In this case, the calculated head roll angle 74 is also positive (clockwise relative to the direction of travel) if it is tilted toward the outside (right) of the turn. Meanwhile, the head roll angle 74b of passenger 51b has a waveform similar to that of head roll angle 74a, but its absolute value is larger. This is because, as mentioned above, individual differences exist: passenger 51b's head swings more significantly in response to a given lateral acceleration 72.

[0097] In this vehicle behavior, an example of the temporal evolution of MSI is shown as the motion sickness sensitivity index 76, as shown in the third graph: motion sickness sensitivity indices 76a and 76b. When vehicle 1 approaches a left curve, the occupant's head experiences lateral acceleration. Furthermore, due to inertia, the head swings toward the outside of the curve, generating a head roll angle. Due to these two effects, the increasing trend is particularly pronounced in the second interval (A-B) and the fourth interval (C-D), where lateral acceleration 72 and head roll angle 74 vary. The rate of increase gradually decreases in the interval where lateral acceleration 72 and head roll angle 74 remain constant. Furthermore, MSI is also caused by fore-aft acceleration and pitch angle. Therefore, if vehicle 1 is accelerating or decelerating in a curve, or before or after it, the MSI will begin to increase immediately before the first interval.

[0098] The evolution of motion sickness sensitivity index 76a of occupant 51a, whose head is less likely to move, is smaller than that of motion sickness sensitivity index 76b of occupant 51b, whose head is more likely to move. According to the principle of MSI, occupant 51b is more likely to feel motion sickness.

[0099] Figure 15 This is an explanatory diagram showing an example of the behavior of the target value correction unit of the vehicle integrated control device according to the first embodiment. Figure 15 The vertical axis of the first graph represents speed 71 , the vertical axis of the second graph represents head roll angle 74 , the vertical axis of the third graph represents motion sickness sensitivity index 76 , and the horizontal axes of the first to third graphs represent positions.

[0100] Figure 15 The example shown here uses a change in speed 71 to create the final target value 29. First, the behavior under normal circumstances, where acquisition accuracy is high, will be described. When passenger 51b negotiates a left curve at the same speed 71a as passenger 51a, passenger 51b's head roll angle 74b is higher than passenger 51a's head roll angle 74a. As a result, motion sickness sensitivity index 76b is higher than motion sickness sensitivity index 76a. Therefore, target value correction unit 28 generates a final target value 29 for speed 71a that is lower than target value 22, similar to speed 71b. Specifically, a modified longitudinal acceleration (not shown) is generated to initiate deceleration earlier than passenger 51a, creating the final target value 29. As a result, lateral acceleration 72 when negotiating a curve is reduced, and occupant 51b's head roll angle 74b becomes equal to head roll angle 74b', which is equivalent to head roll angle 74a. This also reduces motion sickness sensitivity index 76b, which is equal to motion sickness sensitivity index 76b', which is equivalent to motion sickness sensitivity index 76a. Specifically, for occupant 51b, who is highly susceptible to motion sickness, the speed when negotiating a curve is reduced, which suppresses the resulting head shaking and mitigates the onset of motion sickness.

[0101] Next, the behavior in the case of low acquisition accuracy will be described. In the case of low acquisition accuracy, the amplitude of the head roll angle 74 of occupant 51b acquired by the head posture acquisition unit 23 becomes smaller than it should be. Therefore, it is assumed here that head roll angle 74b' is recognized to the same degree as head roll angle 74a. If the head swing prediction unit 27 does not consider acquisition accuracy, based on the prediction result of the head swing prediction unit 27, the motion sickness sensitivity index 76b' of occupant 51b becomes equal to the motion sickness sensitivity index 76a. Therefore, the target value correction unit 28 can determine that the speed 71a of the target value 22 remains unchanged, resulting in the possibility that occupant 51b will experience motion sickness.

