Control device

Abstract

The control device is provided with: an acquisition unit that acquires the acceleration of the vehicle body in the vertical direction of the vehicle; and an estimation unit that estimates, using a neural network, the relative speed of the wheels with respect to the vehicle body in the vertical direction of the vehicle, the relative speed corresponding to a value obtained by multiplying the acceleration acquired by the acquisition unit by the mass of the vehicle, the magnitude of the mass being the same as or different from the magnitude of the neural network at the time of learning. And a neural network that learns so as to estimate the relative speed in accordance with an input of a numerical value obtained by multiplying the acceleration and the mass, and that controls the suspension device on the basis of the estimated relative speed.

Description

Technical Field

[0001] An embodiment of the present invention relates to a control device. Background Art

[0002] Conventionally, there is known a technique for estimating a relative speed of a wheel with respect to a vehicle body in a vertical direction of a vehicle, namely, a so-called stroke speed, and controlling the damping force of a shock absorber of a suspension device based on the estimated stroke speed.

[0003] Prior art literature

[0004] Patent Literature

[0005] Patent Document 1: Japanese Patent Application Laid-Open No. 10-913

[0006] Patent Document 2: Japanese Patent Application Publication No. 2016-117326

[0007] Technical problem to be solved by the invention

[0008] When estimating physical quantities such as travel speed using a neural network, it is necessary to create a neural network for each vehicle specification, which requires a lot of effort. Summary of the invention

[0009] Therefore, one of the problems of the embodiments of the present invention is to reduce the time and effort required to create a neural network even when there are multiple specifications of vehicles, for example.

[0010] Technical means for solving technical problems

[0011] A control device involved in an embodiment of the present invention controls a suspension device between a vehicle body and the wheels of the vehicle, and includes: an acquisition unit, which acquires the acceleration of the vehicle body in the up-down direction of the vehicle; and an estimation unit, which uses a neural network to estimate the relative speed of the wheels with respect to the vehicle body in the up-down direction of the vehicle, the relative speed corresponding to a numerical value obtained by multiplying the acceleration acquired by the acquisition unit with the mass of the vehicle, the magnitude of the mass being the same or different when the neural network is learned, the neural network is learned in a manner that estimates the relative speed corresponding to the input of the numerical value obtained by multiplying the acceleration and the mass, and the suspension device is controlled based on the estimated relative speed.

[0012] Effects of the Invention

[0013] According to the control device according to the embodiment of the present invention, for example, the same neural network can be used for vehicles of multiple specifications with different masses. Therefore, even when there are multiple specifications of vehicles, the time and effort required to create the neural network can be reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a schematic diagram showing a schematic structure of an example of a vehicle according to the first embodiment.

[0015] Figure 2 It is a schematic diagram showing a schematic structure of an example of a suspension device in a vehicle according to the first embodiment.

[0016] Figure 3 This is a functional block diagram of a control device of an example of the vehicle according to the first embodiment.

[0017] Figure 4 It is an explanatory diagram for explaining physical quantities of the vehicle according to the first embodiment.

[0018] Figure 5 This is a diagram showing an example of a neural network used for research.

[0019] Figure 6 : is a diagram showing an example of the estimation result of the travel speed based on the neural network for research, and is a diagram showing the case where the mass of the vehicle is used as a reference.

[0020] Figure 7 1 is a diagram showing an example of the estimation result of the travel speed based on the neural network for research, and is a diagram showing a case where the mass of the vehicle is half of the reference.

[0021] Figure 8 1 is a diagram showing an example of the estimation result of the travel speed based on the neural network for research, and is a diagram showing the case where the mass of the vehicle is 1.5 times the reference.

[0022] Fig. 9 1 is a diagram showing an example of the estimated result of the stroke speed based on the motion equation for research, and is a diagram showing a case where the mass of the vehicle is used as a reference.

[0023] Fig.10 1 is a diagram showing an example of the estimated result of the stroke speed based on the motion equation for research, and is a diagram showing a case where the mass of the vehicle is half of the reference.

[0024] Fig.11 1 is a diagram showing an example of the estimated result of the stroke speed based on the motion equation for research, and is a diagram showing a case where the mass of the vehicle is 1.5 times the reference mass.

[0025] Fig.12This is a diagram used to illustrate a neural network used in research.

[0026] Fig.13 It is a diagram showing a neural network involved in the first embodiment.

[0027] Fig.14 : is a diagram showing an example of the estimation result of the travel speed based on the neural network according to the first embodiment, and is a diagram showing the case where the mass of the vehicle is used as a reference.

