Tire state estimation device, tire state estimation method, and tire state estimation program

Abstract

A tire state estimation device 10 comprises: an acquisition unit 11B that performs a load test on a tire in which a conductive sensing rubber, capable of outputting an electric signal corresponding to deformation of a tread in at least a part of a partial region inside the position of the ground contact surface of the tread, is provided in the partial region, and thereby acquires an electric signal; and an estimation unit 11C that estimates a physical quantity relating to stress in the ground contact surface of the tire by using a tire state estimation model 13D that has been trained in advance with an electric signal acquired by the acquisition unit 11B or an electric resistance value obtained by the electric signal serving as input information and with the physical quantity serving as output information.

Description

Tire condition estimation device, tire condition estimation method, and tire condition estimation program

[0001] The present disclosure relates to a tire condition estimation device, a tire condition estimation method, and a tire condition estimation program.

[0002] Conventionally, the following technologies have existed as technologies for measuring the stress applied to the grounding surface of a vehicle tire.

[0003] Japanese Patent Application Laid-Open No. 2021-89203 discloses a tire grounding characteristic measurement device aimed at increasing the resolution in the rotational axis direction without depending on the size of the sensor.

[0004] This tire grounding characteristic measurement device includes a cylindrical rotating drum rotatable around a rotation axis, drum driving means for rotationally driving the rotating drum around the rotation axis, and a sensor array embedded in the cylindrical surface of the rotating drum for measuring the stress applied to the tire contacting the cylindrical surface. The sensor array has a plurality of stress measurement regions arranged in a vertical row in the rotation axis direction of the cylindrical surface and capable of independently measuring stress. A plurality of the sensor arrays are provided at different positions in the circumferential direction in which the rotating drum rotates. The positions of the stress measurement regions in the rotation axis direction of the stress measurement regions respectively included in the plurality of sensor arrays are arranged at different positions from each other among the plurality of sensor arrays. Stress measurement means, a processing device for calculating tire grounding characteristics, which are characteristics in the grounding region contacting the rotating drum of the tread surface of the tire, based on the stress measured by at least two of the plurality of sensor arrays included in the stress measurement means, and an output device for outputting the tire grounding characteristics calculated by the processing device.

[0005] However, in conventional technologies including the technology described in Japanese Patent Application Laid-Open No. 2021-89203, generally, in order to measure a physical quantity related to the stress applied to the grounding surface of a tire, a sensor for measuring the physical quantity is arranged separately from the tire, so there is room for improvement in terms of the measurement accuracy of the physical quantity.

[0006] This disclosure has been made in view of the above points, and aims to provide a tire condition estimation device, a tire condition estimation method, and a tire condition estimation program that can contribute to improving the accuracy of measuring physical quantities related to stress at the contact surface of a tire.

[0007] A tire condition estimation device according to a first aspect of the present disclosure includes: an acquisition unit that acquires an electrical signal by performing a load test on a tire on which a conductive sensing rubber is provided in a part of the interior region of the contact surface of the tread, and which is capable of outputting an electrical signal corresponding to the deformation of the tread in at least a part of the part of the interior region of the contact surface of the tread; and an estimation unit that takes the electrical signal acquired by the acquisition unit, or the electrical resistance value obtained by the electrical signal, as input information, and estimates the physical quantity using a tire condition estimation model that has been learned in advance as output information for the physical quantity relating to the stress at the contact surface of the tire.

[0008] A tire condition estimation device according to a second aspect of this disclosure is a tire condition estimation device according to a first aspect, wherein the tire condition estimation model is a model based on a recurrent neural network.

[0009] A tire condition estimation device according to a third aspect of this disclosure is a tire condition estimation device according to a second aspect, wherein the model based on the recurrent neural network is a model based on reservoir computing.

[0010] A tire condition estimation device according to a fourth aspect of the present disclosure is a tire condition estimation device according to any one of the first to third aspects, wherein the tire is a tire in which a sensing member, in which the sensing rubber is laminated in part, is bonded to a part of the tread at a part of the contact surface position on the tread, as a substitute for the tread in that part.

[0011] A tire condition estimation device according to a fifth aspect of the present disclosure is a tire condition estimation device according to a fourth aspect, wherein the sensing member includes a tread rubber made of the same material as the tread, a sensing rubber laminated on the tread rubber, and a metal plate in contact with the end of the sensing rubber for extracting the electrical signal generated by the sensing rubber to the outside.

[0012] A tire condition estimation device according to a sixth aspect of this disclosure is a tire condition estimation device according to a fifth aspect, wherein the sensing member further includes a coating rubber for pressing the metal plate against the sensing rubber.

[0013] A tire condition estimation method according to a seventh aspect of this disclosure involves performing a load test on a tire in which a conductive sensing rubber is provided in a part of the interior region of the contact surface of the tread, capable of outputting an electrical signal corresponding to the deformation of the tread in at least a part of the said part of the region, thereby acquiring an electrical signal. The acquired electrical signal, or the electrical resistance value obtained from said electrical signal, is used as input information, and a pre-learned tire condition estimation model is used as output information to estimate the physical quantity related to the stress at the contact surface of the tire. The computer then performs the following processing.

[0014] The eighth aspect of the present disclosure describes a tire condition estimation program that performs a load test on a tire in which a conductive sensing rubber is provided in a part of the interior of the contact surface of the tread, capable of outputting an electrical signal corresponding to the deformation of the tread in at least a part of the said part of the part of the tire. The program acquires an electrical signal by performing a load test on the tire, and uses the acquired electrical signal or the electrical resistance value obtained from the electrical signal as input information, and uses a pre-learned tire condition estimation model to estimate a physical quantity related to the stress at the contact surface of the tire as output information.

