阅读说明：本技术 用于提取轮胎特性的变化的方法 (Method for extracting variations in tyre characteristics ) 是由 K·B·辛格 M·A·阿拉特 于 2020-08-28 设计创作，主要内容包括：提供了一种用于提取支撑车辆的轮胎的特性变化的方法。该方法包括：从安装在轮胎上的至少一个传感器提取所选择的轮胎特性。将所选择的轮胎特性传输到远程处理器并存储在与远程处理器通信的历史数据日志中。选择至少一个感兴趣的轮胎特性,并且将时间序列分解模型应用于来自历史数据日志的数据,以根据所选择的感兴趣的轮胎特性中的潜在趋势来描绘外来输入。将学习模型应用于所选择的感兴趣的轮胎特性中的潜在趋势,以对所选择的感兴趣的轮胎特性和轮胎的状况之间的关系进行建模。从学习模型输出轮胎的状况的预测值。(A method for extracting a change in a characteristic of a tire supporting a vehicle is provided. The method comprises the following steps: the selected tire characteristic is extracted from at least one sensor mounted on the tire. The selected tire characteristics are transmitted to the remote processor and stored in a historical data log in communication with the remote processor. At least one tire characteristic of interest is selected and a time series decomposition model is applied to data from the historical data log to delineate extraneous input according to potential trends in the selected tire characteristic of interest. A learning model is applied to the potential trends in the selected tire characteristic of interest to model a relationship between the selected tire characteristic of interest and a condition of the tire. The predicted value of the condition of the tire is output from the learning model.)

1. A method for extracting a variation in characteristics of tyres supporting a vehicle, characterized in that it comprises the steps of:

extracting selected tire characteristics from at least one sensor mounted on the tire;

transmitting the selected tire characteristic to a remote processor;

storing the selected tire characteristic in a historical data log in communication with the remote processor;

selecting at least one tire characteristic of interest;

applying a time series decomposition model to data from the historical data log to delineate extraneous inputs according to potential trends in the selected tire characteristic of interest;

applying a learning model to the potential trends in the selected tire characteristic of interest to model a relationship between the selected tire characteristic of interest and a condition of the tire; and

outputting a predicted value of the condition of the tire from the learning model.

2. A method for extracting a variation in a characteristic of a tyre supporting a vehicle as claimed in claim 1, wherein said step of extracting a tyre characteristic from at least one sensor mounted on said tyre comprises the steps of:

mounting a processor on at least one of the vehicle and the tire;

receiving tire characteristics from a sensor unit mounted on the tire in the processor mounted on at least one of the vehicle and the tire; and

executing a data analysis tool on the processor mounted on at least one of the vehicle and the tire to retrieve the selected characteristic.

3. A method for extracting variations in characteristics of tyres for supporting vehicles according to claim 1, characterized in that it further comprises the steps of: the selected tire characteristic is extracted from a sensor mounted on the vehicle.

4. A method for extracting variations in characteristics of tyres for supporting vehicles according to claim 1, characterized in that it further comprises the steps of: the selected tire characteristic is extracted from the vehicle CAN bus.

5. Method for extracting variations in characteristics of tyres for supporting vehicles according to claim 1, characterized in that it comprises the following steps: augmenting the historical data log with contextual information, the contextual information including at least one of weather, road irregularities, and road topology data.

6. The method for extracting a characteristic change of a tire of a supporting vehicle of claim 1, wherein the deep learning model comprises at least one of a machine learning model, a deep learning model, and a statistical model.

7. A method for extracting a characteristic variation of a tyre supporting a vehicle as claimed in claim 1, characterized in that the condition of the tyre comprises at least one of a state of wear of the tyre and a load of the tyre.

8. A method for extracting variations in characteristics of tyres for supporting vehicles according to claim 1, characterized in that it further comprises the steps of: transmitting the predicted value to at least one of a service center and a fleet manager.

9. Method for extracting variations in characteristics of tyres for supporting vehicles according to claim 1, characterized in that it comprises the following steps: transmitting the predicted value to a display device accessible to a user of the vehicle.

10. Method for extracting variations in characteristics of tyres for supporting vehicles according to claim 1, characterized in that it comprises the following steps: transmitting the predicted value to at least one of an electronic control unit and a vehicle control system of the vehicle.

Technical Field

The present invention generally relates to tire monitoring. More particularly, the present invention relates to systems and methods for sensing particular tire characteristics to predict or estimate certain conditions of a tire. In particular, the present invention is directed to a method for extracting changes in tire characteristics over the life of a tire to improve the accuracy of a system that estimates tire conditions.