[0028] Fig.15 1 is a diagram showing an example of the estimation result of the travel speed based on the neural network according to the first embodiment, and is a diagram showing a case where the mass of the vehicle is half of the reference.

[0029] Fig.16 : is a diagram showing an example of the estimation result of the travel speed based on the neural network according to the first embodiment, and is a diagram showing a case where the mass of the vehicle is 1.5 times the reference.

[0030] Fig.17 1 is a diagram showing an example of the estimation result of the stroke speed based on the motion equation of the comparative example, and is a diagram showing the case where the mass of the vehicle is used as a reference.

[0031] Fig.18 1 is a diagram showing an example of the estimation result of the stroke speed based on the motion equation of the comparative example, and is a diagram showing a case where the mass of the vehicle is half of the reference.

[0032] Fig.19 1 is a diagram showing an example of the estimation result of the stroke speed based on the motion equation of the comparative example, and is a diagram showing the case where the mass of the vehicle is 1.5 times the reference.

[0033] Fig. 20 This is a flowchart showing an example of a physical quantity estimation process executed by the control device according to the first embodiment.

[0034] Fig.21 It is a diagram showing a neural network involved in the second embodiment.

[0035] Fig. 22 It is a diagram showing a neural network involved in the third embodiment.

[0036] Fig.23 It is a diagram showing a neural network according to the fourth embodiment.

[0037] Explanation of symbols

[0038] 1…Vehicle

[0039] 2…Body

[0040] 3…Wheels

[0041] 4…Suspension

[0042] 4a…Coil spring (spring)

[0043] 4b…Shock absorber

[0044] 10…Control device

[0045] 10a…Getting

[0046] 10b…Presumption

[0047] 121…Neural Networks. DETAILED DESCRIPTION

[0048] Hereinafter, the embodiments will be described in detail with reference to the accompanying drawings. In addition, the present invention is not limited to the embodiments. The following multiple embodiments include the same components. Thus, the same components are marked with common symbols and repeated descriptions are omitted.

[0049] <First embodiment>

[0050] Figure 1 : is a schematic diagram showing the schematic structure of an example of a vehicle 1 involved in the first embodiment. In the present embodiment, the vehicle 1 can be, for example, a car (internal combustion engine car) driven by an internal combustion engine (engine, not shown), a car (electric car, fuel cell car, etc.) driven by an electric motor (motor, not shown), or a car (hybrid car) driven by both of them. In addition, the vehicle 1 can be equipped with various transmission devices and various devices (systems, components, etc.) required to drive the internal combustion engine and the electric motor. In addition, the method, number, layout, etc. of the devices related to the drive of the wheels 3 in the vehicle 1 can be variously set. In addition, in the present embodiment, as an example, the vehicle 1 is a four-wheeled vehicle (four-wheeled car) having two left and right front wheels 3F and two left and right rear wheels 3R. In addition, in Figure 1 In the figure, the front side (direction Fr) in the front-rear direction of the vehicle is the left side. The vehicle 1 is an example of an object. In addition, the object may be a moving body, a device, etc. other than the vehicle 1.

[0051] In this embodiment, as an example, the vehicle control system 100 of the vehicle 1 includes: a control device 10, a steering device 11, a steering angle sensor 12, a yaw rate sensor 13, a brake system 61, etc. In addition, the vehicle control system 100 includes a suspension device 4, a rotation sensor 5, an acceleration sensor 6, etc. corresponding to the two front wheels 3F, and includes a suspension device 4, a rotation sensor 5, an acceleration sensor 6, etc. corresponding to the two rear wheels 3R. Figure 1In addition, the vehicle 1 also has basic components, but only the structure related to the vehicle control system 100 and the control related to the structure are described here. In addition, the wheel 3 is used as a general term for the two front wheels 3F and the two rear wheels 3R. In addition, the relative speed of the wheel 3 of the vehicle 1 with respect to the body 2 of the vehicle 1 is also called the stroke speed, the part of the vehicle 1 on the body 2 side relative to the suspension device 4 is called the sprung part, and the part of the vehicle 1 on the wheel 3 side relative to the suspension device 4 is called the unsprung part. In addition, in the following description, unless otherwise specified, the relative speed is the relative speed of the wheel 3 with respect to the body 2 in the up and down direction of the vehicle 1.