[0015] The technology of this disclosure has the effect of contributing to improving the accuracy of measuring physical quantities related to stress at the tire contact surface.

[0016] This figure shows an example of the overall configuration of the tire condition estimation system according to the embodiment. This figure shows an example of the hardware configuration of the tire condition estimation device according to the embodiment. This is a schematic diagram showing an example of the tire condition estimation model according to the embodiment. This figure is used to explain the tire condition estimation method according to the embodiment, and is a cross-sectional view showing the tire used in the numerical analysis using the finite element method and multiple positions in the depth direction inside it. This figure is used to explain the tire condition estimation method according to the embodiment, and is a cross-sectional view showing the C direction, Z direction, and R direction. This is a graph showing an example of the three-dimensional pressure characteristics on the surface and inside of a tire using the finite element method under the first condition according to the embodiment. This is a graph showing an example of the three-dimensional pressure characteristics on the surface and inside of a tire using the finite element method under the second condition according to the embodiment. This is a plan view showing the configuration of the member for examination according to the embodiment. This is a perspective view showing the configuration of the member for examination according to the embodiment and the state of the test using the member for examination. This is a graph showing the time-dependent changes in the measured and estimated values ​​of the three-component force distribution (load distribution) on the sensing rubber and the NMSE when all of the first resistance value, second resistance value, and third resistance value obtained from the member for examination according to the embodiment are used. This is a perspective view showing an example of the configuration of the sensing member according to the embodiment. This is a perspective view showing an example of the manufacturing process of a sensing member according to the embodiment. This is a perspective view showing an example of the mounting state of the sensing member according to the embodiment to a tire. This is a diagram showing an example of the comparison result between the estimated value and the measured value of the three-component force distribution using the sensing member and tire state estimation model according to the embodiment. This is a block diagram showing an example of the functional configuration of the tire state estimation device according to the embodiment during the training of the tire state estimation model. This is a block diagram showing an example of the functional configuration of the tire state estimation device according to the embodiment during the operation of the tire state estimation model. This is a flowchart showing the flow of the training process according to the embodiment. This is a flowchart showing the flow of the tire state estimation process according to the embodiment.

[0017] The following describes examples of embodiments of the disclosed technology with reference to the drawings. In the following description, the case where the three-component force distribution at the tire contact surface is applied as a physical quantity relating to the stress at the tire contact surface of the disclosed tire will be explained. Note that the same or equivalent components and parts are given the same reference numerals in each drawing. Also, the dimensional ratios in the drawings are exaggerated for illustrative purposes and may differ from actual ratios.

[0018] The tire condition estimation system according to this embodiment estimates the three-component force distribution at the contact surface of the tread of a vehicle tire. First, the overall configuration of the tire condition estimation system will be described.

[0019] Figure 1 shows an example of the overall configuration of the tire condition estimation system 90 according to this embodiment. As shown in Figure 1, the tire condition estimation system 90 of this embodiment is composed of a tire condition estimation device 10, a test tire 50 to be used for estimation of the three-component force distribution, and a sensing member 60 arranged in a part of the tread area of ​​the tire 50. Examples of the tire condition estimation device 10 include information processing devices such as personal computers and server computers.

[0020] In the tire condition estimation system 90 according to this embodiment, the tire condition estimation device 10 estimates the three-component force distribution at the contact surface of the tire 50 using the electrical resistance value indicated by the electrical signal output from the sensing member 60, under load test conditions (hereinafter referred to as "test condition") in which the tire 50 is in contact with the road surface and rotated. For this purpose, the tire 50 is provided with a slip ring 52, and the sensing member 60 and the tire condition estimation device 10 are electrically connected via the slip ring 52. It goes without saying that, although not shown in the figures, electrical elements such as an A / D (analog / digital) converter that converts the electrical signal into digital data and an amplifier are provided between the slip ring 52 and the tire condition estimation device 10.

[0021] As described above, this embodiment explains the case in which the sensing member 60 and the tire condition estimation device 10 are connected via a slip ring 52, but the embodiment is not limited to this configuration. For example, the tire 50 may be provided with a power supply for powering the sensing member 60, and a memory backed up by the power supply. In this configuration, during the test state, the electrical resistance value indicated by the electrical signal output from the sensing member 60 is converted into digital data in real time and stored in the memory. Then, in this configuration, after the test state is completed, the electrical resistance value is transferred from the memory to the tire condition estimation device 10.

[0022] Figure 2 is a block diagram showing an example of the hardware configuration of the tire condition estimation device 10 according to this embodiment. As shown in Figure 2, the tire condition estimation device 10 includes a CPU (Central Processing Unit) 11, a memory 12 as a temporary storage area, a non-volatile storage unit 13, an input unit 14 such as a keyboard and mouse, a display unit 15 such as a liquid crystal display, a media read / write device (R / W) 16, and a communication interface (I / F) unit 18. The CPU 11, memory 12, storage unit 13, input unit 14, display unit 15, media read / write device 16, and communication I / F unit 18 are connected to each other via bus B. The media read / write device 16 reads information written on the recording medium 17 and writes information to the recording medium 17.

[0023] The storage unit 13 is implemented by an HDD (Hard Disk Drive), SSD (Solid State Drive), flash memory, etc. The storage unit 13, as a storage medium, stores the learning processing program 13A and the tire state estimation program 13B. The learning processing program 13A and the tire state estimation program 13B are stored (installed) in the storage unit 13 when the recording medium 17 on which the programs are written is set in the media read / write device 16, and the media read / write device 16 reads the programs from the recording medium 17. The CPU 11 reads the learning processing program 13A and the tire state estimation program 13B from the storage unit 13 as appropriate, loads them into the memory 12, and sequentially executes the processes of each program. This executes the learning process and tire state estimation process, which will be described in detail later.

