Tire physical information estimation system and tire physical information estimation method

Abstract

To provide a tire physical information estimation system and a tire physical information estimation method which can estimate physical information concerning a tire while simply constructing a calculation model.SOLUTION: A tire physical information estimation system 100 comprises a physical information estimation part 32 and a correction processing part 33. The physical information estimation part 32 has a learning-type calculation model 32a extending from an input layer to an output layer, which estimates tire physical information that is generated by movements of a tire 10. The correction processing part 33 corrects the tire physical information estimated by the physical information estimation part 32. The calculation model 32a has a feature extraction part that executes convolution calculation in calculation in the course from the input layer to the output layer, which outputs tire physical information in at least two axial directions normalized from the output layer. The correction processing part 33 corrects the tire physical information estimated by the physical information estimation part 32, on the basis of a parameter set depending on a vehicle.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present invention relates to a tire physical information estimation system and a tire physical information estimation method.

Background Art

[0002] Recently, research has been conducted on a system that inputs information measured in tires, vehicles, etc. into a learning-based arithmetic model and estimates tire physical information such as tire forces.

[0003] Patent Document 1 describes a conventional tire physical information estimation system. The tire physical information estimation system includes a physical information estimation unit and a data acquisition unit. The physical information estimation unit has a learning-based arithmetic model from an input layer to an output layer for estimating physical information related to a tire generated by the movement of the tire. The data acquisition unit acquires input data to the input layer. The arithmetic model has a feature extraction unit that executes a convolution operation in an intermediate operation from the input layer to the output layer to extract a feature amount.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] In the tire physical information estimation system described in Patent Document 1, real-time performance is ensured in the estimation of tire physical information, but there is a problem that it takes time and cost to build a learning model for each of a plurality of vehicle types and models, for example.

[0006] The present invention has been made in view of such circumstances, and an object thereof is to provide a tire physical information estimation system and a tire physical information estimation method capable of simply constructing a calculation model and estimating physical information related to a tire.

Means for Solving the Problems

[0007] One aspect of the present invention is a tire physical information estimation system. It has a learning-based calculation model from an input layer to an output layer, a physical information estimation unit that estimates tire physical information generated by the movement of the tire, and a correction processing unit that corrects the tire physical information estimated by the physical information estimation unit. The calculation model has a feature extraction unit that executes a convolution operation in an intermediate calculation from the input layer to the output layer, outputs at least biaxial tire physical information normalized from the output layer, and the correction processing unit corrects the tire physical information estimated by the physical information estimation unit based on parameters set according to the vehicle.

Effects of the Invention

[0008] According to the present invention, a calculation model can be simply constructed to estimate physical information related to a tire.

Brief Description of the Drawings

[0009]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Mode for Carrying Out the Invention

[0010] Hereinafter, the present invention will be described based on preferred embodiments with reference to FIGS. 1 to 11. The same or equivalent components and members shown in each drawing are denoted by the same reference numerals, and repeated explanations are appropriately omitted. In addition, the dimensions of the members in each drawing are appropriately enlarged or reduced for easy understanding. Also, some of the members that are not important for explaining the embodiments in each drawing are omitted from the display.

[0011] (Embodiment) FIG. 1 is a schematic diagram for explaining the outline of a tire physical information estimation system 100 according to an embodiment. The tire physical information estimation system 100 includes a sensor 20 disposed on a tire 10 and a tire physical information estimation device 30. Further, the tire physical information estimation system 100 may include a server device 40 or the like that acquires and accumulates tire physical information such as the tire force F estimated by the tire physical information estimation device 30, the road surface friction coefficient, and the moments about three axes acting on the tire 10 via a communication network 91, and monitors the tire physical information.

[0012] The sensor 20 measures physical quantities of the tire 10 such as acceleration, strain, tire air pressure, and tire temperature in the tire 10, and outputs the measured data to the tire physical information estimation device 30. The tire physical information estimation device 30 estimates the tire physical information based on the data measured by the sensor 20, and corrects it based on parameters set according to the vehicle. The tire physical information estimation device 30 uses the data measured by the sensor 20 in the calculation for estimating the tire physical information, but may also acquire information from the vehicle side such as vehicle acceleration from the vehicle control device 90 and use it in the calculation for estimating the tire physical information.