[0102] In contrast, when the head swing prediction unit 27 outputs a prediction result that has been corrected by making the learned values ​​of the coefficients (spring coefficient 91b', damper coefficient 92b') smaller than the learned values ​​of the coefficients that are the usual prediction results without considering the acquisition accuracy, that is, the learned values ​​of the coefficients (spring coefficient 91b", damper coefficient 92b") that are more likely to swing, the target value correction unit 28 uses the prediction result corrected by the head swing prediction unit 27, and as a result, generates a final target value 29 that reduces the speed 71a of the target value 22 by the same speed 71b". Specifically, since the head of the occupant 51b is more likely to swing, the target value correction unit 28 determines that if the occupant 51b passes the left curve at the same speed 71a as the occupant 51a, the head roll angle of the occupant 51b may become larger than that of the occupant 51a, and thus generates the final target value 29 that is reduced by making the speed 71b" smaller than the speed 71a of the target value 22. Specifically, a modified longitudinal acceleration (not shown) is generated as final target value 29 to start decelerating earlier than occupant 51a. Consequently, head roll angle 74b" becomes smaller than original head roll angle 74b, and motion sickness sensitivity index 76b" is expected to be smaller than 76b. This prevents insufficient motion sickness reduction effects for occupant 51b, whose head is judged to be less prone to movement despite being prone to movement due to poor acquisition status.

[0103] Setting speed 71b" to the same value as speed 71b is also an option, but the inherently low accuracy of acquiring head roll angle 74b must be taken into consideration. Specifically, if occupant 51b is not prone to moving their head, setting speed 71b" to the same value as speed 71b would result in an overcorrection. While this would reliably prevent motion sickness, it could cause the occupant to perceive the vehicle's movement as slow. Therefore, while the acquisition accuracy is low and it is impossible to determine the occupant's type, prioritizing the possibility that changes to the target value would be insufficient if the occupant were highly sensitive to motion sickness, ideally, the correction amount should be set to a level somewhere in between.

[0104] Furthermore, this figure illustrates an example in which the final target value 29 is generated such that the motion sickness sensitivity index 76b becomes a motion sickness sensitivity index 76b' equivalent to the motion sickness sensitivity index 76a. However, the final target value 29 can also be generated such that the absolute value or rate of change of the motion sickness sensitivity index 76 falls within a predetermined range. The following description also uses the MSI as the motion sickness sensitivity index and an example in which the final target value 29 is generated such that the motion sickness sensitivity index 76b becomes a motion sickness sensitivity index 76b' equivalent to the motion sickness sensitivity index 76a. However, as described above, behavioral changes can be made.

[0105] Furthermore, in Figure 4A In such a case where there are multiple passengers, the passenger with the highest motion sickness sensitivity index is selected, and the final target value 29 is generated so as to suppress the motion sickness sensitivity index of the passenger.

[0106] Figure 16 This is an explanatory diagram showing an example of the behavior of the target value correction unit of the vehicle integrated control device according to the first embodiment.

[0107] Figure 16 2 shows an example in which the roll angle 73 of the vehicle 1 is changed as the final target value 29 without changing the speed 71 . Figure 16 In, with Figure 15 The difference is that the vertical axis of the first graph is the roll angle, and the others are the same as Figure 15 Since they are the same, the explanation will focus on the differences.

[0108] When occupant 51b negotiates a left curve while riding in a vehicle at roll angle 73a, which is appropriate for target value 22 for occupant 51a, motion sickness sensitivity index 76b becomes higher than motion sickness sensitivity index 76a. Consequently, target value correction unit 28 generates final target value 29, which is tilted toward the inside of the curve, similar to roll angle 73b. As a result, lateral acceleration 72 due to inertia experienced by occupant 51b's head during the curve is reduced, and motion sickness sensitivity index 76b becomes 76b', equivalent to 76a.

[0109] Here, when the acquisition accuracy is low, the target value correction unit 28 uses the prediction result corrected by the head swing prediction unit 27, and as a result, generates a final target value 29 obtained by correcting the roll angle 73a of the target value 22 as in the roll angle 73b". In this case, the head roll angle 74b" and the motion sickness sensitivity index 76b" are based on the Figure 15 The same method is used to illustrate the diagram.