[0024] Furthermore, the memory unit 13 stores the training data 13C and the tire state estimation model 13D that has been trained using the training data 13C.

[0025] The tire state estimation model 13D according to this embodiment is pre-learned by a sensing member 60 provided on the tire 50, using the electrical resistance value indicated by the electrical signal obtained during the above-described test state as input information, and outputting a physical quantity related to the stress on the contact surface of the tire 50, specifically the three-component force distribution.

[0026] In this embodiment, the tire state estimation model 13D is a reservoir computing model (ESN (Echo State Network)) which is similar to a model using a recurrent neural network (RNN (RNN)), as an example schematically shown in Figure 3. This is because the reservoir computing model is a network in which time-series information is accumulated and echoes like a reservoir, making it suitable for reproducing the internal state of a tire while it is running (rotating). In this embodiment, the tanh (hyperbolic tangent) function, which is a nonlinear function, is applied as the activation function, and the hyperparameters are determined by grid search. The reason for applying the tanh function, which is a nonlinear function, as the activation function is that it is suitable for reproducing the internal state of the tire. However, this is not the only form; for example, a sigmoid function may be applied as the activation function, or the hyperparameters may be determined by random search.

[0027] The training data 13C is a dataset that contains a large number of sets of electrical resistance values ​​and the corresponding three-component force distributions, which are used when training the tire state estimation model 13D.

[0028] The learning data 13C according to this embodiment is acquired in advance using a test machine that has substantially the same configuration as the tire condition estimation system 90 shown in Figure 1, except that a three-component force pressure sensor is provided at the contact point with the tire 50 on the road surface.

[0029] Using this testing machine, tests are performed on a large number of tires equipped with sensing elements 60. A large number of sets of electrical resistance values ​​and three-component force distributions obtained from the sensing elements 60 are acquired and pre-registered as training data 13C.

[0030] Here, with reference to Figures 4 to 10, the principle of the method for estimating the three-component force distribution using the sensing member 60 on the tire contact surface, according to the tire condition estimation device 10 of this embodiment, will be explained.

[0031] To obtain the three-component force distribution at the tire's contact surface, it is preferable to install a sensor such as a three-component force pressure sensor (hereinafter simply referred to as "sensor") on the tire's contact surface and take measurements, in order to obtain a highly accurate three-component force distribution. However, in this case, the sensor is quickly destroyed by wear from the road surface or impact from the road surface.

[0032] Therefore, the developers of the technology disclosed herein considered providing a thin, plate-shaped sensing rubber inside the tread of the target tire as a sensor to obtain the three-component force distribution. However, it is difficult to manufacture all of the target tires with the sensing rubber already embedded. Therefore, the developers of the technology disclosed herein considered replacing a portion of the tire tread with a block-shaped sensing member 60 (see also Figure 11) containing the sensing rubber.

[0033] To realize this sensing member 60, the developers of the technology disclosed hereby first investigated the optimal position of the sensing rubber in the depth direction from the tire tread surface using numerical analysis with the finite element method (FEM).

[0034] Figure 4 shows the tire 50 used in this study, along with cross-sectional views showing multiple locations within its interior in the depth direction. Note that Figure 4 shows only the upper half of one half of the tire in the width direction, with the tire equatorial plane CL as the boundary.

[0035] As shown in the left diagram of Figure 4, the tire 50 used in the numerical analysis in this embodiment has a belt 3 and a tread layer 4 in this order on the radially outer side of the crown portion of the carcass 2.

[0036] Belt 3 consists of four belt layers 3a to 3d: an inclined belt 3bc consisting of two layers, a first belt layer 3c and a second belt layer 3b located inside the first belt layer 3c in the tire radial direction (located adjacent in this example); an outer belt layer 3d located outside the inclined belt 3bc in the tire radial direction; and an inner belt layer 3a located inside the inclined belt 3bc in the tire radial direction. Each belt layer is made of belt cord covered with belt coating rubber.

[0037] Furthermore, the tire 50 has a center cushion rubber 5 between the inner belt layer 3a (innermost belt layer) and the carcass 2, and the tread layer 4 has at least two shoulder main grooves 7 located on both outer sides in the tire width direction. In addition, the tire 50 has shoulder cushion rubber 6 positioned between the widest belt layer (second belt layer 3b), which has the widest width in the tire width direction among the multiple belt layers, and the carcass 2, and in the region on the outer side in the tire width direction of the belt layer (inner belt layer 3a) that is located radially inward from the widest belt layer 3b. Moreover, the tire 50 has an inner liner 8 that forms the inner surface of the tire, positioned on the inner side of the tire cavity from the carcass 2.

[0038] Using the tire 50 configured as described above, the developers of the technology disclosed herein conducted a finite element method study. As shown in the right-hand figure of Figure 4, numerical analysis was performed on a portion of the tire width direction, straddling the tire equatorial plane CL at the top of the tire 50, and targeting the region from the tread layer 4 to the inner liner 8. In this study, four positions were assumed for the sensing rubber: the position on the surface of the tire 50 that contacts the road surface (hereinafter referred to as "surface"), the central position in the depth direction of the tread layer 4 (hereinafter referred to as "middle"), the bottom surface of the tread layer 4 (hereinafter referred to as "bottom"), and the top surface of the inner liner 8 (hereinafter referred to as "I / L"). Time-series data of the force acting on the sensing rubber when it is placed at these positions was derived using the finite element method. The following two conditions were applied as conditions for this study.

