Air pressure estimation apparatus and air pressure estimation program

Abstract

To improve the estimation accuracy of an air pressure of a wheel.SOLUTION: An air pressure estimation apparatus comprises an execution device and a storage device. The storage device stores a relationship regulation model that outputs an index value indicating an air pressure of a wheel of a vehicle when a plurality of types of input data is input. Here, the relationship regulation model is generated in advance by machine learning. The plurality of types of input data includes unsprung acceleration, which is acceleration in a vertical direction to the wheel, and position coordinates, which are coordinates of a point at which the vehicle is located. The execution device acquires the plurality of types of input data (S11). The execution device outputs the index value by inputting the acquired plurality of types of input data to the relationship regulation model (S13).SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present invention relates to a pneumatic pressure estimation device and a pneumatic pressure estimation program.

Background Art

[0002] The vehicle of Patent Document 1 includes a pneumatic pressure estimation device and an acceleration sensor. The acceleration sensor is attached to a hub that supports a wheel. The acceleration sensor detects the under-spring acceleration, which is the acceleration in the vertical direction with respect to the wheel. The pneumatic pressure estimation device acquires the under-spring acceleration detected by the acceleration sensor. Subsequently, the pneumatic pressure estimation device specifies the magnitude of the amplitude for each frequency of the vibration in the vertical direction with respect to the wheel based on the acquired under-spring acceleration. Then, the pneumatic pressure estimation device estimates the pneumatic pressure of the wheel based on the magnitude of the amplitude for each specified frequency.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] When a vehicle such as that in Patent Document 1 is running, for example, the wheel vibrates in the vertical direction due to the unevenness of the road surface at the location where the vehicle is located. As a result, even if the pneumatic pressure of the wheel is the same, the under-spring acceleration changes. Therefore, in a pneumatic pressure estimation device such as that in Patent Document 1, even if the pneumatic pressure of the wheel is estimated based only on the under-spring acceleration, there is a possibility that the estimation accuracy is low.

Means for Solving the Problems

[0005] The air pressure estimation device for solving the above problems includes an execution device and a storage device. The storage device stores a relationship defining model that outputs an index value indicating the air pressure of the vehicle's wheels when a plurality of types of input data are input. The plurality of types of input data include the unsprung acceleration, which is the vertical acceleration with respect to the wheel, and the position coordinates, which are the coordinates of the location where the vehicle is located. The execution device executes acquiring the plurality of types of input data and outputting the index value by inputting the acquired plurality of types of input data into the relationship defining model.

[0006] The air pressure estimation program for solving the above problems is applied to an air pressure estimation device that includes an execution device and a storage device and estimates the air pressure of the vehicle's wheels. The storage device stores a relationship defining model that outputs an index value indicating the air pressure when a plurality of types of input data are input. The plurality of types of input data include the unsprung acceleration, which is the vertical acceleration with respect to the wheel, and the position coordinates, which are the coordinates of the location where the vehicle is located. The execution device is caused to execute acquiring the plurality of types of input data and outputting the index value by inputting the acquired plurality of types of input data into the relationship defining model.

Advantages of the Invention

[0007] According to the above configuration, an index value indicating the air pressure of the wheel based on the unsprung acceleration and the vehicle's position coordinates is output from the relationship defining model. Here, the relationship defining model defines the relationship between the unsprung acceleration and the vehicle's position coordinates and the air pressure of the wheel. Therefore, the index value indicating the air pressure of the wheel output from the relationship defining model is a value that takes into account not only the unsprung acceleration but also the vehicle's position coordinates, in other words, the influence of the road surface at the location where the vehicle is located. As a result, the estimation accuracy of the air pressure can be improved compared to a configuration that does not take into account the influence of the road surface at the location where the vehicle is located.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Mode for Carrying Out the Invention

[0009] <Schematic Configuration of the Estimation System> Hereinafter, an embodiment of the present invention will be described with reference to FIGS. 1 and 2. First, the schematic configuration of the vehicle 100 will be described.

[0010] As shown in FIG. 1, the vehicle 100 includes an internal combustion engine 10, a torque converter 20, an automatic transmission 30, a differential 41, a plurality of drive wheels 42, and a hydraulic mechanism 50. The internal combustion engine 10 includes four cylinders 11 and a crankshaft 12. The cylinder 11 is a space for burning a mixture of fuel and intake air. The crankshaft 12 rotates due to the combustion of the mixture in the cylinder 11.