[0052] The control device 10 obtains signals, data, etc. from various parts of the vehicle control system 100, and performs control of various parts of the vehicle control system 100 and various operations. The control device 10 is also called a state quantity estimation device, a vehicle state quantity estimation device, and a physical quantity estimation device. In addition, the control device 10 is configured as a computer, and has an operation processing unit (microcomputer, ECU (Electronic Control Unit), etc., not shown), a storage unit 10d (for example, ROM (Read Only Memory), RAM (Random Access Memory), flash memory, etc., refer to Figure 3 ) etc. The operation processing unit reads a program stored in (installed in) a nonvolatile storage unit 10d (such as a ROM, flash memory, etc.), and performs operation processing according to the program, and can be used as Figure 3 Each part shown functions (acts). In addition, the storage unit 10d can store data (tables (data groups), functions, etc.) used for various operations related to control, operation results (including values ​​in the middle of the operation), etc. In addition, the storage unit 10d stores the neural network 121. The details of the neural network 121 will be described later.

[0053] The steering device 11 includes, for example, a steering wheel, and steers (turns) the two front wheels 3F. The steering angle sensor 12 detects the steering angle (steering angle, cutting angle, turning angle) of the front wheels 3F, and outputs steering angle data indicating the detected steering angle.

[0054] The yaw rate sensor 13 detects the yaw rate of the vehicle 1 (body 2 ), and outputs yaw rate information indicating the detected yaw rate.

[0055] Figure 2 1 is a schematic diagram showing a schematic structure of an example of a suspension device 4 in a vehicle according to the first embodiment. Figure 2As shown, the suspension device 4 is interposed between the wheel 3 and the vehicle body 2 to suppress the vibration and impact from the road surface from being transmitted to the vehicle body. The suspension device 4 has a coil spring 4a and a shock absorber 4b. The shock absorber 4b can electrically control (adjust) the damping force (damping characteristics). Specifically, the shock absorber 4b has an actuator 4bb that operates based on the input current. The actuator 4bb can change the opening degree of the hole of the piston provided in the shock absorber 4b, or change the opening degree between the valve core and the valve seat. Thus, the flow rate of the lubricating oil flowing between the two oil chambers separated by the piston in the shock absorber 4b is controlled, thereby adjusting the damping force of the shock absorber 4b. The suspension device 4 is provided in each of the four wheels 3 (two front wheels 3F and two rear wheels 3R), and the control device 10 can control the damping force of each of the four wheels 3. The control device 10 can control the four wheels 3 to be in a state where the damping force is different from each other. In detail, the damping coefficient is controlled in order to control the damping force (=damping coefficient×stroke speed) of the shock absorber 4b. More specifically, the damping coefficient is controlled by controlling the current flowing through the actuator 4bb of the shock absorber 4b.

[0056] The rotation sensor 5 can output a signal corresponding to the rotation speed (angular velocity, rotation speed, rotation state) of each of the four wheels 3. The control device 10 can calculate the speed of the vehicle 1 based on the detection result of the rotation sensor 5. In addition, a rotation sensor (not shown) for detecting the rotation of a crankshaft, an axle, etc. may be provided separately from the rotation sensor 5 for the wheel 3, and the control device 10 may also obtain the speed of the vehicle 1 based on the detection result of the rotation sensor.

[0057] The acceleration sensor 6 is provided on the vehicle body 2 corresponding to each of the four wheels 3 and the suspension device 4, and detects the acceleration of the vehicle body 2 in the vehicle 1. In detail, each acceleration sensor 6 is provided at an acceleration detection target portion located directly above each wheel 3 in the vehicle body 2 (directly above the tire). That is, in the present embodiment, there are four acceleration detection target portions. In the present embodiment, as an example, the acceleration sensor 6 can obtain the acceleration in the up-and-down direction of the vehicle 1, the acceleration in the front-and-back direction (long side direction) of the vehicle 1, and the acceleration in the width direction (vehicle width direction, short side direction, left-right direction) of the vehicle 1.

[0058] The above-described configuration of the vehicle control system 100 is merely an example, and various modifications can be made to implement the vehicle control system 100. Well-known devices can be used as each device constituting the vehicle control system 100. In addition, each configuration of the vehicle control system 100 can be shared with other configurations.

[0059] In addition, in the present embodiment, as an example, the control device 10 is implemented by the cooperation of hardware and software (program). Figure 3As shown in FIG. 1 , the program can function (operate) as an acquisition unit 10a, an estimation unit 10b, an attenuation control unit 10c, etc. That is, as an example, the program can include Figure 3 Modules corresponding to each block except the storage unit 10d.

[0060] The acquisition unit 10a acquires acceleration data indicating the acceleration of the vehicle 1 from each acceleration sensor 6. The acceleration of the vehicle 1 may be acceleration in any direction. In addition, the acceleration of the vehicle 1 may be acceleration in multiple directions. That is, the acceleration of the vehicle 1 may be one or more of the acceleration in the front-rear direction of the vehicle 1, the acceleration in the left-right direction of the vehicle 1, and the acceleration in the up-down direction of the vehicle 1. In addition, the acquisition unit 10a acquires a current value (current value data) which is the value of the current input to the shock absorber 4b from the attenuation control unit 10c.