[0039]

[0040] Specifically, the first condition is that a load of 21.97 kN is applied to the tire 50 at 100% and free rolling is performed, and the second condition is that a lateral force (SF) of 0.3 G is applied to the tire 50 in addition to the first condition. In the following, as shown in Figure 5, the three directions in the three-component force distribution at this time will be referred to as the C direction, which is the rotation angle direction (tractive direction) of the tire 50, the Z direction, which is the width direction (lateral direction) of the tire 50, and the R direction, which is the radial direction of the tire 50. Figure 5 is a diagram used to explain the tire state estimation method according to this embodiment, and is a cross-sectional view showing the C direction, the Z direction, and the R direction.

[0041] Figure 6 shows a graph illustrating an example of the three-dimensional pressure characteristics on the surface and inside of tire 50 using the finite element method under the first condition. Figure 7 also shows a graph illustrating an example of the three-dimensional pressure characteristics on the surface and inside of tire 50 using the finite element method under the second condition. The graphs in Figures 6 and 7 show time-series data of the force applied to the sensing rubber at three locations in the tire width direction shown in Figure 4: position 2 on the serial side, position 5 in the center, and position 9 on the non-serial side.

[0042] As shown in Figures 6 and 7, the time-series changes in the force applied to the sensing rubber when it is placed at the middle and bottom positions show a very high correlation with the changes when the sensing rubber is placed on the surface, for all positions 2, 5, and 9, under both the first and second conditions. In particular, the correlation is extremely high at position 5, which is the center in the tire width direction.

[0043] In contrast, the time-series change in the force applied to the sensing rubber when it is placed in the I / L position remains almost flat (almost unchanged) at all positions and conditions, which is clearly different from the change observed when the sensing rubber is placed on the surface.

[0044] In addition, although representative graphs were illustrated in FIGS. 6 and 7, the above tendency is not limited to those shown in these graphs, and a tendency substantially the same as that shown in FIGS. 6 and 7 was observed at any position from position numbers 1 to 10 in the tire width direction shown in FIG. 4.

[0045] On the other hand, the tire state estimation method according to the present embodiment is a method for estimating the three-force distribution using the tire state estimation model 13D. Therefore, as long as the arrangement position has a certain degree of correlation with the time-series change of the force on the surface, even if the difference is minute, the three-force distribution can be accurately estimated.

[0046] Therefore, as shown in the following table, as a result, it was obtained that there is no problem in estimating the three-force distribution on the ground contact surface of the tire if the sensing rubber is arranged at either the middle or the bottom position.

[0047]

[0048] Next, the developers of the technology of the present disclosure examined the number of electrodes for outputting an electrical signal from the sensing rubber that can accurately estimate the three-force distribution in the sensing member 60 incorporating the sensing rubber. For this purpose, the developers of the technology of the present disclosure created the flat plate-shaped examination member 80 shown in FIGS. 8 and 9. FIG. 8 is a plan view showing the configuration of the examination member 80 according to the present embodiment, and FIG. 9 is a perspective view showing the configuration of the examination member 80 according to the present embodiment and the state of the test using the examination member 80.

[0049] As shown in FIGS. 8 and 9, in this examination member 80, a rectangular sensing rubber 40 in a plan view is arranged on the upper surface of a flat plate-shaped surface pressure distribution measurement sensor 42. Here, the sensing rubber 40 is made rectangular because it is assumed that the shape of the sensing member 60 is a rectangular parallelepiped as shown in FIG. 11 as an example in order to make it easy to cut out a part of the tread portion of the tire and embed it. Here, as the surface pressure distribution measurement sensor 42, the product name RK1A-Z01 SV manufactured by Nissha Co., Ltd. was used.

[0050] Furthermore, the testing member 80 is provided with a positive electrode 32 for applying a positive voltage to one end of one of the four sides of the sensing rubber 40. Similarly, the testing member 80 is provided with a first electrode 34A, a second electrode 34B, and a third electrode 34C, one each at the end of the remaining three sides of the sensing rubber 40, in a counterclockwise direction in plan view.

[0051] The developers of the technology disclosed herein used the study member 80 to conduct the following study using electrical resistance values ​​corresponding to electrical signals obtained individually from each electrode, the first electrode 34A, the second electrode 34B, and the third electrode 34C, as shown in Figure 9 as an example.

[0052] First, a predetermined positive voltage was applied to the positive electrode 32 of the test member 80, and a load was applied individually to each of the first region 40A, second region 40B, third region 40C, and fourth region 40D of the sensing rubber 40 (hereinafter referred to as the "loaded state"). In this loaded state, time-series data of electrical resistance values ​​was acquired from each electrode, and the three-component force distribution was acquired from the surface pressure distribution measurement sensor 42. This process was repeated many times to create a large number of pairs of electrical resistance values ​​and three-component force distributions from each electrode.

[0053] Next, using the large number of pairs of electrical resistance values ​​and three-component force distributions that were created, learning was performed for each individual and multiple combination of electrical resistance values, thereby generating a learning model using a reservoir computing method for each of those individual and multiple combinations.

[0054] Subsequently, the member 80 for examination was again subjected to a load, and time-series data of electrical resistance values ​​were acquired from each electrode, while the three-component force distribution was acquired from the surface pressure distribution measurement sensor 42. This allowed for the acquisition of pairs of electrical resistance values ​​and three-component force distributions for each electrode. The acquired electrical resistance values ​​were then input into the corresponding learning models, both individually and in combination, to estimate the three-component force distribution for each learning model. The error between the estimated three-component force distribution and the three-component force distribution obtained by the surface pressure distribution measurement sensor 42 was then derived. In the following, the electrical resistance value obtained from the first electrode 34A will be referred to as the "first resistance value," the electrical resistance value obtained from the second electrode 34B as the "second resistance value," and the electrical resistance value obtained from the third electrode 34C as the "third resistance value."