Background

Tire estimation systems and methods are known in the art. Such systems and methods estimate tire conditions such as tire wear state and/or tire load. To perform the estimation, these systems and methods consider specific tire characteristics, such as tire inflation pressure, tire temperature, tread depth, road conditions, and the like.

In the prior art, direct measurement techniques are employed. Direct measurement techniques involve the use of sensors to try and directly measure characteristics such as tire inflation pressure, tire temperature, tread depth, and road conditions. For example, a pressure transducer disposed in a tire measures tire pressure, a thermocouple disposed in the tire measures tire temperature, a wear sensor disposed in the tire tread measures tread depth, and a vibration sensor or accelerometer measures road conditions. Data collected by such sensors has been transmitted to a processor having a memory to enable collection of the data. The collected data may then be sent to a display unit to display the measured tire characteristics, and/or to an electronic storage device for analysis and/or review.

Such direct measurement techniques may not enable tire conditions such as wear states or loads to be predicted in an accurate, reliable, or economical manner. To overcome such drawbacks, indirect estimation techniques have been developed.

Indirect techniques involve inputting measured tire characteristics (such as tire inflation pressure, tire temperature, tread depth, and road conditions) from sensor data into a statistical model stored on a processor. The model performs an analysis of the data to estimate or predict tire conditions, such as tire wear state and tire load.

While indirect estimation techniques have been successful in estimating or predicting tire conditions, they do not account for deviations or variances in certain tire characteristics that change over the life of the tire. Since such characteristics are not taken into account, the accuracy of the prior art indirect estimation techniques is undesirably reduced.

As a result, there is a need in the art for a method that improves the accuracy of an indirect estimation system that estimates tire conditions by taking into account and extracting changes in tire characteristics over the life of the tire.

Disclosure of Invention

According to an aspect of exemplary embodiments of the present invention, a method for extracting a characteristic variation of a tire supporting a vehicle is provided. The method comprises the following steps: the selected tire characteristic is extracted from at least one sensor mounted on the tire and transmitted to a remote processor. The selected tire characteristics are stored in a historical data log in communication with the remote processor. At least one tire characteristic of interest is selected. A time series decomposition model is applied to the data from the historical data log to delineate (delinteate) extraneous inputs according to potential trends in the selected tire characteristic of interest. A learning model is applied to the potential trends in the selected tire characteristic of interest to model a relationship between the selected tire characteristic of interest and a condition of the tire. The predicted value of the condition of the tire is output from the learning model.

Scheme 1. a method for extracting a variation in characteristics of a tyre supporting a vehicle, said method comprising the steps of:

extracting selected tire characteristics from at least one sensor mounted on the tire;

transmitting the selected tire characteristic to a remote processor;

storing the selected tire characteristic in a historical data log in communication with the remote processor;

selecting at least one tire characteristic of interest;

applying a time series decomposition model to data from the historical data log to delineate extraneous inputs according to potential trends in the selected tire characteristic of interest;

applying a learning model to the potential trends in the selected tire characteristic of interest to model a relationship between the selected tire characteristic of interest and a condition of the tire; and

outputting a predicted value of the condition of the tire from the learning model.

Scheme 2. the method for extracting a characteristic variation of a tyre supporting a vehicle according to scheme 1, wherein said step of extracting tyre characteristics from at least one sensor mounted on said tyre comprises the steps of:

mounting a processor on at least one of the vehicle and the tire;

receiving tire characteristics from a sensor unit mounted on the tire in the processor mounted on at least one of the vehicle and the tire; and

executing a data analysis tool on the processor mounted on at least one of the vehicle and the tire to retrieve the selected characteristic.

Scheme 3. the method for extracting a variation in characteristics of a tyre supporting a vehicle according to scheme 1, further comprising the steps of: the selected tire characteristic is extracted from a sensor mounted on the vehicle.

Scheme 4. the method for extracting a variation in characteristics of a tyre supporting a vehicle according to scheme 1, further comprising the steps of: the selected tire characteristic is extracted from the vehicle CAN bus.

Scheme 5. the method for extracting a variation in characteristics of a tyre supporting a vehicle according to scheme 1, further comprising the steps of: augmenting the historical data log with contextual information, the contextual information including at least one of weather, road irregularities, and road topology data.

Scheme 6. the method for extracting a characteristic change of a tire of a supporting vehicle according to scheme 1, wherein the deep learning model includes at least one of a machine learning model, a deep learning model, and a statistical model.