[0110] As mentioned earlier, motion sickness can be caused not only by lateral acceleration and roll angle when entering a curve, but also by fore-aft acceleration and pitch angle during acceleration and deceleration. Therefore, excessive deceleration before a curve can actually cause motion sickness in some passengers. Therefore, it's possible to adjust target values ​​other than speed when negotiating a curve.

[0111] Figure 17 This is an explanatory diagram showing an example of the behavior of an actuator for realizing the behavior of the target value correction unit of the vehicle integrated control device of the first embodiment.

[0112] Figure 17 In the following, an example of a method of changing the roll angle by the motor 12 is described. Figure 16 When the roll angle is generated toward the inside of the turn (to the left when facing the direction of travel), first Figure 17 As shown on the left side, the motor 12 is used to move the left front wheel 11 FL and right rear wheel 11 RR Generate positive driving force, let the right front wheel 11 FR and left rear wheel 11 RL As a result, the vehicle 1 generates a negative driving force. Figure 17 As shown on the right side of the figure, the suspension forces are in the DOWN direction on the left and in the UP direction on the right. These suspension forces tilt the vehicle 1 to the left, thereby achieving posture control of the vehicle 1 that conforms to the final target value 29 generated as an internal command. Alternatively, an actuator may be mounted on the suspension 15 to directly generate the suspension forces.

[0113] According to the vehicle integrated control device 2 of the present embodiment 1 described above, the coefficient of the head swing prediction is appropriately changed according to the acquisition status, especially the acquisition accuracy, of the head posture acquisition unit, thereby preventing the situation in which the target value change becomes insufficient due to the increased sensitivity to motion sickness when the acquisition accuracy is low, thereby more reliably achieving vehicle motion control that seeks to reduce motion sickness. Example 2

[0114] Example 2 is a variation of Example 1. While Example 1 focused on acquisition accuracy when describing head posture acquisition, acquisition accuracy is generally defined as detection accuracy when the head is visible. On the other hand, changes in the relative position between camera 100 and occupant 51's head can sometimes cause the occupant's head to become completely invisible. Therefore, Example 2 considers this situation when the head is completely invisible when determining acquisition status.

[0115] Figure 18 It is an explanatory diagram showing an example of the behavior of the acquisition state determination unit of the vehicle integrated control device according to the second embodiment. Figure 18The vertical axis of the upper graph represents the head posture acquisition accuracy, and the horizontal axis represents the time. Figure 18 In the lower graph, the vertical axis represents the learned value of the head firmness coefficient, and the horizontal axis represents time.

[0116] In this embodiment, the head posture acquisition accuracy is defined as 0 when the occupant's head posture becomes completely invisible. Figure 18 In the example, with Figure 4B and Figure 8A As in the example, the camera 100 is set on the front surface of the vehicle 1 and faces rearward. It is assumed that the passenger 51 sits in the left rear seat and initially leans forward to observe the front but starts to observe the scenery on the left side (right side in the image) halfway.

[0117] Focus on Figure 18 The upper graph shows the temporal evolution of head pose acquisition accuracy 95. When occupant 51 is looking straight ahead, head pose acquisition accuracy 95 is high. However, as occupant 51 gradually moves to the left (right in the image), head pose acquisition accuracy 95 decreases. After occupant 51 begins to observe the scenery to the left, the passenger seat headrest completely obscures occupant 51's head, causing head pose acquisition accuracy 95 to drop to 0. Thus, acquisition state determination unit 25 sets the acquisition accuracy to 0 when the head is no longer visible.

[0118] at this time, Figure 18 The chart below shows Figure 10B The figure shows an example of learning the coefficients of the spring 91 and the damper 92 for the head motion model. Originally, the head swing prediction unit 27 estimated a high coefficient based on the head motion of the occupant 51. However, the occupant 51's head became completely invisible, causing the acquisition accuracy to drop to zero. Shortly after this point, the coefficients were corrected and calculated to achieve a low coefficient, meaning a prediction result that is more prone to swing.