[0055] As a result, the results shown in the following table were obtained. Here, the average NMSE (Normalized Mean Square Error) was applied to each of the first region 40A, second region 40B, third region 40C, and fourth region 40D under conditions where loads were individually applied.

[0056]

[0057] As shown in this table, the average NMSE when using the first resistance value, second resistance value, and third resistance value individually is all 0.5 or higher, indicating that it is not necessarily possible to estimate the three-component force distribution with high accuracy.

[0058] In contrast, the average NMSE for two combinations of the first, second, and third resistance values, and for all combinations, was less than 0.5 in all cases, indicating that the three-component force distribution can be estimated with high accuracy in all cases. In particular, the average NMSE for all combinations of the first, second, and third resistance values ​​was 0.17, indicating that the three-component force distribution can be estimated with very high accuracy.

[0059] Figure 10 is a graph showing the measured and estimated changes over time in the three-component force distribution (load distribution) on the sensing rubber 40, along with the NMSE, when all three resistance values ​​(first, second, and third) are used.

[0060] As shown in Figure 10, the NMSE when a load was applied to the first region 40A was 0.207, and the NMSE when a load was applied to the second region 40B was 0.261. Similarly, the NMSE when a load was applied to the third region 40C was 0.072, and the NMSE when a load was applied to the fourth region 40D was 0.156. As a result, the average NMSE was 0.17.

[0061] Therefore, in the tire condition estimation system 90 according to this embodiment, the sensing member 60 is provided with electrodes corresponding to all electrodes, including the positive electrode 32 and the first electrode 34A to the third electrode 34C, and the three-component force distribution is estimated using all electrical resistance values, including the first resistance value, the second resistance value, and the third resistance value. However, the system is not limited to this form, and for example, the three-component force distribution may be estimated using a combination of the first resistance value and the second resistance value, or a combination of the first resistance value and the third resistance value.

[0062] In this embodiment, brass plates, which are metal plates plated with brass, were used as the positive electrode 32 and the first electrodes 34A to the third electrodes 34C. This is because brass plates adhere strongly to rubber and are less likely to peel off when strain is applied.

[0063] Next, the developers of the technology disclosed herein considered the layer structure and manufacturing method of the sensing member 60.

[0064] As a result, the results shown in the following table were obtained. In the following table, sensing rubber is denoted as "SR" and tread rubber is denoted as "tread".

[0065]

[0066] As shown in this table, three types of layer configurations were considered: a first layer configuration in which tread rubber, sensing rubber, brass plate, and tread rubber are laminated in that order; a second layer configuration in which sensing rubber is interposed between the brass plate and tread rubber in the first layer configuration; and a third layer configuration in which coating rubber is interposed between the brass plate and tread rubber in the first layer configuration.

[0067] Furthermore, four manufacturing methods were investigated: a first method in which raw rubber (both tread rubber and sensing rubber) is vulcanized without being cut to the specified dimensions; a second method in which raw rubber is cut to the specified dimensions, a brass plate is bent into a V-shape and inserted into the center, and then vulcanized; a third method in which polyimide tape is not applied to the sides before vulcanization; and a fourth method in which the sides are covered with polyimide tape from the beginning before vulcanization.

[0068] As a result, as shown in this table, it was found that a suitable sensing member 60 that can be actually applied can be created by a combination of the third layer configuration and the second manufacturing method, or a combination of the third layer configuration and the fourth manufacturing method.

[0069] Furthermore, the reason for providing a coating rubber in the third layer configuration is as follows:

[0070] In this embodiment, the sensing rubber preferably has a lower electrical resistance than the tread rubber in order to facilitate the extraction of electrical signals. For this reason, a large amount of carbon black is blended into the sensing rubber. In this case, if additives such as zinc oxide are added, the sensing rubber becomes very hard and abnormal values ​​in the electrical signals are more likely to occur, so such additives should not be added.

[0071] Therefore, the sensing rubber is prone to peeling off from the brass plate, which is a metal plate. To suppress this peeling, it was envisioned that a coating rubber would be provided.

[0072] Furthermore, for similar reasons, the coating rubber according to this embodiment is formulated with zinc oxide. This further enhances the effect of preventing delamination between the sensing rubber and the brass plate.

[0073] Thus, in this embodiment, a brass plate is used as the electrode of the sensing member, but the invention is not limited to this form. Any material that has high adhesion to rubber, is resistant to peeling when deformation occurs, and can conduct electrical signals can be used as an electrode.

[0074] Figure 11 shows a perspective view illustrating an example of the configuration of the sensing member 60 according to this embodiment, which was created based on the above considerations.

[0075] As shown in Figure 11, the sensing member 60 according to this embodiment includes a tread rubber 62 made of the same material as the tread layer 4 of the tire 50, and a sensing rubber 64 laminated on the tread rubber 62 for detecting the pressure generated inside the tread of the tire 50 as an electrical signal. The sensing member 60 also includes a metal plate (a brass plate in this embodiment) 66 which is in contact with the end of the sensing rubber 64 and is an electrode for extracting the electrical signal generated by the sensing rubber 64 to the outside, and a coating rubber 68 for pressing the metal plate 66 against the sensing rubber 64. Furthermore, the sensing member 60 includes a tread rubber 70 made of the same material as the tread layer 4 of the tire 50, which is laminated on the side of the coating rubber 68 opposite to the sensing rubber 64.

[0076] Although Figure 11 only shows two metal plates 66 which serve as electrodes, as mentioned above, in reality, there are four metal plates 66, one for each of the four sides of the sensing rubber 64.

[0077] Figure 12 shows a perspective view illustrating an example of the manufacturing process for the sensing member 60 according to this embodiment.