[0011] The torque converter 20 includes an input shaft 21 and an output shaft 22. The torque converter 20 transmits the driving force of the input shaft 21 to the output shaft 22 via a fluid. At this time, the torque converter 20 decelerates the rotation of the input shaft 21 and outputs it from the output shaft 22. The first end of the input shaft 21 is connected to the crankshaft 12. Further, the torque converter 20 includes a lock-up clutch (not shown). The second end of the input shaft 21 is connected to the first end of the output shaft 22 via the lock-up clutch. In a state where the lock-up clutch is engaged, the input shaft 21 and the output shaft 22 rotate integrally.

[0012] The automatic transmission 30 includes an input shaft 31 and an output shaft 32. The first end of the input shaft 31 is connected to the second end of the output shaft 22 in the torque converter 20. The second end of the input shaft 31 is connected to the first end of the output shaft 32 via a clutch and gears (not shown). The second end of the output shaft 32 is connected to the left and right drive wheels 42 via a differential 41. The automatic transmission 30 can change the gear ratio, which is the ratio of the rotational speed of the input shaft 31 to the rotational speed of the output shaft 32. Here, the gear ratio of the automatic transmission 30 is the ratio indicating the number of rotations of the input shaft 31 when the output shaft 32 makes one rotation. Therefore, the higher the gear ratio, the higher the rotational speed of the input shaft 31 with respect to the output shaft 32. An example of the automatic transmission 30 is a stepped automatic transmission. Therefore, the automatic transmission 30 changes the gear ratio by changing the gear position.

[0013] The hydraulic mechanism 50 is attached to the automatic transmission 30. The hydraulic mechanism 50 supplies oil to the automatic transmission 30. And the automatic transmission 30 is controlled by the oil supplied from the hydraulic mechanism 50.

[0014] As shown in FIG. 1, the vehicle 100 includes an accelerator operation amount sensor 71, a vehicle speed sensor 72, and a GNSS receiver 73. The vehicle 100 also includes a sprung acceleration sensor 74, a plurality of unsprung acceleration sensors 75, a payload sensor 76, and a display 79.

[0015] The accelerator operation amount sensor 71 detects the accelerator operation amount ACC, which is the operation amount of the accelerator pedal operated by the driver of the vehicle 100. The vehicle speed sensor 72 detects the vehicle speed SP, which is the speed of the vehicle 100.

[0016] The GNSS receiver 73 detects the position coordinates PC, which are the coordinates of the location where the vehicle 100 is located, by communicating with GNSS satellites (not shown). Note that "GNSS" is an abbreviation for Global Navigation Satellite System.

[0017] The sprung acceleration sensor 74 is attached to the vehicle body of the vehicle 100 that is located vertically above the suspension device. In other words, the sprung acceleration sensor 74 is attached to the vehicle body of the vehicle 100 that is connected to the wheels of the vehicle 100 via the suspension device. In the present embodiment, the sprung acceleration sensor 74 is located near the driver's seat of the vehicle 100. Further, the sprung acceleration sensor 74 is a so-called triaxial sensor. That is, the sprung acceleration sensor 74 can detect the longitudinal acceleration GX, the lateral acceleration GY, and the vertical acceleration GZ. The longitudinal acceleration GX is the acceleration along the longitudinal axis of the vehicle 100. The lateral acceleration GY is the acceleration along the lateral axis of the vehicle 100. The vertical acceleration GZ is the acceleration along the vertical axis of the vehicle 100. Therefore, the vertical acceleration GZ detected by the sprung acceleration sensor 74 corresponds to the sprung acceleration GZA that is the vertical acceleration of the vehicle body of the vehicle 100.

[0018] The unsprung acceleration sensor 75 is attached near the wheels of the vehicle 100. As a specific example, the unsprung acceleration sensor 75 is attached to the hub that supports the wheels of the vehicle 100. The unsprung acceleration sensor 75 detects the unsprung acceleration GZB that is the acceleration along the vertical axis of the vehicle 100. Therefore, the unsprung acceleration GZB detected by the unsprung acceleration sensor 75 is the vertical acceleration with respect to the wheels of the vehicle 100. In the present embodiment, the vehicle 100 is provided with a total of four unsprung acceleration sensors 75 corresponding to the total of four wheels provided in the vehicle 100. In FIG. 1, one unsprung acceleration sensor 75 is illustrated representatively.