Scheme 7. the method for extracting a characteristic change of a tire supporting a vehicle according to scheme 1, wherein the condition of the tire includes at least one of a wear state of the tire and a load of the tire.

Scheme 8. the method for extracting a variation in characteristics of a tyre supporting a vehicle according to scheme 1, further comprising the steps of: transmitting the predicted value to at least one of a service center and a fleet manager.

Scheme 9. the method for extracting a variation in characteristics of a tyre supporting a vehicle according to scheme 1, further comprising the steps of: transmitting the predicted value to a display device accessible to a user of the vehicle.

Scheme 10. the method for extracting a variation in characteristics of a tyre supporting a vehicle according to scheme 1, further comprising the steps of: transmitting the predicted value to at least one of an electronic control unit and a vehicle control system of the vehicle.

Drawings

The invention will be described by way of example and with reference to the accompanying drawings, in which:

FIG. 1 is a schematic perspective view of a vehicle including a tire having a tire state of wear estimation system employing an exemplary embodiment of the method for extracting a change in a tire characteristic of the present invention;

FIG. 2 is a schematic diagram of a tire state of wear estimation system employing an exemplary embodiment of the method for extracting a change in a tire characteristic of the present invention;

FIG. 3 is a schematic diagram illustrating aspects of an analysis module of the tire wear state estimation system shown in FIG. 2;

FIG. 4 is a tabular representation of the change in characteristics over the life of a tire;

FIG. 5 is a graphical representation of the change in characteristics over the life of a tire; and

FIG. 6 is a schematic diagram illustrating exemplary steps of a method of the present invention for extracting a change in a tire characteristic.

Like numbers refer to like parts throughout the drawings.

Definition of

An "ANN" or "artificial neural network" is an adaptive tool for nonlinear statistical data modeling that changes its structure based on external or internal information flowing through the network during a learning phase. An ANN neural network is a non-linear statistical data modeling tool for modeling complex relationships between inputs and outputs or for finding patterns in data.

"axial" and "axially" mean lines or directions parallel to the axis of rotation of the tire.

"CAN bus" is an abbreviation for controller area network.

"circumferential" means a line or direction extending along the perimeter of the surface of the annular tread perpendicular to the axial direction.

"equatorial Center Plane (CP)" means a plane perpendicular to the axis of rotation of the tire and passing through the center of the tread.

"footprint" means the contact patch or contact area created by a tire tread having a flat surface as the tire rotates or rolls.

"inboard" means the side of the tire closest to the vehicle when the tire is mounted on the wheel and the wheel is mounted on the vehicle.

"lateral" means an axial direction.

"outboard" means the side of the tire furthest from the vehicle when the tire is mounted on the wheel and the wheel is mounted on the vehicle.

"radial" and "radially" mean directions radially toward or away from the axis of rotation of the tire.

By "rib" is meant a circumferentially extending rubber strip on the tread defined by at least one circumferential groove and either a second such groove or a lateral edge, the strip being laterally undivided by a full-depth groove.

"tread element" or "traction element" means a rib or block element defined by a shape having adjacent grooves.

Detailed Description

Referring to fig. 1-6, an exemplary embodiment of a method for extracting a change in a tire characteristic of the present invention is indicated at 100. To illustrate an exemplary environment in which the method 100 of the present invention is employed, a tire state of wear estimation system is indicated at 10 and will now be described. The Tire state of wear estimation system 10 is described in greater detail in an application entitled "Tire state of wear estimation System and method with footprint Length" filed concurrently with The present application by The Goodyear Tire & Rubber Company, The same assignee, and incorporated herein in its entirety.

With particular reference to FIG. 1, a tire wear state system 10 estimates tread wear on each tire 12 supporting a vehicle 14. It will be understood that the vehicle 14 may be any vehicle type and is shown by way of example as a passenger car. The tires 12 are of conventional construction and each is mounted on a respective wheel 16, as known to those skilled in the art. Each tire 12 includes a pair of sidewalls 18 (only one shown) extending to a circumferential tread 20 that wears over time due to road wear. The innerliner 22 is positioned on the inner surface of the tire 12 and, when the tire is mounted on the wheel 16, forms an interior cavity 24 that is filled with a pressurized fluid, such as air.

The sensor unit 26 is attached to the innerliner 22 of each tire 12 by means such as adhesive, and measures certain characteristics of the tire, such as tire pressure 38 (fig. 2) and temperature 40. For this reason, the sensor unit 26 preferably includes a pressure sensor and a temperature sensor, and may have any known configuration. The sensor unit 26 preferably also includes electronic memory capacity for storing Identification (ID) information for each tire 12, referred to as tire ID information and indicated at 42. The sensor unit 26 preferably also measures the centerline length 28 of the footprint of the tire 12.