[0078] As shown in the left diagram of Figure 12, first, the raw rubber is cut to the specified size, and then, as shown in the middle diagram of Figure 12, the sides are secured with polyimide tape and vulcanized. At this time, the mold is made to the specified size. As a result, the sensing member 60 is completed as shown in the right diagram of Figure 12.

[0079] Figure 13 shows a perspective view illustrating an example of how the sensing member 60 according to this embodiment is attached to the tire 50.

[0080] As shown in the left diagram of Figure 13, first, a notch 54 is created by cutting out the central part of the tread layer 4 of the tire 50 in the tire width direction. Needless to say, the size of the notch 54 is such that the sensing member 60 fits inside.

[0081] Then, as shown in the right-hand diagram of Figure 13, the sensing member 60 is fitted into the notch 54 and bonded. In this embodiment, the sensing member 60 is bonded using cyanoacrylate adhesive, but it goes without saying that the invention is not limited to this form.

[0082] As mentioned above, when estimating the three-component force distribution at the contact surface of a tire 50 equipped with a sensing member 60, the sensing member 60 and the tire condition estimation device 10 are electrically connected via a slip ring 52, as shown in Figure 1 as an example.

[0083] Figure 14 shows an example of a comparison between estimated and measured values ​​of the three-component force distribution using the sensing member 60 and tire state estimation model 13D according to this embodiment. In Figure 14, the time-series estimated and measured values ​​of pressure in the Z direction (indicated as "Fz" in Figure 14) and C direction (indicated as "Fc" in Figure 14) at the four positions P1 to P4 shown in the lower left figure are shown in the right figure.

[0084] As shown in Figure 14, the time-series changes of the estimated and measured values ​​in this case are almost identical, indicating that the three-component force distribution at the tire's contact surface can be estimated with high accuracy by using the sensing member 60 according to this embodiment.

[0085] Next, with reference to Figure 15, the functional configuration of the tire state estimation device 10 according to this embodiment during the training of the tire state estimation model 13D will be described. Figure 15 is a block diagram showing an example of the functional configuration of the tire state estimation device 10 according to this embodiment during the training of the tire state estimation model 13D.

[0086] As shown in Figure 15, the tire state estimation device 10 includes a learning processing unit 11A during the training of the tire state estimation model 13D. The CPU 11 of the tire state estimation device 10 functions as the learning processing unit 11A by executing the learning processing program 13A.

[0087] The learning processing unit 11A according to this embodiment acquires training data to be used for training the tire state estimation model 13D.

[0088] In this embodiment, the learning processing unit 11A acquires learning data by reading learning data 13C from the storage unit 13. However, this is not the only configuration; the tire condition estimation device 10 may be connected to a network such as the Internet, and learning data may be acquired by reading it from an external device connected to the network.

[0089] The learning processing unit 11A according to this embodiment takes the electrical resistance value included in the acquired learning data as input information and the three-component force distribution corresponding to the said electrical resistance value included in the acquired learning data as output information to learn the tire state estimation model 13D.

[0090] Next, with reference to Figure 16, the functional configuration of the tire condition estimation device 10 according to this embodiment during the operation of the tire condition estimation model 13D will be described. Figure 16 is a block diagram showing an example of the functional configuration of the tire condition estimation device 10 according to this embodiment during the operation of the tire condition estimation model 13D.

[0091] As shown in Figure 16, the tire condition estimation device 10 during operation of the tire condition estimation model 13D includes an acquisition unit 11B and an estimation unit 11C. The CPU 11 of the tire condition estimation device 10 executes the tire condition estimation program 13B, thereby enabling the acquisition unit 11B and the estimation unit 11C to function.

[0092] The acquisition unit 11B according to this embodiment acquires an electrical signal by performing a load test on a tire 50, which is provided with a conductive sensing rubber 64 in a part of the interior region of the contact surface of the tread, capable of outputting an electrical signal corresponding to the deformation of the tread in at least a part of the said part of the region.

[0093] Then, the estimation unit 11C according to this embodiment takes the electrical signal acquired by the acquisition unit 11B, or the electrical resistance value obtained from said electrical signal (in this embodiment, the electrical resistance value), as input information, and uses a pre-learned tire state estimation model 13D to estimate the physical quantity related to the stress on the contact surface of the tire 50 (in this embodiment, the three-component force distribution) as output information.

[0094] As described above, the tire 50 according to this embodiment is a tire in which a sensing member 60, which has sensing rubber 64 laminated on a portion of it, is bonded to a portion of the tread at a certain area of ​​the contact surface of the tread, as a substitute for the tread in that portion.

[0095] Next, with reference to Figure 17, the operation of the tire state estimation device 10 when training the tire state estimation model 13D will be explained. Figure 17 is a flowchart showing an example of the learning process flow according to this embodiment.

[0096] The CPU 11 of the tire condition estimation device 10 executes the learning process 13A, which is how the learning process shown in Figure 17 is performed. The learning process shown in Figure 17 is performed when the user gives an instruction to start the execution of the learning process 13A via the input unit 14. To avoid confusion, the following description will focus on the case where a sufficient amount of learning data 13C for training the tire condition estimation model 13D has already been registered in the storage unit 13.

[0097] In step S100 of Figure 17, the CPU 11 acquires all the learning data 13C by reading it from the storage unit 13.

[0098] In step S102, the CPU 11 takes the electrical resistance values ​​(first resistance value to third resistance value in this embodiment) included in the acquired learning data as input information and the three-component force distribution corresponding to the electrical resistance values ​​included in the learning data as output information to learn the tire state estimation model 13D, and then terminates this learning process.

[0099] Through the above learning process, the tire state estimation model 13D will be trained using the training data 13C.