[0061] As an example, the estimation unit 10b is input with the predetermined data (hereinafter also referred to as input data) acquired by the acquisition unit 10a. The estimation unit 10b estimates the travel speed corresponding to the input data using the neural network 121. The input data includes at least acceleration data. In addition to the acceleration data, the input data may also include at least one of yaw rate data, current value data, steering angle data, and wheel speed data. In addition, the input data is not limited to the above.

[0062] Here, a vehicle 1 may have multiple specifications (types, models) for one model. For example, multiple specifications include: gasoline / two-wheel drive specification, hybrid power / two-wheel drive specification, gasoline / four-wheel drive specification, hybrid power / four-wheel drive specification, etc. The gasoline / two-wheel drive specification is that the driving source is a gasoline engine and the driving wheels are two wheels. The hybrid power / two-wheel drive specification is that the driving source is a gasoline engine and an electric motor and the driving wheels are two wheels. The gasoline / four-wheel drive specification is that the driving source is a gasoline engine and the driving wheels are four wheels. The hybrid power / four-wheel drive specification is that the driving source is a gasoline engine and an electric motor and the driving wheels are four wheels. As an example, vehicles 1 of such multiple specifications have different masses. Therefore, in the past, in order to estimate the travel speed of each vehicle 1, even if it is a single model, a neural network is created for each vehicle 1. Creating a neural network requires measuring input data and teacher data for learning, and creating multiple neural networks requires time.

[0063] Therefore, the inventors of the present application have studied a structure (method) that can reduce the time and effort required for creating a neural network even when there are a plurality of specifications of the vehicle 1. The following describes the contents of the study conducted by the inventors.

[0064] The inventors compared the neural network with the equations of motion that represent the motion of the vehicle 1. Here, Figure 4 1 is an explanatory diagram for explaining physical quantities of the vehicle 1 according to the first embodiment. Figure 4As shown, the mass of the vehicle 1 is Mb, the sprung displacement amount which is the displacement amount of the vehicle body 2 in the vertical direction is xb, the spring constant of the coil spring is Kc, the damping coefficient of the shock absorber 4b is Cx, the unsprung displacement amount which is the displacement amount of the wheel 3 in the vertical direction is xt, the spring constant of the tire of the wheel 3 is Kt, and the displacement amount of the lower end of the tire is xr. In this case, the motion equation of the vehicle 1 is expressed by the following equation (1).

[0065] Mb(xb)″=-Kc(xb―xt)-Cx((xb)′-(xt)′)·····(1)

[0066] In addition, "'" and """ in the formula (1) represent the first differential and the second differential, respectively.

[0067] Therefore, the estimated stroke speed ((xb)′-(xt)′) is expressed by equation (2).

[0068] ((xb)′-(xt)′)=-(Mb(xb)″) / Cx-(Kc(xb―xt)) / Cx·····(2)

[0069] Figure 5 2 is a diagram showing an example of a neural network 200 for research. Figure 5 As shown, the input data of the neural network 200 used for the study are the sprung acceleration and the current value of the shock absorber 4b, and the output data is the stroke speed.

[0070] The neural network 200 and equations (1) and (2) are created based on the reference mass of the vehicle 1 .

[0071] The inventors calculated the estimated value of the stroke speed based on the neural network 200 and equation (2) under the condition of changing the mass. Figure 6 to Figure 11 Shown. Figure 6 1 is a diagram showing an example of the estimation result of the travel speed based on the neural network 121 for research, and is a diagram showing a case where the mass of the vehicle 1 is used as a reference. Figure 7 1 is a diagram showing an example of the estimation result of the travel speed based on the neural network 121 for research according to the first embodiment, and is a diagram showing a case where the mass of the vehicle 1 is half of the reference mass. Figure 8 1 is a diagram showing an example of the estimation result of the travel speed based on the neural network 121 for research, and is a diagram showing a case where the mass of the vehicle 1 is 1.5 times the reference. Fig. 9 1 is a diagram showing an example of the estimation result of the stroke speed based on the motion equation for research, and is a diagram showing a case where the mass of the vehicle 1 is used as a reference. Fig.101 is a diagram showing an example of the estimated result of the stroke speed based on the motion equation for research, and is a diagram showing a case where the mass of the vehicle 1 is half of the reference mass. Fig.11 1 is a diagram showing an example of the estimated result of the stroke speed based on the motion equation for research, and is a diagram showing a case where the mass of the vehicle 1 is 1.5 times the reference mass.

