System and method for vehicle tire performance modeling and feedback

Abstract

Systems and methods for vehicle tire performance modeling and feedback are provided. A computer-implemented system and method for vehicle tire performance monitoring, modeling, and feedback is disclosed. Vehicle data, including movement data and / or location data, is collected for a vehicle and / or at least one tire associated with the vehicle. A current tire wear state of the at least one tire is determined in real-time based at least in part on the collected data, various embodiments of which are disclosed herein. One or more tire performance characteristics, such as tire traction, replacement time, and / or future tire wear, are predicted based at least in part on the determined tire wear state and the collected data. Real-time feedback is selectively provided based on the predicted one or more tire performance characteristics and / or the determined current tire wear state. The feedback can include an alert to an operator or fleet manager, or can take the form of an automated control invention.

Description

[0001] This application is a divisional application of invention application filed on March 30, 2020, with application number 2020800311911 and entitled "System and Method for Modeling and Feedback of Vehicle Tire Performance". Technical Field

[0002] This invention relates generally to modeling and predicting tire performance and providing feedback based on that modeling and prediction. More specifically, embodiments of the invention disclosed herein relate to systems and methods for implementing tire wear and / or tire traction models for wheeled vehicles, including but not limited to motorcycles, consumer vehicles (e.g., buses and light trucks), commercial vehicles, and off-road (OTR) vehicles. Background Technology

[0003] Predicting tire wear and corresponding tire traction is an important tool for anyone who owns or operates a vehicle, especially in the context of vehicle fleet management. As tires are used, the tread typically becomes thinner, and overall tire performance changes.

[0004] In addition, irregular tread wear can occur for a variety of reasons, which may cause users to replace tires earlier than originally intended. Vehicles, drivers, and individual tires are all different from each other, and this can cause tires to wear at very different rates. For example, high-performance tires used in sports cars wear out faster than tires used in family cars. However, a wide variety of factors can cause tires to wear out earlier than expected, and / or cause irregular tire wear and produce noise or vibration. Two common causes of premature and / or irregular tire wear are improper inflation pressure and out-of-specification alignment conditions.

[0005] Tire wear is known to develop in a non-linear manner throughout the tire's lifespan. A major reason for this is that the tread blocks become harder as the tread wears down over time. Additionally, tread patterns are typically designed to have a smaller void area as the tire wears. Any or both of these characteristics contribute to a slower wear rate.

[0006] Most tire wear predictions focus on initial wear rate, that is, the wear rate when the tire is brand new. This is at least in part because the tire industry generally focuses on new tire performance due to the need to meet original equipment manufacturer (OEM) requirements. To predict the tire's performance over its entire lifespan, new wear models are required.

[0007] However, tire wear is a complex phenomenon to model. Accurate models using finite element analysis (FEA) exist, but these simulations typically take weeks to complete. If it is desired to simulate wear rates at several different tread depths, this would further require computationally expensive simulations lasting several months.

[0008] The goal is to provide users with essentially real-time predictions about the performance and capabilities of their tires.

[0009] It is also expected that the tire's traction capacity will be estimated, and such feedback will be provided as input to models used for other useful / effective prediction or control loops.

[0010] It is also desirable to estimate the tire tread depth and provide such feedback as input to models for other useful / effective predictions, such as traction, fuel efficiency, durability, etc. Accurate tread depth prediction is the first step in predicting many other areas of tire performance.

[0011] It is also expected that these services will be provided as part of a distributed and relatively automated tire-as-a-service model, without the need for manual tread depth measurements (such as those typically provided by field engineers and / or with specialized equipment).

[0012] It is known that high-frequency vehicle and / or tire data is generated to determine the vehicle's condition at a given time. However, continuously collecting streaming data results in excessively large data point volumes, which is generally impractical from a data transmission, storage, and processing perspective. There is also a desire to improve the state of knowledge based on tread depth measurements, thereby providing real-time feedback to users (e.g., individual drivers, fleet managers, other equivalent end-users) based on the ability to predict wear life remaining in the tire based on several periodic measurements, and thus enabling users to extract maximum value from the tires. Summary of the Invention

[0013] In the first exemplary embodiment disclosed herein, the foregoing objective can be achieved via a computer-implemented method for modeling and predicting tire performance and providing feedback based on such modeling and prediction. The method includes: collecting vehicle data of a vehicle and / or tire data of at least one tire associated with the vehicle; and determining, in real time, the current tire wear condition of the at least one tire, at least in part based on the collected data. One or more tire performance characteristics are predicted, at least in part based on the determined tire wear condition and the collected data. Real-time feedback is selectively provided based on the predicted one or more tire performance characteristics and / or the determined current tire wear condition.