[0100] Next, with reference to Figure 18, the operation of the tire condition estimation device 10 when the tire condition estimation model 13D is in operation will be explained. Figure 18 is a flowchart showing an example of the flow of the tire condition estimation process according to this embodiment.

[0101] The tire condition estimation process shown in Figure 18 is executed when the CPU 11 of the tire condition estimation device 10 executes the tire condition estimation program 13B. The tire condition estimation process shown in Figure 18 is executed when the user gives an instruction to start the execution of the tire condition estimation program 13B via the input unit 14. To avoid confusion, the following explanation assumes that the training of the tire condition estimation model 13D has been completed, the tire condition estimation system 90 is in the state shown in Figure 1, and a load test is being performed on the tire 50.

[0102] In step S200 of Figure 18, the CPU 11 reads the tire condition estimation model 13D from the storage unit 13 and expands it into the memory 12, thereby obtaining the tire condition estimation model 13D.

[0103] In step S202, the CPU 11 acquires electrical resistance values ​​from the sensing element 60 of the tire 50 as input data via the communication I / F unit 18 in a time series.

[0104] In step S204, the CPU 11 uses the tire condition estimation model 13D acquired in step S200 to estimate the three-component force distribution corresponding to the input data acquired in step S202 as output data.

[0105] In step S206, the CPU 11 outputs information indicating the estimation result and terminates the tire state estimation process. In this embodiment, as output of the estimation result, information indicating the estimated three-component force distribution is displayed by the display unit 15 and stored in the storage unit 13, but it goes without saying that the embodiment is not limited to this form.

[0106] As described above, according to the tire state estimation device of this embodiment, an electrical signal is acquired by performing a load test on a tire that is equipped with conductive sensing rubber in a part of the interior of the contact surface of the tread, which is capable of outputting an electrical signal corresponding to the deformation of the tread in at least a part of that part of the interior. The acquired electrical signal, or the electrical resistance value obtained from the electrical signal, is used as input information, and a pre-learned tire state estimation model is used to estimate the physical quantity related to the stress at the contact surface of the tire as output information. This contributes to improving the measurement accuracy of the physical quantity related to the stress at the contact surface of the tire compared to when the sensor that measures the physical quantity is placed separately from the tire.

[0107] Furthermore, according to the tire condition estimation device of this embodiment, the tire condition estimation model is a model based on a recurrent neural network. Therefore, it is possible to estimate physical quantities related to stress at the tire contact surface by making more effective use of time-series information.

[0108] Furthermore, the tire state estimation device according to this embodiment uses a reservoir computing model instead of a regression neural network model. As a result, it is possible to reproduce the internal state of the tire during rolling, which changes over time, with high accuracy. Consequently, time-series information can be used more effectively to estimate physical quantities related to stress at the tire's contact surface.

[0109] Furthermore, according to the tire condition estimation device of this embodiment, the tire is a tire in which a sensing member, in which sensing rubber is laminated in part, is bonded to a part of the tread at a certain location on the contact surface of the tread, as a substitute for the tread in that part.Therefore, compared to the case in which sensing rubber is provided on the tire itself, it is possible to estimate physical quantities related to the stress on the contact surface of the tire at a lower cost.

[0110] Furthermore, according to the tire condition estimation device of this embodiment, the sensing member includes a tread rubber made of the same material as the tread, a sensing rubber laminated on the tread rubber for detecting the pressure generated inside the tire tread as an electrical signal, and a metal plate in contact with the end of the sensing rubber for extracting the electrical signal generated by the sensing rubber to the outside. Therefore, compared to technologies that measure physical quantities related to stress on the tire's contact surface separately from the tire, or that provide sensing rubber on the tire itself, this device can contribute to improving the accuracy of measuring physical quantities related to stress on the tire's contact surface.

[0111] In particular, according to the tire condition estimation device of this embodiment, the sensing member further includes a coating rubber for pressing a metal plate against the sensing rubber. As a result, the coating rubber can prevent the metal plate from peeling off, and thus the yield of the sensing member can be improved compared to the case where the coating rubber is not used.

[0112] Furthermore, according to the sensing member of this embodiment, the metal plate is made of brass. Therefore, the metal plate used as an electrode can be strongly bonded to the rubber, and peeling can be suppressed when distortion occurs.

[0113] In particular, according to the sensing member of this embodiment, the metal plate is composed of a positive electrode for voltage application and a plurality of electrodes, each in contact with different ends of the sensing rubber, for extracting electrical signals to the outside at the contact points. Therefore, compared to cases where there is only one or two electrodes for extracting electrical signals to the outside, it is possible to estimate physical quantities related to stress at the tire's contact surface with higher accuracy.

[0114] Furthermore, according to the sensing member of this embodiment, the coating rubber is made of zinc oxide. Therefore, the effect of preventing peeling between the sensing rubber and the brass plate can be further enhanced.

[0115] Furthermore, according to the sensing member of this embodiment, the sensing rubber has a carbon black content that differs from that of the tread rubber. As a result, the electrical resistance value of the sensing rubber can be made different from that of the tread rubber, making it easier to extract the electrical resistance value.

[0116] Furthermore, in this embodiment, the sensing member described above is used as a substitute for a portion of the tread located on the contact surface. Therefore, the tire can be manufactured at a lower cost compared to the case where sensing rubber is provided in the tire itself.

[0117] In particular, in this embodiment, the sensing member is bonded to the tire as a substitute for a portion of the tread located on the contact surface. Therefore, compared to the case where sensing rubber is provided on the tire itself, it is possible to estimate physical quantities related to the stress on the tire's contact surface more easily.