[0013] The tire physical information estimation device 30 outputs tire physical information such as the estimated tire force F, road surface friction coefficient, and moments around three axes acting on the tire 10 to, for example, the vehicle control device 90. The vehicle control device 90 uses the tire physical information input from the tire physical information estimation device 30 for, for example, estimating the braking distance, applying it to vehicle control, and further notifying the driver of information regarding the safe driving of the vehicle. The vehicle control device 90 can also provide information regarding the safe driving of the vehicle in the future using map information, weather information, and the like. Further, when the vehicle control device 90 has a function of automatically driving the vehicle, the tire physical information estimation system 100 provides the estimated tire physical information to the vehicle control device 90 as data used for vehicle speed control and the like in automatic driving.

[0014] FIG. 2 is a block diagram showing the functional configuration of the tire physical information estimation system 100 according to the embodiment. The sensor 20 of the tire physical information estimation system 100 includes an acceleration sensor 21, a strain gauge 22, a pressure gauge 23, a temperature sensor 24, etc., and measures physical quantities in the tire 10. These sensors measure physical quantities related to the deformation and movement of the tire 10 as physical quantities of the tire 10.

[0015] The acceleration sensor 21 and the strain gauge 22 mechanically move together with the tire 10 while measuring the acceleration and the amount of strain generated in the tire 10, respectively. The acceleration sensor 21 is disposed, for example, on the tread, side, bead, and wheel of the tire 10, and measures the acceleration in three axes in the circumferential direction, axial direction, and radial direction of the tire 10.

[0016] The strain gauge 22 is disposed on the tread, side, bead, etc. of the tire 10, and measures the strain at the disposed location. Further, the pressure gauge 23 and the temperature sensor 24 are disposed, for example, on the air valve of the tire 10, and measure the tire air pressure and the tire temperature, respectively. The temperature sensor 24 may be disposed directly on the tire 10 in order to accurately measure the temperature of the tire 10. The tire 10 may be attached with, for example, an RFID 11 to which unique identification information is given in order to identify each tire.

[0017] The tire physical information estimation device 30 includes a data acquisition unit 31, a physical information estimation unit 32, a correction processing unit 33, and a communication unit 34. The tire physical information estimation device 30 is an information processing device such as a PC (personal computer), for example. Each unit in the tire physical information estimation device 30 can be realized, in terms of hardware, by electronic elements such as a computer CPU and mechanical parts, and in terms of software, by a computer program or the like. Here, however, functional blocks realized by their cooperation are depicted. Therefore, it is understood by those skilled in the art that these functional blocks can be realized in various forms by a combination of hardware and software.

[0018] The data acquisition unit 31 acquires information on acceleration, strain, air pressure, and temperature measured by the sensor 20 through wireless communication or the like. The communication unit 34 communicates with external devices such as the vehicle control device 90 and the server device 40 through wired or wireless communication or the like. The communication unit 34 transmits the physical quantity of the tire 10 measured by the sensor 20 and the tire physical information estimated for the tire 10 to an external device via a communication line, such as CAN (Controller Area Network), the Internet, or the like.

[0019] The physical information estimation unit 32 has an arithmetic model 32a, inputs the information from the data acquisition unit 31 into the arithmetic model 32a, and estimates tire physical information such as the tire force F, the road surface friction coefficient, and the moments about the three axes acting on the tire 10. As shown in FIG. 2, the tire force F has three-axis direction components of the longitudinal force Fx in the longitudinal direction, the lateral force Fy in the lateral direction, and the vertical load Fz in the vertical direction of the tire 10. The physical information estimation unit 32 may calculate all of these three-axis direction components, or may perform the calculation of at least any one component or the calculation of two components by an arbitrary combination.

[0020] The arithmetic model 32a uses a learning model such as a neural network. FIG. 3 is a schematic diagram showing the configuration of the arithmetic model 32a. The arithmetic model 32a is of the CNN (Convolutional Neural Network) type and is a learning model having convolutional operations and pooling operations used in a so-called LeNet which is its prototype. FIG. 3 shows an example in which acceleration data in three-axis directions is used as input data to the arithmetic model 32a and the normalized tire forces in the three-axis directions are output.