[0014] Additional advantageous features are also implemented in an exemplary variant of the aforementioned first embodiment, wherein a second embodiment of a computer-implemented method for estimating tire wear condition is disclosed herein, and includes accumulating information in a data storage device about the probability distribution corresponding to each of a plurality of tire wear factors. Vehicle data and / or tire data, including movement data and location data, collected in association with the vehicle, are transmitted from the vehicle to a remote server. At least one observation corresponding to one or more of the plurality of factors is generated based on the transmitted vehicle data. A Bayesian estimate of the tire wear condition of at least one tire associated with the vehicle at a given time is generated, based at least on the generated at least one observation and the stored information about the probability distribution.

[0015] An exemplary aspect of the aforementioned second embodiment may include storing information on updated probability distributions corresponding to a plurality of factors that contribute to tire wear of at least one tire associated with the vehicle, based at least on at least one of the generated observations.

[0016] Another exemplary aspect of the aforementioned second embodiment may include predicting tire wear condition based on one or more future parameters of at least one tire associated with the vehicle. For example, tire wear condition may be predicted relative to an upcoming period of vehicle travel or relative to an upcoming distance traveled.

[0017] Another exemplary aspect of the aforementioned second embodiment may include the replacement time of at least one tire associated with the vehicle based on the current tire wear condition or the predicted tire wear condition relative to a tire wear threshold associated with at least one tire associated with the vehicle.

[0018] In another exemplary aspect of the aforementioned second embodiment, information about multiple probability distributions can be reflected in an array of time series characterization curves.

[0019] Another exemplary aspect of the aforementioned second embodiment may include: receiving one or more tire wear input values ​​from a user via a user interface associated with a remote server; and generating at least one observation of one or more factors among a plurality of factors based on the one or more tire wear input values.

[0020] Another exemplary aspect of the aforementioned second embodiment may include: receiving one or more tire wear input values ​​generated by one or more sensors mounted in or on a respective tire in at least one tire; and generating at least one observation of one or more factors among a plurality of factors based on the one or more tire wear input values.

[0021] Another exemplary aspect of the aforementioned second embodiment may include: receiving one or more tire wear input values ​​generated by sensors located outside the vehicle; and generating at least one observation of one or more factors among a plurality of factors based on the one or more tire wear input values.

[0022] In another exemplary aspect of the aforementioned second embodiment, at least one of the tire wear input values ​​generated by sensors located outside the vehicle includes a tread depth measurement result.

[0023] Another exemplary aspect of the aforementioned second embodiment may include generating an estimated tire wear condition using a baseline value and a range corresponding to the estimated confidence level.

[0024] A system for estimating tire wear condition can be provided according to the second embodiment described above. The system may include a data storage network having information stored thereon about probability distributions corresponding to each of a plurality of tire wear factors. For each of a plurality of vehicles, distributed computing nodes are linked to one or more onboard sensors respectively configured to collect vehicle data. A server-based computing network is provided, comprising a computer-readable medium having instructions residing thereon, and the instructions being executable by one or more processors to direct the performance of the aspects previously described with respect to the second embodiment.

[0025] Additional advantageous features are also implemented in another exemplary variant of the first embodiment described above, wherein a third embodiment of a computer-implemented method for analyzing tire wear models using a brush-type model is disclosed herein. The brush-type model is a simplified tire model that models tread elements as individual “bristles,” which greatly reduces the complexity of modeling the contact interface between the road and the rubber. This model can capture the first-order effects (tread block hardening and contact area increase) that occur in a real tire during actual tire wear.

[0026] According to a third exemplary embodiment, an initial tread depth of the tires associated with the vehicle is determined, and an initial tire wear rate is determined based at least in part on this initial tread depth. One or more tire conditions are measured as a time-series input to a predictive tire wear model. The current wear rate is normalized based on this input regarding the initial tire wear rate, wherein the tire wear state can be predicted for one or more specified future parameters.

[0027] In one aspect of the aforementioned third embodiment, the current wear rate is determined at least in part based on a brush-type tire wear model for the contact interface between the tire's matrix material and the road surface, wherein the interface is represented as a plurality of independently deformable elements.

[0028] In another aspect of the aforementioned third embodiment, the measured one or more tire conditions include the detected contact area and void area corresponding to the tire tread depth.