[0072] like Figure 6 and Fig. 9 As shown in , when the input quality is set as the reference quality, both the neural network 200 and the equation (2) obtain results in which the estimated value is close to the true value. Figure 7 and Fig.10 As shown in FIG. 1 , when the input quality is set to half of the reference, both the neural network 200 and the equation (2) obtain a result that the deviation between the estimated value and the true value becomes larger. Figure 8 and Fig.11 As shown, when the input quality is set to 1.5 times the reference, both the neural network 200 and the formula (2) obtain a result that the deviation between the estimated value and the true value becomes larger. That is, when an erroneous value is intentionally input as an input value, the deviation between the estimated value and the true value becomes larger. This deviation is considered to be an error.

[0073] Based on the above results, the inventors found that the deviation between the estimated value and the true value, i.e., the error of the estimated value, is similar to that in the case where an erroneous value is intentionally input as an input value in the neural network 200 and in equation (2). Fig.12 As shown, the inventors believe that there is a portion 200a corresponding to the parameter of quality in the neural network 200, and therefore believe that the parameter of quality can be removed from the neural network 200. Furthermore, the inventors have thought of a structure in which the parameter of quality is removed from the neural network 121 of this embodiment.

[0074] The neural network 121 is described below. Fig.131 is a diagram showing a neural network 121 according to the first embodiment. The neural network 121 is a structure that is pre-learned so that the relative speed of the wheel 3 of the vehicle 1 relative to the body 2 of the vehicle 1 in the vertical direction of the vehicle 1, that is, the stroke speed, is estimated based on the input data. For example, the neural network 121 is composed of LSTM (Long short-term memory). When learning the neural network 121, the input data is the numerical value of the force obtained by multiplying the sprung acceleration and the mass and the current value of the shock absorber 4b. The sprung acceleration is the acceleration detected by the acceleration sensor 6, that is, the acceleration of the body 2 of the vehicle 1 in the vertical direction. The mass is the reference mass. That is, the mass is a constant. The output data as the teacher data is, for example, the relative speed of the wheel 3 of the vehicle 1 relative to the body 2 of the vehicle 1 measured by the stroke speed sensor (not shown), that is, the measured value (true value) of the stroke speed. The sprung acceleration is an example of the first physical quantity, the mass is an example of the second physical quantity, and the stroke speed is an example of the third physical quantity.

[0075] Fig.14 1 is a diagram showing an example of the estimation result of the travel speed based on the neural network 121 according to the first embodiment, and is a diagram showing a case where the mass of the vehicle 1 is used as a reference. Fig.15 1 is a diagram showing an example of the estimation result of the travel speed based on the neural network 121 according to the first embodiment, and is a diagram showing a case where the mass of the vehicle 1 is half of the reference mass. Fig.16 1 is a diagram showing an example of the estimation result of the travel speed based on the neural network 121 according to the first embodiment, and is a diagram showing a case where the mass of the vehicle 1 is 1.5 times the reference.

[0076] like Figure 14 to Figure 16 As shown, for the neural network 121, even if the quality of the calculation used for the input data is changed, a result is obtained in which the estimated value and the true value are substantially consistent.

[0077] Here, even if the mass eliminated from the neural network 121 is not multiplied by the sprung acceleration, and only the mass is input as input data to the neural network 121, the neural network 121 does not recognize the mass because the mass is not learned as a parameter. Therefore, even if the mass is changed and input to the neural network 121, the deviation between the estimated value and the true value will increase. In contrast, in the present embodiment, since the mass is multiplied by the sprung acceleration, the multiplied value of the mass and the sprung acceleration is recognized by the neural network 121 as a parameter. Therefore, even if the mass is changed, the deviation between the estimated value and the true value is suppressed from increasing. Therefore, for multiple vehicles 1 with different masses, the same neural network 121 can be used. That is, the neural network 121 produced for one of the multiple specifications of vehicles 1 can be used for other vehicles 1 with different specifications. In such a neural network 121, the mass (second physical quantity) of the vehicle 1 can be the same as that when the neural network 121 is learned, or it can be different. For example, when the neural network 121 is created for a gasoline / two-wheel-drive vehicle 1, the mass of the vehicle 1 used for the neural network 121 is the same as that during learning for the neural network 121 provided for the gasoline / two-wheel-drive vehicle 1. On the other hand, for the neural network 121 provided for the vehicle 1 of other specifications (hybrid / two-wheel-drive, gasoline / four-wheel-drive, hybrid / four-wheel-drive), the mass of the vehicle 1 used for the neural network 121 is different from that during learning.