[0021] The arithmetic model 32a includes an input layer 50, a feature extraction unit 51, an intermediate layer 52, a fully connected unit 53, and an output layer 54. The arithmetic model 32a is a multi-task type in which the feature extraction unit 51 is shared in the three-axis directions and fully connected units 53 are provided for each of the three-axis directions, and outputs the normalized values of the tire forces Fx, Fy, and Fz in the three-axis directions from the output layer 54. The normalized tire forces in the three-axis directions output from the output layer 54 are denoted as Fxn, Fyn, and Fzn.

[0022] For example, when the tire forces Fx and Fy generated during vehicle travel are -5000 N or more and 5000 N or less, the normalized tire forces Fxn and Fyn output from the output layer 54 are the values obtained by dividing the tire forces Fx and Fy by a constant value of 5000. Thereby, the normalized tire forces Fxn and Fyn vary within the range of -1 or more and 1 or less.

[0023] For example, when the tire force Fz generated during vehicle travel is 2000 N or more and 8000 N or less, the normalized tire force Fzn output from the output layer 54 is the value obtained by subtracting the tire force Fz from the median value of 5000 and then dividing by a constant value of 3000. Thereby, the normalized tire force Fzn varies within the range of -1 or more and 1 or less.

[0024] The normalization of the tire physical information such as the tire force F in the three-axis directions, the road surface friction coefficient, and the moments about the three axes acting on the tire 10 is not limited to the above examples, and can be appropriately set according to the nature of the tire physical information, the range of possible values, and the like.

[0025] Time-series data of the accelerations in the three-axis directions acquired by the data acquisition unit 31 are input to the input layer 50. The acceleration data is measured in a time series by the sensor 20, and data within a certain time interval is cut out by a window function to serve as input data. The input data is, for example, 224 acceleration data included in a certain time interval in each axis direction.

[0026] The acceleration measured by the tire 10 is periodic for each rotation of the tire 10. The time interval of the input data cut out by the window function is, for example, a time corresponding to the rotation period of the tire 10, and it is preferable to give the input data itself periodicity. Note that the window function may cut out the input data in a time interval shorter or longer than one rotation of the tire 10, and as long as the cut-out input data includes periodic information, the learning of the arithmetic model 32a is possible.

[0027] The feature extraction unit 51 extracts feature amounts using a convolution operation 51a and a pooling operation 51b, and transmits them to each node of the intermediate layer 52. In the example of the feature extraction unit 51 shown in FIG. 3, the first convolution operation is executed on the input data using 20 filters. The convolution operation 51a executes a convolution operation while moving a filter with respect to time-series input data such as acceleration data. The convolution operation 51a is executed for each of the three-axis acceleration data which are a plurality of input data, but the operation scale can be reduced by sharing the same filter for each axis direction.

[0028] In the convolution operation 51a, the filter length is set to 5, but it may be appropriately set to a size of about 1 to 5. Note that in the convolution operation, for data of a continuous filter length in time-series input data (for example, A1, A2, A3), each value (f1, f2, f3) in the filter is multiplied respectively, and the values obtained by multiplication are added, resulting in A1×f1 + A2×f2 + A3×f3. Note that zero-padding may be performed by adding "0 (zero)" data to the ends of the input data to execute the convolution operation. Also, the moving amount of the filter in the convolution operation is usually set to 1 input data, but it can be appropriately changed in order to reduce the operation model 32a.

[0029] The pooling operation 51b executes the first maximum value pooling operation on the data after the first convolution operation. The pooling operation 51b selects, for example, the larger value out of two values arranged in time series.

[0030] The second convolution operation 51c executes a convolution operation on the data after the pooling operation 51b using, for example, 50 filters. The filter length in the convolution operation 51c may be the same as that in the convolution operation 51a or may be different. Also for the convolution operation 51c, the operation scale can be reduced by sharing the same filter for each axis direction.

[0031] The pooling operation 51d performs a second max pooling operation on the data after the convolution operation 51c. Similar to the pooling operation 51b, the pooling operation 51d selects the larger value out of, for example, two values arranged in a time series. The feature extraction unit 51 obtains, as a calculation result, 64 data, that is, 64×3ch data in each axial direction by the convolution operation and the pooling operation, and outputs it to each node of the intermediate layer 52.

[0032] The fully connected part 53 fully connects the data from each node of the intermediate layer 52 in two layers for each of the three axial directions, and outputs the normalized tire forces Fxn, Fyn, and Fzn to each node of the output layer 54. The fully connected part 53 performs an operation by a fully connected path that executes a linear operation using weighting, etc. However, in addition to the linear operation, a non-linear operation may be performed using an activation function or the like.