[0119] Various methods are contemplated for setting the changed numerical values, such as setting the coefficient based on the head movement behavior of the person most susceptible to motion sickness, setting the coefficient to the lowest value among the coefficient fluctuations of occupant 51, or restoring the coefficient to the default value based on the head movement behavior so that the amplitude of occupant 51's head swinging becomes the amplitude of normal (average) human body swinging. For a person already identified as prone to motion sickness, the coefficient can be set based on the head movement behavior of the person most susceptible to motion sickness, or restored to the default coefficient if learning of the coefficient for that occupant has not yet been fully completed, thereby enabling different responses to be taken depending on the occupant's situation.

[0120] According to the vehicle integrated control device 2 of Example 2, in addition to the acquisition accuracy defined based on the premise that the head is visible, exceptions are also set for actions when the head is not visible. This makes it possible to achieve vehicle motion control that appropriately reduces motion sickness of the occupant 51 in a wider variety of situations. Example 3

[0121] Figure 19 This is a functional block diagram of the vehicle integrated control device of Example 3.

[0122] Embodiment 3 is a modification of Embodiment 1 or 2. The difference from Embodiment 1 or 2 is that the output is corrected by the target value correction unit 28 instead of the head swing prediction unit 27 according to the acquired state.

[0123] In Example 3, the head swing prediction unit 27 outputs a prediction result without considering the acquisition status. The target value correction unit 28 then adjusts the correction amount of the target value 22 based on the acquisition status. Specifically, if the acquisition accuracy is low, the target value correction unit 28 adjusts the correction amount to minimize fluctuation compared to the correction amount that would be obtained without considering the acquisition accuracy.

[0124] For example, in Figure 15 In the example described above, even when the acquisition accuracy of occupant 51b is low, the head swing prediction unit 27 outputs the same prediction results (coefficient learning values) for occupants 51a and 51b, disregarding the acquisition status. When the target value correction unit 28 disregards acquisition accuracy, speed 71a is not changed, so the correction amount for speed 71a in target value 22 is 0. In contrast, when acquisition accuracy is low, the target value correction unit 28 of this embodiment adjusts the correction amount to a value that is less prone to fluctuation than the correction amount that disregards acquisition accuracy, determining the correction amount so that speed 71b" is obtained.

[0125] Figure 16 The same is true in the case of the example. When the acquisition accuracy is not considered, the roll angle 73a is kept unchanged and the correction amount is 0. On the contrary, when the acquisition accuracy is low, the roll angle is corrected to 73b".

[0126] In addition, when combined with Example 2, when the target value correction unit 28 corrects the output according to the acquisition state, when the head of the occupant 51 as the acquisition state is no longer visible, the correction is performed in a manner that makes it less likely to swing than when the head of the occupant 51 is visible.

[0127] Furthermore, while the previous description assumes that the acquisition accuracy and the correction amount to the target value 22 are uniquely associated, even with the same acquisition accuracy, the degree of correction to the target value 22 may vary depending on the duration of the ride, for example. For example, even for the same person, the longer the ride, the less susceptible they may become to motion sickness due to adaptation. Therefore, even with the same acquisition accuracy, the degree of correction to the target value 22 may be reduced as the ride time increases.

[0128] Furthermore, the present invention encompasses various variations and is not limited to the above-described embodiments. For example, the above-described embodiments are provided for the purpose of providing a detailed description of the present invention in an easily understandable manner and are not necessarily limited to all of the described structures. Furthermore, a portion of the structure of one embodiment may be replaced with a structure of another embodiment, and a structure of another embodiment may be added to a structure of another embodiment. Furthermore, a portion of the structure of each embodiment may be added, deleted, or replaced with another structure. Explanation of symbols

[0129] 1…Vehicle 11…Wheels 12…Motor 13…Brake mechanism 13a…wheel cylinder 13b…Brake control device 14…Steering mechanism 14a…Steering controls 14b…Steering motor 15…Suspension 16…Accelerator pedal 16a…travel sensor 16b…Acceleration control device 17…Brake pedal 18…steering wheel 18a…Steering torque detection device 18b…Steering angle detection device 19…External sensors 2…Vehicle integrated control device 21…Target value generation unit 22…target value 23…Head posture acquisition unit 24…Occupant head posture 25…Acquisition status determination unit 26…Vehicle Motion Acquisition Department 27…Head Swing Prediction Department 28…Target value correction unit 29…final target value 3…External control device 4…Combination sensor 51…crew 61…motion sickness 71…speed 72…lateral acceleration 73…roll angle 74…Head tilt angle 75…Driving track 76…Motion Sickness Sensitivity Index 91…spring 92…Damper 93…Inertia 94…Displacement 95…Head posture acquisition accuracy 100... cameras 101…optical axis 110…masks 111…book 112…window.