[0078] here, Figure 14 to Figure 16 The waveform shown is similar to Figure 6 to Figure 8 The waveforms shown are different because the model of the vehicle 1 used for the simulation is different. Figure 14 to Figure 16 The waveform of the travel speed of the motion equation (comparative example) produced by the vehicle 1 model shown in the waveform is Figure 17 to Figure 19 Shown in.

[0079] return Figure 3 The damping control unit 10c controls the damping force of the shock absorber 4b based on the stroke speed estimated by the estimating unit 10b. Specifically, the damping control unit 10c determines the current value to be input to the shock absorber 4b, and inputs the current value to the shock absorber 4b.

[0080] Next, refer to Fig. 20 , an example of the physical quantity estimation direction executed by the control device 10 is described. Fig. 20 This is a flowchart showing an example of the physical quantity estimation process executed by the control device 10 according to the first embodiment.

[0081] like Fig. 20As shown, first, the acquisition unit 10a acquires input data (data) (S1). The input data is, for example, the sprung acceleration and the current value of the shock absorber 4b. Next, the estimation unit 10b inputs the value obtained by the acquisition unit 10a multiplying the sprung acceleration by the mass of the vehicle 1 and the current value acquired by the acquisition unit 10a into the neural network 121, thereby estimating the physical quantity as the estimation object through the neural network 121 (S2). The physical quantity is, for example, the stroke speed. The mass is stored in the storage unit 10d. That is, the storage unit 10d stores the second physical quantity input to the neural network 121.

[0082] As described above, in the present embodiment, the control device 10 includes an acquisition unit 10a and an estimation unit 10b. The acquisition unit 10a acquires a first physical quantity as a variable related to the vehicle 1 (object). The estimation unit 10b estimates a third physical quantity corresponding to a numerical value obtained by multiplying the first physical quantity acquired by the acquisition unit 10a by a second physical quantity that is the same or different in size from that when the neural network 121 is learned, using a neural network 121, which estimates the third physical quantity related to the vehicle 1 in accordance with the input of a numerical value obtained by multiplying the first physical quantity by the second physical quantity related to the vehicle 1. As an example, the first to third physical quantities are physical quantities related to the movement of the vehicle 1.

[0083] According to such a configuration, for example, the same neural network 121 can be used for a plurality of vehicle specifications 1 having different second physical quantities. Therefore, even when there are a plurality of vehicle specifications 1, the time required to create the neural network 121 can be reduced.

[0084] The first physical quantity is the acceleration of the vehicle body 2 in the vertical direction of the vehicle 1, and the second physical quantity is the mass of the vehicle 1. The third physical quantity is the relative speed (stroke speed) of the wheels 3 with respect to the vehicle body 2 in the vertical direction of the vehicle 1.

[0085] According to such a configuration, even when a plurality of specifications of different masses exist in the vehicle 1 , the time and effort required to prepare the neural network 121 can be reduced.

[0086] <Second embodiment>

[0087] Fig.21 It is a diagram showing a neural network 121 according to the second embodiment.

[0088] like Fig.21 As shown, the present embodiment is different from the first embodiment mainly in that the input data input to the neural network 121 are the numerical value of the force obtained by multiplying the sprung acceleration by the mass of the vehicle 1 and the damping force of the shock absorber 4 b .

[0089] The damping force of the shock absorber 4b is calculated based on the current value of the shock absorber 4b and the stroke speed estimated last by the estimating unit 10b in the estimation of the stroke speed (third physical quantity) performed by the estimating unit 10b. Specifically, the damping force is calculated based on the damping coefficient of the shock absorber corresponding to the current value of the shock absorber 4b and the stroke speed estimated last by the estimating unit 10b. The estimating unit 10b multiplies the damping coefficient of the shock absorber 4b or a quantity related to the damping coefficient by the stroke speed (relative speed) estimated last time, thereby estimating a new stroke speed (relative speed). The stroke speed (third physical quantity) estimated last time by the estimating unit 10b is equivalent to the first physical quantity in the estimation of the stroke speed performed by the estimating unit 10b this time. In addition, the current value of the shock absorber 4b is an example of the second physical quantity.

[0090] As described above, in the present embodiment, the estimating unit 10 b uses the third physical quantity estimated last time as the first physical quantity in estimating the third physical quantity.

[0091] According to such a configuration, the latest third physical quantity can be estimated using the third physical quantity estimated last time by the estimating unit 10 b .

[0092] The second physical quantity is a quantity (current value) related to the damping coefficient of the shock absorber 4b. The third physical quantity is a relative speed of the wheel 3 with respect to the vehicle body 2 in the vertical direction of the vehicle 1, that is, a stroke speed.