Turning to FIG. 2, the sensor unit 26 includes a transmission device 34 for transmitting the measured characteristics of tire pressure 38, tire temperature 40, and centerline length 28, as well as tire ID information 42 to the processor 36. The transmission device 34 may comprise an antenna for wireless transmission or a wire for wired transmission. The processor 36 may be integrated into the sensor unit 26 or may be a remote processor that may be mounted on the vehicle 14 or cloud-based. The tire ID information 42 enables the tire configuration database 44 to be electronically accessed 46. The tire build database 44 stores tire build data 50 and is in electronic communication with the processor 36, thereby enabling transmission 48 of the tire build data to the processor.

The analysis module 52 is stored on the processor 36 and receives the tire pressure 38, the tire temperature 40, the tire centerline length 28, the tire ID information 42, and the tire build data 50. Analysis module 52 analyzes these inputs to generate an estimate of the tire wear state (indicated at 54). Referring to fig. 3, the analysis module 52 preferably also receives data from a vehicle-mounted collection unit 56, including vehicle speed 58 as calculated from Global Positioning System (GPS) data and inertial measurements 60 of the vehicle 14 from the accelerometer.

The event filter 62 is applied to data received from the collection unit 56 mounted on the vehicle. More specifically, vehicle conditions are reviewed in an event filter 62, including vehicle speed 58, as measured from GPS data, and inertial measurements 60. These measured values are compared to threshold values (including upper and lower limits). If the measured value is outside of the threshold, the system 10 does not proceed because the vehicle 14 is likely to be operating beyond normal or predictable conditions. If the measured values are within the threshold values, the measured data for tire pressure 38, tire temperature 40, centerline length 28, and vehicle speed 58 are sent to an anti-normalization filter 64.

The anti-normalization filter 64 is employed to account for and eliminate the effects of the inflation pressure 38, temperature 40, and vehicle speed 58 on the centerline length 28 of the tire 12. In the denormalization filter 64, a pre-trained regression model is used to account for the effects of charge pressure 38, temperature 40, and vehicle speed 58. Regardless of the operating conditions of the vehicle and tires, the centerline length 28 is returned to the predefined nominal conditions, namely the predefined inflation pressure 38, temperature 40, and vehicle speed 58.

In addition, the fastest wearing portion of the tire 12 may not always be at the centerline. For many tires, the fastest wear may be at the shoulder. However, the difference between the wear rate of the tire 12 at the centerline and at the shoulder typically depends on the tire build data 50, including the tire Footprint Shape Factor (FSF), mold design drop (mold design drop), tire belt/breaker angle, and/or overlay material. Thus, the tire build data 50 from the tire build database 44 is input into the anti-normalization filter 64 and used in conjunction with the centerline length measurement 28 from the sensor unit 26 to estimate the length at the shoulder, which may be the fastest wearing portion of the tread 20.

The de-normalization filter 64 generates a normalized footprint length 66. Since the centerline length 28 of the tire 12 may also be affected by vehicle loading, the effect of loading on the normalized footprint length 66 must be considered and eliminated. To eliminate the effect of loading on the normalized print length 66, a historical print measurement database 68 is accessed. A historical print measurement database 68 is in electronic communication with and storable on the processor 36 and contains a historical log of print measurements 70. The normalized footprint length 66 is related to the history log 70 and takes the average of the values.

The average of these values is applied to a time filter 72. The temporal filter 72 is a module that applies the steps of the method 100 for extracting a change in a tire characteristic, which will be described in more detail below. The time filter 72 produces a regularized footprint length 74 of the tire 12. The regularized footprint length 74 is input into a predictive model 76 that applies a non-linear regression model to generate the estimated wear state 54 of the tire 12.

Referring to fig. 4-6, a method 100 for extracting changes in tire characteristics improves the accuracy of indirect estimation systems, such as the wear estimation system 10, by taking into account and extracting changes in tire characteristics over the life of the tire 12. Tire characteristics (such as tire inflation pressure 38, tire temperature 40, tread depth, and road conditions) vary over the life of the tire 12. These characteristics are affected by more than one factor.