[0019] The payload sensor 76 detects the payload L of the vehicle 100. In the present embodiment, the payload L is the weight of the goods and passengers loaded on the vehicle 100. The display 79 is located near the driver's seat of the vehicle 100. The display 79 can display various information.

[0020] As shown in FIG. 1, the vehicle 100 includes a control device 90. The control device 90 acquires various information from an accelerator operation amount sensor 71, a vehicle speed sensor 72, and a GNSS receiver 73. Further, the control device 90 acquires various information from a sprung acceleration sensor 74, a plurality of unsprung acceleration sensors 75, and a payload sensor 76.

[0021] The control device 90 includes an execution device 91 and a storage device 92. An example of the execution device 91 is a CPU. The storage device 92 includes a read-only ROM, a volatile RAM that can be read and written, and a non-volatile storage that can be read and written. The storage device 92 stores various programs and various data in advance. Specifically, the storage device 92 stores a control program 92A in advance as one of various programs. Further, the storage device 92 stores a relationship specification model M in advance as one of various data. The relationship specification model M describes the relationship between predetermined input data and an index value indicating the air pressure of the wheels of the vehicle 100 in a form recognizable by the execution device 91. The relationship specification model M outputs an index value indicating the air pressure of the wheels of the vehicle 100 when a plurality of types of input data are input. In the present embodiment, the relationship specification model M is generated in advance by machine learning. A specific description of the relationship specification model M will be described later. The execution device 91 executes various processes described later by executing the control program 92A stored in the storage device 92. In the present embodiment, the control device 90 is an example of an air pressure estimation device. Further, the control program 92A is an example of an air pressure estimation program.

[0022] The execution device 91 of the control device 90 calculates a target driving force, which is a target value of the driving force of the vehicle 100, based on the accelerator operation amount ACC and the vehicle speed SP. Subsequently, the execution device 91 calculates a target output, which is a target value of the output of the internal combustion engine 10, based on the target driving force. Then, the execution device 91 outputs a control signal corresponding to the target output to the internal combustion engine 10. As a result, the internal combustion engine 10 is controlled according to the target output. Also, the execution device 91 calculates a target gear position, which is a target value of the gear position of the automatic transmission 30, based on the target driving force. Then, the execution device 91 outputs a control signal corresponding to the target gear position to the hydraulic mechanism 50. As a result, the automatic transmission 30 is controlled by controlling the hydraulic mechanism 50.

[0023] <Estimation control> Next, with reference to FIG. 2, the estimation control executed by the control device 90 will be described. This estimation control is control for estimating the air pressure of the wheels of the vehicle 100. In the present embodiment, the execution device 91 of the control device 90 starts the estimation control at each predetermined control cycle on the condition that the vehicle 100 is running. Note that the execution device 91 executes the estimation control for each wheel corresponding to a total of four wheels provided in the vehicle 100.

[0024] As shown in FIG. 2, when the execution device 91 of the control device 90 starts the estimation control, it executes the process of step S11. In step S11, the execution device 91 acquires the time-series data of the position coordinates PC, the time-series data of the acceleration GZA above the spring, and the time-series data of the acceleration GZB below the spring. In the present embodiment, the above time-series data includes the data at each time point from the time point of the process in step S11 to a predetermined regular period before. Here, it is assumed that there are N time points at regular intervals as the above plurality of time points. Also, among the N time points, in order from the older ones, they are the first time point, the second time point, ···, the Nth time point. Note that "N" is an integer of 2 or more. Therefore, in step S11, the execution device 91 acquires N position coordinates PC from the first time point to the Nth time point as the time-series data of the position coordinates PC. Also, the execution device 91 acquires N accelerations GZA above the spring from the first time point to the Nth time point as the time-series data of the acceleration GZA above the spring. As described above, the acceleration GZA above the spring is the vertical acceleration GZ detected by the acceleration sensor 74 above the spring. Further, the execution device 91 acquires N accelerations GZB below the spring from the first time point to the Nth time point as the time-series data of the acceleration GZB below the spring. Here, the acceleration GZB below the spring is the value detected by the acceleration sensor 75 below the spring corresponding to the wheel targeted in the current estimation control.