[0033] The physical information estimation unit 32 can restore and estimate the tire forces Fx, Fy, and Fz by performing the inverse operation of normalization on the normalized tire forces Fxn, Fyn, and Fzn output to the output layer 54.

[0034] In addition to the normalized tire forces in the three axial directions, each node of the output layer 54 may output tire physical information such as the road surface friction coefficient and the moments around the three axes acting on the tire 10. The road surface friction coefficient output to the output layer 54 and the tire physical information such as the moments around the three axes acting on the tire 10 may be normalized values. The output layer 54 may output one type or a plurality of types of tire physical information by an arbitrary combination among the tire physical information such as the tire forces in the three axial directions, the road surface friction coefficient, and the moments around the three axes acting on the tire 10.

[0035] Also, in estimating the road surface friction coefficient, the output layer 54 may output an estimated value of the road surface friction coefficient, or may classify the road surface friction coefficient into categories such as dry, wet, snow-covered, or frozen states and output which category it belongs to.

[0036] Return to FIG. 2. The arithmetic model 32a uses what has been learned in one vehicle in another vehicle. The correction processing unit 33 of the tire physical information estimation device 30 corrects the output of the arithmetic model 32a (tire physical information output by the arithmetic model 32a) based on parameters set according to the vehicle, and estimates it as the final tire physical information.

[0037] The arithmetic model 32a constructs one model by learning in the same vehicle type such as a compact car, a light automobile, and a truck, and uses it in other vehicles of the same vehicle type. For example, the arithmetic model 32a learned in a vehicle of one vehicle name belonging to a compact car is used as the arithmetic model 32a for tire physical information in a vehicle of another vehicle name belonging to the same compact car.

[0038] The arithmetic model 32a can be learned by mounting tires 10 of specifications corresponding to the vehicle on an actual vehicle and test-driving the vehicle. The specifications of the tire 10 include information regarding the performance of the tire, such as, for example, tire size, tire width, aspect ratio, tire strength, tire outer diameter, load index, and manufacturing date.

[0039] The arithmetic model 32a may be learned by performing a rotation test while changing the road surface friction coefficient of the ground contact surface of the tire 10. Furthermore, the arithmetic model 32a can also be learned by mounting the tire 10 on an actual vehicle and test-driving the vehicle on road surfaces with different road surface friction coefficients.

[0040] As parameters set according to the vehicle, the correction processing unit 33 can use the standard deviation and the average value regarding the estimation of tire physical information in the vehicle equipped with the tire 10. An example of estimating the tire physical information in another vehicle names B and C belonging to the same compact car using the arithmetic model 32a learned in a vehicle of vehicle name A belonging to a compact car will be described. FIGS. 4(a) to 4(c) are graphs showing the correlation between the estimated values and the measured values in the same vehicle based on the arithmetic model 32a learned in the vehicle of vehicle name A. FIGS. 4(a) to 4(c) show the correlation regarding the tire forces Fx, Fy, and Fz, respectively.

[0041] Figures 5(a) to 5(c) are graphs showing the correlation between the estimated values and the measured values in the vehicle of vehicle name B. Figures 5(a) to 5(c) show the correlations for the tire forces Fx, Fy, and Fz, respectively. For the estimation of the tire forces Fx, Fy, and Fz in the vehicle of vehicle name B, the operation model 32a learned in the vehicle of vehicle name A is used.

[0042] Figures 6(a) to 6(c) are graphs showing the correlation between the estimated values and the measured values in the vehicle of vehicle name C. Figures 6(a) to 6(c) show the correlations for the tire forces Fx, Fy, and Fz, respectively. For the estimation of the tire forces Fx, Fy, and Fz in the vehicle of vehicle name C, the operation model 32a learned in the vehicle of vehicle name A is used.

[0043] Figure 7 is a chart showing the mean absolute error of the estimated values and the measured values. Although the mean absolute error regarding the tire forces Fx, Fy, and Fz in the vehicles of vehicle name B and vehicle name C is larger than that in the vehicle of vehicle name A, it is within a certain range. Therefore, the operation model 32a learned in the vehicle of vehicle name A can be used to estimate the tire physical information in vehicle name B and vehicle name C.