Claims

1. A vehicle integrated control device, characterized in that: have: a target value generating unit that generates a target value for motion control of the vehicle; a head posture acquiring unit for acquiring the head posture of the occupant; an acquisition state determination unit configured to determine an acquisition state of the head posture; a vehicle motion acquisition unit that acquires vehicle motion information indicating a state of motion of the vehicle; a head swing prediction unit that predicts and outputs the occupant's head swing based on at least one of the target value and the vehicle motion information and the head posture; as well as a target value correction unit configured to correct and output the target value in a manner to reduce the head swing based on a result of the predicted head swing; The head swing prediction unit or the target value correction unit corrects an output based on the acquired state.

2. The vehicle integrated control device according to claim 1, wherein: The head swing prediction unit corrects the head swing prediction result according to the acquired state.

3. The vehicle integrated control device according to claim 2, wherein: The acquisition state determination unit determines the acquisition accuracy of the head posture as the acquisition state. When the acquisition accuracy is low, the head swing prediction unit performs correction so that the prediction result becomes more likely to swing compared to a prediction result without considering the acquisition accuracy.

4. The vehicle integrated control device according to claim 1, wherein: The target value correction unit corrects a correction amount of the target value according to the acquisition state.

5. The vehicle integrated control device according to claim 4, characterized in that: The acquisition state determination unit determines the acquisition accuracy of the head posture as the acquisition state. When the acquisition accuracy is low, the target value correction unit corrects the correction amount so as to be less likely to fluctuate than the correction amount obtained without considering the acquisition accuracy.

6. The vehicle integrated control device according to claim 1, wherein: The head posture acquiring unit acquires sensing information obtained by sensing the head of the occupant, The acquisition status determination unit determines the acquisition status based on the sensing information.

7. The vehicle integrated control device according to claim 6, wherein: The acquisition state determination unit determines the acquisition state based on the number of feature points of the occupant's head in the sensing information.

8. The vehicle integrated control device according to claim 6, wherein: The head posture acquisition unit acquires an image as the sensing information, The acquisition state determination unit determines the acquisition state based on a position of the occupant's head within the image.

9. The vehicle integrated control device according to claim 6, wherein: The head posture acquisition unit acquires an image as the sensing information, The acquisition state determination unit determines the acquisition state based on a ratio of the occupant's head portion in the image.

10. The vehicle integrated control device according to claim 6, wherein: The acquisition state determination unit determines the acquisition state based on a relative positional relationship between a sensor that acquires the sensing information and the head of the occupant.

11. The vehicle integrated control device according to claim 10, wherein: The acquisition state determination unit determines the acquisition state based on a vertical distance from an optical axis of the sensor.

12. The vehicle integrated control device according to claim 10, wherein: The acquisition state determination unit determines the acquisition state based on a distance from the sensor to the occupant's head.

13. The vehicle integrated control device according to claim 1, wherein: When the head swing prediction unit or the target value correction unit corrects the output based on the acquired state, if the acquired state is that the occupant's head is no longer visible, the correction is performed in a manner that makes the predicted result easier to swing than when the head is visible, or in a manner that makes it less likely to swing than when the head is visible.

14. The vehicle integrated control device according to claim 1, wherein: When the acquired state is that the head of the occupant is no longer visible, the head swing prediction unit predicts that the amplitude of the head swing of the occupant becomes the amplitude of the average human body swing.

15. The vehicle integrated control device according to claim 1, wherein: The target value correction unit corrects at least one of the target values, including a running speed, front-rear, left-right, and top-bottom acceleration, jerk, roll, pitch, and yaw angles, angular velocity, and angular acceleration of the vehicle.

Images