[0025] Also, in step S11, the execution device 91 acquires the load weight L at the time point of the process in step S11. Then, the execution device 91 calculates the total weight WV of the vehicle 100 based on the load weight L. For example, the execution device 91 calculates the sum of the load weight L and the initial value of the weight of the vehicle 100 as the total weight WV. After step S11, the execution device 91 proceeds with the process to step S12.

[0026] In step S12, the execution device 91 generates the time-series data of the position coordinates PC, the time-series data of the acceleration GZA above the spring, the time-series data of the acceleration GZB below the spring, and the total weight WV obtained in step S11 as input variables of the relationship definition model M. Specifically, the execution device 91 sequentially substitutes the values of the acceleration GZB below the spring from the first time point to the Nth time point into the input variables x(1) to x(N) one by one. Similarly to the above, the execution device 91 sequentially substitutes the values of the position coordinates PC from the first time point to the Nth time point into the input variables x(N + 1) to x(2×N) one by one. At this time, the execution device 91 substitutes the values after converting the position coordinates PC into one-dimensional values, for example. Further, the execution device 91 sequentially substitutes the values of the acceleration GZA above the spring from the first time point to the Nth time point into the input variables x(2×N + 1) to x(3×N) one by one. Furthermore, the execution device 91 substitutes the value of the total weight WV into the input variable x(3×N + 1). Hereinafter, "3×N + 1" is described as "Z". That is, "Z" is the number of input variables generated in step S12.

[0027] In the present embodiment, each of the time-series data of the position coordinates PC, the time-series data of the acceleration GZA above the spring, the time-series data of the acceleration GZB below the spring, and the total weight WV is input data input to the relationship definition model M. After step S12, the execution device 91 proceeds with the process to step S13.

[0028] In step S13, the execution device 91 inputs the input variables x(1) to x(Z) and the input variable x(0) as a bias parameter to the relationship definition model M, thereby outputting the value of the output variable y(i) indicating the tire pressure of the wheel. Here, the output variable y(i) is an index value indicating the tire pressure of the wheel.

[0029] An example of the relationship specification model M is a function approximator, which is a fully connected feedforward neural network with one hidden layer. In this relationship specification model M, each of the "m" values obtained by linearly mapping the input variables x(1) to x(Z) and the input variable x(0) as a bias parameter by the coefficients wFjk (j = 1 to m, k = 0 to Z) is substituted into the activation function f. As a result, the values of the nodes in the hidden layer are determined. Further, each of the values obtained by linearly mapping the values of the nodes in the hidden layer by the linear mapping defined by the coefficient wSij (i = 1) is substituted into the activation function g, thereby determining the output variable y(1). In the present embodiment, an example of the activation function f is the ReLU function. Also, an example of the activation function g is the sigmoid function. That is, the output variable y(1) can vary in the range from "0" to "1". Note that the smaller the output variable y(1), the lower the air pressure of the wheel.

[0030] The relationship specification model M is generated in advance, for example, as follows. First, engineers and the like drive the vehicle 100 under various conditions at various locations. At this time, in the same manner as above, time-series data of the position coordinate PC, time-series data of the acceleration GZA above the spring, time-series data of the acceleration GZB below the spring, and the total weight WV are acquired. Then, in the same manner as above, the input variables x(1) to x(Z) are generated. Also, the air pressure of the wheels of the vehicle 100 is actually measured to grasp the air pressure of the wheels. Then, the output variable y(1) corresponding to the grasped air pressure of the wheels is generated. The relationship specification model M is generated by machine learning using the data generated as described above. That is, the relationship specification model M is generated in advance by machine learning using, as teacher data, the combination of the time-series data of the position coordinate PC, the time-series data of the acceleration GZA above the spring, the time-series data of the acceleration GZB below the spring, and the total weight WV, and the air pressure of the wheels. After step S13, the execution device 91 proceeds to step S21.

[0031] In step S21, the execution device 91 determines whether the output variable y(1) is equal to or greater than a predetermined specified value A. Here, the specified value A indicates the lower limit value allowable as the air pressure of the wheels of the vehicle 100. If, in step S21, the execution device 91 determines that the output variable y(1) is equal to or greater than the specified value A (S21: YES), the execution device 91 proceeds with the process to step S31.

[0032] In step S31, the execution device 91 determines that the air pressure of the wheels of the vehicle 100 is normal. After step S31, the execution device 91 ends the current estimation control. On the other hand, if, in step S21, the execution device 91 determines that the output variable y(1) is less than the specified value A (S21: NO), the execution device 91 proceeds with the process to step S32.