[0044] Figure 8 is a graph showing the standard deviation of the estimated values of the tire force F in each vehicle, and Figure 9 is a graph showing the average value of the estimated values of the tire force F in each vehicle. In Figures 8 and 9, the horizontal axis represents the vehicle weight and the vertical axis represents the load. As shown in Figure 8, the standard deviations of the tire force F in each of the vehicles of vehicle name A, vehicle name B, and vehicle name C are on a straight line with an upward slope in each axis direction. Also, as shown in Figure 9, the average values of the tire forces Fx and Fy in each of the vehicles of vehicle name A, vehicle name B, and vehicle name C are near 0, and the average value of Fz is on a straight line with an upward slope.

[0045] The correction processing unit 33 uses, as parameters, the ratio Ps of the standard deviation and the difference ΔPm of the average value in the vehicle of vehicle name B (or vehicle name C) with respect to the standard deviation and the average value in the vehicle of vehicle name A which is the reference for training the operation model 32a. The correction processing unit 33 corrects the estimated value of the tire force F by the operation model 32a in the vehicle of vehicle name B (or vehicle name C), and the corrected tire force Fc is given by the following formula Fc = F × Ps + ΔPm ···(1)

[0046] The correction processing unit 33 can apply formula (1) for each axial direction. Also, the correction processing unit 33 may set the difference ΔPm of the average value to 0 for the tire forces Fx and Fy. Further, as shown in FIGS. 8 and 9, based on the relationship between the standard deviation and the average value with respect to the vehicle weight, for vehicles other than vehicle name A, vehicle name B, and vehicle name C, the ratio Ps of the standard deviation and the difference ΔPm of the average value may be set and corrected based on the weight of the same vehicle.

[0047] The server device 40 acquires from the tire physical information estimation device 30 the physical quantity of the tire 10 measured by the sensor 20, and tire physical information such as the tire force F and the road surface friction coefficient estimated for the tire 10. The server device 40 may accumulate the physical quantity measured by the tire 10 and the tire physical information estimated by the tire physical information estimation device 30 from a plurality of vehicles.

[0048] Next, the operation of the tire physical information estimation system 100 will be described. FIG. 10 is a flowchart showing the procedure of the tire physical information estimation process by the tire physical information estimation device 30. The tire physical information estimation device 30 acquires, by the data acquisition unit 31, physical quantities such as the acceleration, strain, tire air pressure, and tire temperature in the tire 10 measured by the sensor 20 (S1).

[0049] The physical information estimation unit 32 extracts input data for a certain time interval from the data acquired by the data acquisition unit 31 (S2). In the estimation of tire physical information, at least one-axis (for example, circumferential direction) of acceleration data is required as input data. Also, in the estimation of tire physical information, for example, acceleration data for two axes in the circumferential direction and axial direction of the tire 10 may be used as input data, or acceleration data for three axes may be used as input data. Further, at least one or more time-series data among the strain, tire air pressure, and tire temperature in the tire 10 may be included in the input data.

[0050] The feature extraction unit 51 of the arithmetic model 32a executes a process of extracting feature amounts by performing a convolution operation and a pooling operation on the input data (S3). The fully connected unit 53 of the arithmetic model 32a performs an arithmetic operation by full connection on the feature amounts extracted by the feature extraction unit 51 and input to each node of the intermediate layer 52 (S4). Parameters such as weighting used in the full connection operation are determined in the learning of the arithmetic model 32a. By the full connection operation, normalized tire physical information such as, for example, the tire force F, the road surface friction coefficient, and the moments around the three axes acting on the tire 10 is output to each node of the output layer 54.

[0051] The physical information estimation unit 32 estimates the tire physical information by performing an inverse operation of normalization on the normalized tire physical information output to the output layer 54 and restoring it (S5). The correction processing unit 33 corrects the tire physical information estimated by the physical information estimation unit 32 based on parameters set according to the vehicle (S6), and ends the process.

[0052] The correction processing unit 33 of the tire physical information estimation system 100 corrects the tire physical information estimated by the physical information estimation unit 32 based on parameters set according to the vehicle. Thereby, the tire physical information estimation system 100 can easily construct an arithmetic model for another vehicle based on the arithmetic model learned in one vehicle and estimate the tire physical information. Further, the arithmetic model 32a of the tire physical information estimation system 100 can improve the estimation accuracy of the tire physical information by being configured to output at least biaxial tire physical information normalized from the output layer 54.