[0093] According to such a configuration, even when a plurality of specifications with different amounts of correlation with the damping coefficient of the shock absorber 4 b exist in the vehicle 1 , the time and effort required to prepare the neural network 121 can be reduced.

[0094] In addition, the damping coefficient of the damper 4b may be used instead of the current value which is a quantity related to the damping coefficient of the damper 4b.

[0095] <Third embodiment>

[0096] Fig. 22 It is a diagram showing a neural network 121 according to the third embodiment.

[0097] In this embodiment, if Fig. 22 As shown, the main difference from the second embodiment is that the input data input to the neural network 121 is the numerical value of the force obtained by multiplying the sprung acceleration by the mass of the vehicle 1, the damping force of the shock absorber 4b, and the spring force of the coil spring 4a. In addition, in this embodiment, the output data is the stroke speed and the relative displacement of the wheel 3 with respect to the vehicle body 2 in the vertical direction of the vehicle 1 (the so-called stroke displacement).

[0098] The spring force of the coil spring 4a is calculated based on the spring constant of the coil spring 4a and the stroke displacement amount estimated last time by the estimation unit 10b. Specifically, the spring force of the coil spring 4a is calculated by multiplying the spring constant of the coil spring 4a by the stroke displacement amount estimated last time by the estimation unit 10b. The estimation unit 10b estimates a new stroke displacement amount by multiplying the spring constant of the coil spring 4a by the stroke displacement amount estimated last time. The stroke displacement amount (third physical quantity) estimated last time by the estimation unit 10b is equivalent to the first physical quantity in the estimation of the stroke displacement amount performed by the estimation unit 10b this time. In addition, the spring constant of the coil spring 4a is an example of the second physical quantity.

[0099] As described above, in this embodiment, the second physical quantity is the spring constant of the coil spring 4 a (spring). The third physical quantity is the stroke displacement which is the relative displacement of the wheel 3 with respect to the vehicle body 2 in the vertical direction of the vehicle 1 .

[0100] According to such a configuration, even when the vehicle 1 has a plurality of coil springs 4 a having different spring constants, the time and effort required to prepare the neural network 121 can be reduced.

[0101] <Fourth embodiment>

[0102] Fig.23 It is a diagram showing a neural network 121 according to the fourth embodiment.

[0103] In this embodiment, the vehicle 1 has a front stabilizer (not shown) and a rear stabilizer (not shown). Fig.23 As shown, in this embodiment, the main difference from the second embodiment is that the input data input to the neural network 121 are the numerical value of the force obtained by multiplying the sprung acceleration by the mass of the vehicle 1, the front stabilizer spring force, and the rear stabilizer spring force. In addition, in this embodiment, the output data (third physical quantity) are the stroke speed, the FR stroke displacement amount which is the stroke displacement amount of the right front part of the vehicle 1, the FL stroke displacement amount which is the stroke displacement amount of the left front part of the vehicle 1, the RR stroke displacement amount which is the stroke displacement amount of the right rear part of the vehicle 1, and the RL stroke displacement amount which is the stroke displacement amount of the right front part of the vehicle 1.

[0104] The front stabilizer spring force is calculated by multiplying the difference between the FR stroke displacement and the FL stroke displacement estimated by the estimation unit 10b last time by the spring constant (Kf) of the front stabilizer. The rear stabilizer spring force is calculated by multiplying the difference between the RR stroke displacement and the RL stroke displacement estimated by the estimation unit 10b last time by the spring constant (Kr) of the rear stabilizer. The FL stroke displacement, FR stroke displacement, RR stroke displacement, and RR stroke displacement (third physical quantity) estimated by the estimation unit 10b last time are equivalent to the second physical quantity in the current estimation of the stroke displacement by the estimation unit 10b. The spring constant of the front stabilizer and the spring constant of the rear stabilizer are the second physical quantity.

[0105] As described above, in the present embodiment, the second physical quantity is the spring constant of the front stabilizer and the spring constant of the rear stabilizer.

[0106] According to such a configuration, even when the vehicle 1 has a plurality of specifications in which the spring constants of the front stabilizer and the spring constants of the rear stabilizer are different, the time and effort required to prepare the neural network 121 can be reduced.