[0033] In step S32, the execution device 91 determines that the air pressure of the wheels of the vehicle 100 is abnormal. Then, the execution device 91 notifies the driver of the vehicle 100 or the like that the air pressure of the wheels is low. As a specific example, the execution device 91 outputs a control signal to the display 79, and causes the display 79 to display that the air pressure of the wheels is low. After that, the execution device 91 ends the current estimation control.

[0034] <Actions of the present embodiment> When the vehicle 100 is running, for example, the wheels vibrate in the vertical direction due to the unevenness of the road surface at the location where the vehicle 100 is located. As a result, even if the air pressure of the wheels is the same, the unsprung acceleration GZB can change.

[0035] As shown in FIG. 2, in the estimation control, the execution device 91 of the control device 90 acquires, as a plurality of types of input data, time-series data of the position coordinates PC, time-series data of the sprung acceleration GZA, time-series data of the unsprung acceleration GZB, and the total weight WV. Then, the execution device 91 inputs the plurality of types of input data into the relationship defining model M, and outputs an output variable y(1) indicating an index value of the air pressure of the wheels.

[0036] <Effects of this Embodiment> (1) According to this embodiment, an index value indicating the tire pressure of a wheel is output from a relationship defining model M based on the under-spring acceleration GZB and the position coordinate PC. Here, the relationship defining model M defines the relationship between the under-spring acceleration GZB and the position coordinate PC and the tire pressure of the wheel. Therefore, the index value indicating the tire pressure of the wheel output from the relationship defining model M takes into account not only the under-spring acceleration GZB but also the position coordinate PC, in other words, a value taking into account the vertical vibration of the wheel due to the influence of the road surface at the point where the vehicle 100 is located. Thereby, for example, the estimation accuracy of the tire pressure of the wheel can be improved as compared with a configuration that does not take into account the vertical vibration of the wheel due to the influence of the road surface at the point where the vehicle 100 is located.

[0037] (2) In the vehicle 100, for example, the under-spring acceleration GZB at a certain point in time can change due to the influence of the under-spring acceleration GZB before that point in time. Also, for example, the under-spring acceleration GZB at a certain point in time can change not only due to the influence of the road surface at the point where the vehicle 100 is located at that time but also due to the influence of the road surface at the point where the vehicle 100 was located at a time before the above time.

[0038] In this regard, according to this embodiment, an index value indicating the tire pressure of the wheel is output from the relationship defining model M based on the time-series data of the under-spring acceleration GZB and the time-series data of the position coordinate PC. Thereby, by taking into account the temporal changes in the under-spring acceleration GZB and the position coordinate PC, the estimation accuracy of the tire pressure of the wheel can be more reliably improved.

[0039] (3) In the vehicle 100, the under-spring acceleration GZB may change when the vertical vibration of the vehicle body is transmitted to the wheel via the suspension device. According to this embodiment, an index value indicating the tire pressure of the wheel is output from the relationship defining model M based on the time-series data of the above-spring acceleration GZA. Thereby, by taking into account the above-spring acceleration GZA, in other words, by taking into account the vertical vibration of the vehicle body, the estimation accuracy of the tire pressure of the wheel can be improved.

[0040] (4) In the vehicle 100, the way the variation of the acceleration GZB under the spring changes according to the total weight WV. According to the present embodiment, based on the total weight WV of the vehicle 100, an index value indicating the air pressure of the wheel is output from the relationship defining model M. Thereby, for example, the estimation accuracy of the air pressure of the wheel can be improved as compared with a configuration that does not take into account the total weight WV of the vehicle 100.

[0041] <Modification example> The present embodiment can be implemented with the following modifications. The present embodiment and the following modification examples can be implemented in combination with each other within a technically non - conflicting range.

[0042] · In the above - described embodiment, the input data of the relationship defining model M may be changed. For example, the plurality of types of input data input to the relationship defining model M may not include the total weight WV. As a specific example, if the change in the acceleration GZB under the spring according to the total weight WV is small, the input data of the relationship defining model M may not include the total weight WV.

[0043] · For example, the plurality of types of input data input to the relationship defining model M may not include the acceleration GZA above the spring. As a specific example, if the change in the acceleration GZB under the spring due to the vertical vibration of the vehicle body is small, the input data of the relationship defining model M may not include the acceleration GZA above the spring.