[0053] By using the standard deviation and the average value of the tire physical information corresponding to the vehicle as parameters, the correction processing unit 33 can improve the estimation accuracy of the tire physical information. Further, the correction processing unit 33 can enhance the versatility of the arithmetic model 32a by setting parameters based on the weight of the vehicle.

[0054] The tire physical information estimation system 100 can construct a reference arithmetic model 32a by learning according to, for example, the vehicle type, and apply the learned arithmetic model 32a to the arithmetic models 32a of vehicles with different vehicle names belonging to the same vehicle type. Further, the tire physical information estimation system 100 can construct a reference arithmetic model 32a by learning according to the driving method (front-wheel drive, rear-wheel drive, and four-wheel drive) among the same vehicle type, and apply the learned arithmetic model 32a to the arithmetic models 32a of vehicles with different vehicle names belonging to the same driving method of the same vehicle type.

[0055] In addition to the vehicle type, the tire physical information estimation system 100 may construct a reference arithmetic model 32a by learning according to the type of the tire 10, and apply the learned arithmetic model 32a to the arithmetic model 32a for another specification of the tire 10 belonging to the same tire type. For example, as the tire type focusing on the grooves in the tread portion of the tire 10, a reference arithmetic model 32a may be constructed according to all-season tires, studless tires, snow tires, slick tires, and the like.

[0056] In the calculation model 32a of the tire physical information estimation system 100, by providing a fully connected part 53 for each of a plurality of pieces of tire physical information to be output and making the calculation model 32a multitask, the calculation scale can be reduced. The feature extraction part 51 of the calculation model 32a reduces the calculation scale by sharing the filters in the convolutional operations 51a and 51c. Further, the physical information estimation part 32 can provide information necessary for analyzing the behavior such as slip in the tire 10 by estimating all of the tire forces F in the three-axis directions.

[0057] (Modification example) FIG. 11 is a block diagram showing the functional configuration of the tire physical information estimation system 100 according to the modification example. In the modification example shown in FIG. 11, data input to the calculation model 32a is acquired from the vehicle control device 90. Incidentally, both the data from the vehicle control device 90 and the data from the sensor 20 (see FIG. 2) can be used as the data input to the calculation model 32a.

[0058] The vehicle control device 90 acquires, for example, running data such as the running speed of the vehicle, the acceleration in the three-axis directions, and the three-axis angular velocity, and load data such as the weight of the vehicle and the axle load applied to the axles, in a digital tachometer or the like of the vehicle. The vehicle control device 90 outputs these running data and load data to the tire physical information estimation device 30.

[0059] The tire physical information estimation device 30 estimates tire physical information such as the tire force F, the road surface friction coefficient, and the moment around the three axes acting on the tire 10 using the calculation model 32a for the data input from the vehicle control device 90. Incidentally, the calculation model 32a is constructed by performing learning to estimate tire physical information for the data input from the vehicle control device 90 in advance, for example, by test running with an actual vehicle.

[0060] In the above-described embodiments and modifications, the tire physical information estimated by the arithmetic model 32a has been described with respect to the tire force F, the road surface friction coefficient, and the moments about the three axes acting on the tire 10. However, for example, it is also possible to estimate the looseness of fastening parts such as wheel nuts used for attaching the tire 10. Since the vibration due to the looseness of the fastening parts such as wheel nuts appears in the acceleration data measured by the tire 10, an arithmetic model 32a for estimating the looseness of the fastening parts is constructed and learned in a CNN type by comparing with other tire forces F. The tire physical information estimation system 100 can execute the arithmetic operation by the arithmetic model 32a based on input data such as acceleration data acquired during actual vehicle running, and estimate the looseness of the fastening parts of the tire 10 in real time.

[0061] Further, the sensor 20 is not limited to each sensor described with reference to FIG. 1. For example, a microphone provided around the tire 10 or the tire 10 may be used. The arithmetic model 32a may estimate the tire physical information using the voice data collected by a microphone or the like.

[0062] In the above-described embodiments and modifications, the CNN type LeNet model is used for the arithmetic model 32a. However, model structures such as a so-called DenseNet model, ResNet model, MobileNet model, and PeleelNet model may also be used. Further, a module structure such as a Dense Block, Residual Block, or Stem Block may be incorporated into the arithmetic model 32a to construct the model.