For example, as shown in fig. 4, testing confirms that the braking stiffness 102 of the tire 12, which is a measure of tire stiffness, varies with tire wear or tread depth (indicated at 104). The braking stiffness 102 of the tire 12 also varies with other operating conditions, such as ambient temperature 106, road surface conditions 108, and inflation pressure 38. Fig. 5 illustrates how the change in each of these characteristics over time 112 affects the braking stiffness 102. In addition to tire wear or tread depth 104, ambient temperature 106, road surface conditions 108, and inflation pressure 38 affecting the braking stiffness 102 of the tire 12, random variations 114 over time 112 may also affect the braking stiffness.

Such changes in the characteristics of the tire 12 over time 112 make the predictive model susceptible to bias and/or variance. The method 100 for extracting changes in tire characteristics models such changes in tire characteristics after a time scale by using an additive time series. As described above, the tire characteristics include tire inflation pressure 38, tire temperature 40, ambient temperature 106, tread depth 104, road surface condition 108, and the like. For convenience, reference will be made to tire characteristics 120, with the understanding that such reference includes such characteristics.

Referring specifically to fig. 6, method 100 includes a step 150 of extracting relevant tire characteristics from sensors on tire 12 and/or vehicle 14. For example, a processor 116 may be mounted on the vehicle 14 or the tire 12 that receives certain tire characteristics 120, such as the inflation pressure 38 and the tire temperature 40, from the sensor unit 26 (FIG. 1). The processor 116 may also receive characteristics 120 such as ambient temperature 106 and road surface condition 108 from tire mounted sensors, vehicle mounted sensors, and/or a vehicle CAN bus. To perform the extraction, the processor 116 executes a data analysis tool that retrieves the selected characteristics 120 from the sensor(s) and/or the CAN bus.

The processor 116 includes or is in electronic communication with an antenna 118 that provides for transmission of the selected characteristic 120 to a remote processor (step 152), such as a processor in the cloud-based server 122. The cloud-based server 122 includes or is in communication with a historical data log 124 of the extracted tire characteristics 120. Storage of the characteristics 120 in the historical data log 124 is provided at step 154. Step 154 optionally includes augmenting the historical data log 124 with contextual information (contextual information), such as weather, road irregularities, and road topology data.

Next, in step 156, one or more tire characteristics of interest 120 are selected and the time series decomposition model 126 is applied to the data from the historical data log 124. The time series decomposition model 126 delineates or separates extraneous input or data from potential trends in the selected tire characteristic(s) 120 of interest.

The learning model 128 is then applied to the selected potential trends in the tire characteristic(s) 120 of interest (step 158). The learning model 128 may be a machine learning model, a deep learning model, or a statistical model, and models a relationship between the tire characteristic(s) 120 of interest and a condition of the tire 12 to be predicted, such as a tire wear state or a tire load. The learning model 128 outputs the predicted value 130 at step 160. The predicted value 130 is a value that has been filtered to eliminate deviations due to factors affecting the tire 12 over time, and is therefore a value of increased accuracy.

In the example of the wear state estimation system 10 shown in fig. 4 and 5, the temporal filter 72 performs steps 154 and 156 to output the regularized footprint length 74. The regularized footprint length 74 is input into the predictive model 76 to perform steps 158 and 160 to generate the predicted value 130, which is the estimated wear state 54 of the tire 12.

Returning to fig. 6, depending on how the method 100 for extracting a change in tire characteristics is employed, the predicted values 130 may be transmitted in step 162 to one or more destinations via the antenna 132. For example, the predicted values 130 may be transmitted to a service center or fleet manager 134. Additionally, or alternatively, the predicted value 130 may be transmitted to a display device accessible to a user 136 of the vehicle 14, such as a smartphone. Also, or as another alternative, the predicted value 130 may be transmitted to an electronic control unit 138 of the vehicle 14 and/or a vehicle control system (such as a braking system and/or a suspension system) to improve the performance of such a system.

In this manner, the method 100 for extracting changes in tire characteristics improves the accuracy of the indirect estimation system by taking into account and extracting changes in tire characteristics over the life of the tire 12.

It will be understood that the steps of the above-described method 100 for extracting a change in tire characteristic and accompanying structures may be altered or rearranged, or components or steps known to those skilled in the art may be omitted or added, without affecting the overall concept or operation of the present invention. For example, electronic communication may be through a wired connection or wireless communication without affecting the overall concept or operation of the invention. Such wireless communications include Radio Frequency (RF) and Bluetooth @ communications. Additionally, tire characteristics and tire conditions other than those described above and known to those skilled in the art may be employed without affecting the overall concept or operation of the invention.

The invention has been described with reference to the preferred embodiments. Modifications and alterations will occur to others upon a reading and understanding of this specification. It is to be understood that all such modifications and variations are included within the scope of the invention as set forth in the following claims or the equivalents thereof.