[0107] In the above embodiment, the first physical quantity may be the acceleration of the vehicle body 2 in the front-rear direction, the acceleration of the vehicle body 2 in the left-right direction (lateral direction), the speed of the wheels, the yaw rate of the vehicle body 2, etc. In addition, the second physical quantity may be the height of the center of gravity of the vehicle 1. The pitch moment may be calculated by multiplying the acceleration of the vehicle body 2 in the front-rear direction by the height of the center of gravity of the vehicle 1, and the pitch moment may be used as input data of the neural network 121. In addition, the rolling moment may be calculated by multiplying the acceleration of the vehicle body 2 in the left-right direction by the height of the center of gravity of the vehicle 1, and the rolling moment may be used as input data of the neural network 121.

[0108] In the above embodiment, the mass of the vehicle 1 as the second physical quantity may also be a value including the occupants and loads of the vehicle 1. In this case, by providing a detection device for detecting the mass of the occupants and loads of the vehicle 1, the mass of the vehicle 1 including the occupants and loads can be calculated.

[0109] The embodiments are described in the present invention, but these are only provided as examples and are not intended to limit the scope of the invention. These new embodiments can be implemented in various other ways, and various omissions, substitutions, and changes can be made without departing from the scope of the invention. These embodiments and their variations are included in the scope and purpose of the invention, and are included in the invention described in the scope of the patent claim and its equivalent scope.

[0110] [Summary of this embodiment]

[0111] The control device 10 of the present embodiment has at least the following configuration.

[0112] That is, the control device 10 is a control device 10 for controlling the suspension device 4 interposed between the vehicle body 2 of the vehicle 1 and the wheel 3 of the vehicle 1, and includes:

[0113] an acquisition unit 10a that acquires the acceleration of the vehicle body 2 in the up-down direction of the vehicle 1; and

[0114] an estimating unit 10b that estimates a relative speed of the wheel 3 of the vehicle 1 in the vertical direction with respect to the vehicle body 2 using the neural network 121, the relative speed corresponding to a value obtained by multiplying the acceleration obtained by the acquiring unit 10a by a mass, the magnitude of the mass being the same as or different from that at the time of learning the neural network 121, the neural network being learned in a manner that estimates the relative speed in accordance with an input of a value obtained by multiplying the acceleration by the mass of the vehicle 1,

[0115] The suspension device 4 is controlled based on the estimated relative speed.

[0116] According to this configuration, for example, the same neural network 121 can be used for vehicles 1 of a plurality of specifications having different masses. Therefore, even when there are a plurality of specifications of vehicles 1, the time and effort required to create the neural network 121 can be reduced.

[0117] The control device 10 of the embodiment preferably includes a shock absorber 4b interposed between the vehicle body 2 and the wheel 3, and the estimating unit 10b estimates a new relative speed by multiplying the damping coefficient of the shock absorber 4b or a quantity related to the damping coefficient by the last estimated relative speed.

[0118] According to this configuration, even when the vehicle 1 has a plurality of specifications of the shock absorber 4 b having different damping coefficients or amounts related to the damping coefficient, the time required to prepare the neural network 121 can be reduced.

[0119] In addition, the control device 10 of the embodiment preferably includes a spring 4a, which is interposed between the vehicle body 2 and the wheel 3, and estimates the relative displacement of the wheel 3 in the vertical direction of the vehicle relative to the vehicle body 2 based on the relative speed. The estimation unit 10b estimates the new relative displacement by multiplying the spring constant of the spring 4a by the relative displacement estimated last time.

[0120] According to this configuration, even when the vehicle 1 has a plurality of springs 4 a having different spring constants, the time and effort required to prepare the neural network 121 can be reduced.

Claims

1. A control device for controlling a suspension device between a vehicle body and a wheel of the vehicle, wherein: have: an acquisition unit that acquires the acceleration of the vehicle body in the up-down direction of the vehicle; and an estimating unit that estimates a relative speed of the wheel with respect to the vehicle body in a vertical direction of the vehicle using a neural network, the relative speed corresponding to a value obtained by multiplying the acceleration obtained by the obtaining unit by a mass of the vehicle, the mass being the same or different in magnitude from that when the neural network was learned, and the neural network was learned in such a manner that the relative speed was estimated in response to an input of a value obtained by multiplying the acceleration by the mass, The suspension device is controlled based on the estimated relative speed.

2. The control device according to claim 1, wherein: A shock absorber is provided, the shock absorber being interposed between the vehicle body and the wheel, The estimating unit estimates the new relative speed by multiplying the damping coefficient of the shock absorber or a quantity related to the damping coefficient by the last estimated relative speed.

3. The control device according to claim 1, wherein: A spring is provided, the spring being interposed between the vehicle body and the wheel, estimating a relative displacement of the wheel with respect to the vehicle body in the vertical direction of the vehicle based on the relative speed; The estimating unit estimates a new relative displacement amount by multiplying the spring constant of the spring by the relative displacement amount estimated last time.

Images