[0044] · For example, the acceleration GZB under the spring as the input data input to the relationship defining model M may not be time - series data. As a specific example, if the acceleration GZB at a certain point in time is unlikely to change due to the influence of the acceleration GZB under the spring before that point in time, the acceleration GZB under the spring may not be time - series data.

[0045] ·For example, the position coordinates PC as input data to the relationship defining model M do not have to be time series data. As a specific example, if the spring-under acceleration GZB at a certain point in time is unlikely to change due to the influence of the road surface at the point where the vehicle 100 is located at a time point earlier than that time point, the position coordinates PC do not have to be time series data.

[0046] ·In the above embodiment, the relationship defining model M may be changed. For example, the activation function of the relationship defining model M is an example, and the activation function of the relationship defining model M can be changed.

[0047] ·For example, as the relationship defining model M, a neural network with one hidden layer was exemplified, but the number of hidden layers may be two or more. ·For example, as the neural network of the relationship defining model M, a fully connected feedforward neural network was exemplified, but it is not limited to this. As a specific example, the neural network may be a recurrent connection type neural network. Also, for example, the function approximator as the relationship defining model M is not limited to a neural network. As a specific example, the relationship defining model M may be a regression equation without a hidden layer.

[0048] ·For example, the relationship defining model M does not have to be generated by machine learning. As a specific example, the relationship defining model M may be a relational expression determined by experiments and simulations, etc.

[0049] ·In the above embodiment, the air pressure estimation device may be changed. For example, the air pressure estimation device is not limited to the control device 90 of the vehicle 100, and may be a device different from the control device 90 in the vehicle 100. In this case, it is sufficient if the above-mentioned different device stores the relationship regulation model M and the air pressure estimation program in advance. Also, for example, the air pressure estimation device may be a device external to the vehicle 100. Here, an example of a device external to the vehicle 100 is a server capable of communicating with the vehicle 100. In this configuration, the server can execute estimation control by acquiring various information from the vehicle 100. In this case, it is sufficient if the server stores the relationship regulation model M and the air pressure estimation program in advance.

Explanation of Signs

[0050] 10…Internal combustion engine 11…Cylinder 12…Crankshaft 20…Torque converter 21…Input shaft 22…Output shaft 30…Automatic transmission 31…Input shaft 32…Output shaft 41…Differential 42…Drive wheel 50…Hydraulic mechanism 71…Accelerator operation amount sensor 72…Vehicle speed sensor 73…GNSS receiver 74…Sprung acceleration sensor 75…Unsprung acceleration sensor 76…Payload sensor 79…Display 90…Control device 91…Execution device 92…Storage device 92A…Control program M…Relationship regulation model 100…Vehicle

Claims

1. An execution device and a storage device are provided, The storage device stores a relationship defining model that outputs an index value indicating the air pressure of the vehicle wheels when a plurality of types of input data are input, The plurality of types of input data include the unsprung acceleration, which is the vertical acceleration with respect to the wheel, and the position coordinates, which are the coordinates of the location where the vehicle is located, The execution device, obtains the plurality of types of input data, outputs the index value by inputting the obtained plurality of types of input data into the relationship defining model, and executes an air pressure estimation device.

2. The plurality of types of input data include time-series data of the unsprung acceleration and time-series data of the position coordinates, The air pressure estimation device according to Claim 1.

3. The plurality of types of input data include time-series data of the sprung acceleration, which is the vertical acceleration with respect to the vehicle body connected to the wheel via a suspension device, The air pressure estimation device according to Claim 1 or Claim 2.

4. The plurality of types of input data include the total weight of the vehicle, The air pressure estimation device according to Claim 1 or Claim 2.

5. It is applied to an air pressure estimation device that includes an execution device and a storage device and estimates the air pressure of the vehicle wheels. The storage device stores a relationship defining model that outputs an index value indicating the air pressure when a plurality of types of input data are input, The plurality of types of input data include the unsprung acceleration, which is the vertical acceleration with respect to the wheel, and the position coordinates, which are the coordinates of the location where the vehicle is located, In the execution device, obtain the plurality of types of input data, Outputting the index value by inputting the obtained plurality of types of input data into the relationship definition model; causing to execute a pneumatic pressure estimation program.

Images