[0063] Next, the features of the tire physical information estimation system 100 and the tire physical information estimation method according to the embodiment will be described. The tire physical information estimation system 100 according to the embodiment includes a physical information estimation unit 32 and a correction processing unit 33. The physical information estimation unit 32 has a learning-based arithmetic model 32a from an input layer 50 to an output layer 54, and estimates tire physical information generated by the movement of the tire 10. The correction processing unit 33 corrects the tire physical information estimated by the physical information estimation unit 32. The arithmetic model 32a has a feature extraction unit 51 that executes convolution operations 51a and 51c in intermediate operations from the input layer 50 toward the output layer 54, and outputs at least biaxial tire physical information normalized from the output layer 54. The correction processing unit 33 corrects the tire physical information estimated by the physical information estimation unit 32 based on parameters set according to the vehicle. Thereby, the tire physical information estimation system 100 can easily construct an arithmetic model in another vehicle based on an arithmetic model learned in one vehicle and estimate the tire physical information.

[0064] Also, the correction processing unit 33 corrects using the standard deviation and the average value of the tire physical information corresponding to the vehicle as parameters. Thereby, the tire physical information estimation system 100 can improve the estimation accuracy of the tire physical information.

[0065] Also, the correction processing unit 33 sets parameters based on the weight of the vehicle. Thereby, the tire physical information estimation system 100 can enhance the versatility of the arithmetic model 32a.

[0066] Also, the tire physical information is the tire force in three axial directions. Thereby, the tire physical information estimation system 100 can provide information necessary for analyzing the behavior such as slip in the tire 10.

[0067] The tire physical information estimation method includes a physical information estimation step and a correction processing step. The physical information estimation step estimates tire physical information generated by the movement of the tire 10 based on the learning-based operation model 32a from the input layer 50 to the output layer 54. The correction processing step corrects the tire physical information estimated by the physical information estimation unit 32. The operation model 32a has a feature extraction unit 51 that executes convolution operations 51a and 51c in the intermediate operations from the input layer 50 to the output layer 54, and outputs at least biaxial tire physical information normalized from the output layer 54. The correction processing step corrects the tire physical information estimated in the physical information estimation step based on parameters set according to the vehicle. According to this tire physical information estimation method, it is possible to simply construct an operation model in another vehicle based on an operation model learned in one vehicle and estimate the tire physical information.

[0068] As described above, the embodiments of the present invention have been described based on the embodiments. These embodiments are examples, and it is understood by those skilled in the art that various modifications and changes are possible within the scope of the claims of the present invention, and such modifications and changes are also within the scope of the claims of the present invention. Therefore, the descriptions and drawings in this specification should be treated as illustrative rather than restrictive.

Explanation of reference numerals

[0069] 10 Tire, 32 Physical information estimation unit, 32a Operation model, 33 Correction processing unit, 50 Input layer, 51 Feature extraction unit, 51a, 51c Convolution operations, 54 Output layer, 100 Tire physical information estimation system.

Claims

1. A physical information estimation unit having a learning-based operation model from an input layer to an output layer, for estimating tire physical information generated by the movement of a tire, and a correction processing unit for correcting the tire physical information estimated by the physical information estimation unit, comprising: The operation model has a feature extraction unit that executes a convolution operation in an intermediate operation from the input layer to the output layer, and outputs at least biaxial tire physical information normalized from the output layer. The correction processing unit corrects the tire physical information estimated by the physical information estimation unit based on parameters set according to the vehicle. A tire physical information estimation system characterized by the above.

2. The correction processing unit corrects using the standard deviation and average value of tire physical information according to the vehicle as the parameters. The tire physical information estimation system according to claim 1.

3. The correction processing unit sets the parameters based on the weight of the vehicle. The tire physical information estimation system according to claim 1 or 2.

4. The tire physical information is tire force in three axial directions. The tire physical information estimation system according to any one of claims 1 to 3.

5. A physical information estimation step of estimating tire physical information generated by the movement of a tire based on a learning-based operation model from an input layer to an output layer, and a correction processing step of correcting the tire physical information estimated by the physical information estimation step, comprising: The operation model has a feature extraction unit that executes a convolution operation in an intermediate operation from the input layer to the output layer, and outputs at least biaxial tire physical information normalized from the output layer. The correction processing step corrects the tire physical information estimated by the physical information estimation step based on parameters set according to the vehicle. A tire physical information estimation method characterized by the above.

Images