[0118] In the above embodiment, we described a case where only the electrical resistance value obtained by the sensing member 60 is applied as input information for the tire condition estimation model 13D, but the embodiment is not limited to this.

[0119] For example, in addition to the electrical resistance value mentioned above, physical quantities related to the tire 50 itself, such as the internal temperature and internal pressure of the tire 50, may be applied as input information to the tire state estimation model 13D. These physical quantities can be obtained by a TPMS (Tire Pressure Monitoring System).

[0120] Furthermore, physical quantities that affect the tire 50 of a vehicle equipped with the tire 50 (vehicle speed, steering angle, accelerator opening, brake amount, yaw rate, etc.) may be applied as input information to the tire state estimation model 13D, in addition to the electrical resistance value and physical quantities related to the electrical resistance value and the tire 50 itself.

[0121] These configurations allow for the estimation of physical quantities related to stress at the contact surface of the tire 50 with higher accuracy compared to the case where only electrical resistance values ​​are applied as input information for the tire condition estimation model 13D.

[0122] Alternatively, instead of the electrical resistance value, the electrical signal output from the sensing member 60 itself may be used as input information for the tire condition estimation model 13D.

[0123] Furthermore, although the above embodiment describes the case where the sensing member 60 is a rectangular parallelepiped, it is not limited to this form. For example, it may be a triangular prism, a prism with five or more sides, or a cylinder, an elliptical prism, or other shapes. Also, the dimensions of the sensing member 60 are not limited to those shown in Figure 11, and may be determined appropriately according to the type and dimensions of the tire, etc.

[0124] Furthermore, although the above embodiment described a case in which a reservoir computing model is applied as the tire state estimation model 13D, the invention is not limited to this form. For example, a CNN model other than a reservoir computing model may be applied as the tire state estimation model 13D.

[0125] Furthermore, although the above embodiment describes the case in which a three-component force distribution is applied as a physical quantity relating to the stress at the tire contact surface, the embodiment is not limited to this form. For example, the amount of strain at the tire contact surface may be applied as a physical quantity relating to the stress at the tire contact surface.

[0126] Furthermore, in the above embodiment, for example, the hardware structure of the processing unit that executes the learning processing unit 11A, the acquisition unit 11B, and the estimation unit 11C can be any of the following types of processors. As mentioned above, these types of processors include a CPU, which is a general-purpose processor that executes software (programs) and functions as a processing unit, as well as a programmable logic device (PLD), such as an FPGA (Field-Programmable Gate Array), which is a processor whose circuit configuration can be changed after manufacturing, and a dedicated electrical circuit, such as an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration specifically designed to execute a particular process.

[0127] The processing unit may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the processing unit may consist of a single processor.

[0128] Examples of configuring a processing unit with a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, as is common in client and server computers, and this processor functions as the processing unit. Secondly, a configuration using a processor that realizes the functions of the entire system, including the processing unit, on a single IC (Integrated Circuit) chip, as is common in System-on-a-Chip (SoC) systems. Thus, the processing unit is configured as a hardware structure using one or more of the above-mentioned types of processors.

[0129] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits, which are combinations of circuit elements such as semiconductor devices.

[0130] The disclosure of Japanese Patent Application No. 2024-224016, filed on 19 December 2024, is incorporated herein by reference in its entirety. Furthermore, all documents, patent applications, and technical standards described herein are incorporated herein by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually indicated as being incorporated by reference.

Claims

1. A tire state estimation device comprising: an acquisition unit that acquires an electrical signal by performing a load test on a tire provided with conductive sensing rubber in a certain region inside the contact surface of the tread, which is capable of outputting an electrical signal corresponding to the deformation of the tread in at least a part of the said region; and an estimation unit that uses the electrical signal acquired by the acquisition unit, or the electrical resistance value obtained by said electrical signal, as input information, and estimates the physical quantity using a pre-learned tire state estimation model with the physical quantity relating to the stress at the contact surface of the tire as output information.

2. The tire condition estimation device according to claim 1, wherein the tire condition estimation model is a model based on a regressive neural network.

3. The tire condition estimation device according to claim 2, wherein the model based on the recurrent neural network is a model based on reservoir computing.

4. The tire state estimation device according to any one of claims 1 to 3, wherein the tire is a tire in which a sensing member, in which the sensing rubber is laminated in part, is bonded to a part of the tread at a part of the contact surface of the tread, as a substitute for the tread in that part.

5. The tire condition estimation device according to claim 4, wherein the sensing member comprises: tread rubber made of the same material as the tread; sensing rubber laminated on the tread rubber; and a metal plate in contact with the end of the sensing rubber for extracting the electrical signal generated by the sensing rubber to the outside.

6. The tire condition estimation device according to claim 5, wherein the sensing member further includes a coating rubber for pressing the metal plate against the sensing rubber.

7. A tire state estimation method in which a computer performs a load test on a tire provided with conductive sensing rubber in a certain area inside the contact surface of the tread, which is capable of outputting an electrical signal corresponding to the deformation of the tread in at least a part of that area, thereby acquiring the electrical signal, and using the acquired electrical signal or the electrical resistance value obtained by said electrical signal as input information, the computer estimates the physical quantity related to the stress at the contact surface of the tire using a pre-learned tire state estimation model as output information.

8. A tire state estimation program that causes a computer to perform a load test on a tire equipped with conductive sensing rubber in a certain area inside the contact surface of the tread, which is capable of outputting an electrical signal corresponding to the deformation of the tread in at least a part of that area, thereby acquiring the electrical signal, and using the acquired electrical signal or the electrical resistance value obtained by said electrical signal as input information, and using a pre-learned tire state estimation model as output information, to estimate the physical quantity related to the stress at the contact surface of the tire.

Images