[0029] In another aspect of the aforementioned third implementation scheme, one or more specified future parameters are associated with travel time.

[0030] Alternatively, one or more specified future parameters may be associated with the distance traveled.

[0031] In another aspect of the aforementioned third embodiment, the tire replacement time can be predicted based on the predicted tire wear condition relative to one or more predetermined tire wear thresholds associated with the tire.

[0032] In another aspect of the aforementioned third implementation scheme, an alert is generated for the user associated with the vehicle based on the predicted replacement time.

[0033] In another aspect of the aforementioned third embodiment, one or more measurements are received from the user via a user interface.

[0034] In another aspect of the aforementioned third embodiment, one or more of the measured conditions are generated and received from one or more sensors installed in or on the tire.

[0035] In another aspect of the aforementioned third embodiment, conditions are generated and received from one or more measurements by sensors located outside the vehicle. At least one of the tire wear input values ​​generated by the sensors outside the vehicle may include a tread depth measurement result.

[0036] In another aspect of the aforementioned third embodiment, the system can predict tire rotation threshold events and / or alignment threshold events, at least in part, based on time-series inputs and / or predicted tire wear conditions. Therefore, alerts can be generated for a user interface associated with the vehicle. The user interface can be a static display installed in the vehicle, a display for a mobile computing device associated with the driver of the vehicle, etc.

[0037] In another aspect of the aforementioned third embodiment, the optimal tire type for the vehicle can be predicted, at least in part, based on time-series input and / or the predicted tire wear condition. Therefore, an alert can be generated for a user interface associated with the vehicle. The user interface can be a static display installed in the vehicle, a display for a mobile computing device associated with the driver of the vehicle, etc.

[0038] In one embodiment, a system for predicting the progress of vehicle tire wear according to the aforementioned third embodiment is provided, the system including a server functionally linked to a data storage network. The data storage network includes raw tread depths of tires associated with the vehicle and a predictive tire wear model. One or more sensors are provided, and the sensors are configured to provide signals corresponding to measured tire conditions. The server is configured to determine the initial wear rate of the tire based on the raw tread depth and the tire wear model, collect the signals corresponding to the measured tire conditions as time-series inputs to the predictive tire wear model, normalize the current wear rate to the initial wear rate of the tire based on this input, and predict the tire wear state for one or more specified future parameters.

[0039] In one exemplary aspect of the system according to the third embodiment, a brush-type tire wear model for the contact interface between the tire's matrix material and the road surface can be used to model the wear rate, wherein the interface is represented as a plurality of independently deformable elements. Alternative physics-based tire wear models may also be implemented within the scope of this disclosure, including but not limited to the FEA model.

[0040] Additional advantageous features are also implemented in an exemplary variant of the first embodiment described above, wherein a fourth embodiment of a computer-implemented method for estimating the progress of vehicle tire wear is disclosed herein. The method according to the fourth embodiment includes storing the tread depth of a tire associated with a vehicle in a first (e.g., initial or unworn) phase. The method also includes sensing and storing a first set of one or more modal frequencies of the tire in the first phase in response to effects associated with a first modal analysis. In a subsequent second (e.g., at least partially worn) phase, in response to effects associated with a second modal analysis, sensing a second set of one or more corresponding modal frequencies of the tire. The tire wear state can be estimated in the second phase based on a calculated frequency shift between at least one corresponding modal frequency from each of the first and second sets.

[0041] In one exemplary aspect of the aforementioned fourth embodiment, the tire mass is stored in a first stage, wherein the step of estimating the tire wear state in a second stage includes determining the change in tire mass between the first and second stages based on the calculated frequency shift.

[0042] In another exemplary aspect of the aforementioned fourth embodiment, the estimated loss of tire tread is determined relative to the change in tire mass between the first and second stages based on the calculated frequency shift. Alternatively, the estimated loss of tire tread can be determined via a searchable correlation between the observed frequency shift and the change in tire tread of a given tire. The correlation can be retrieved from a data storage device, for example, for a given tire type, or the correlation can be established over time based on historical measurements of tire tread changes and offsets between corresponding modal frequencies associated with a given tire type.

[0043] In another exemplary aspect of the aforementioned fourth embodiment, in response to the excitation of a tire's structural mode, a first set of corresponding modal frequencies and a second set of corresponding modal frequencies are sensed via one or more accelerometers associated with the tire. The one or more accelerometers may be attached to the tire, for example, attached to the tire's inner liner, or mounted to the main shaft of the associated vehicle.

[0044] In another exemplary aspect of the aforementioned fourth embodiment, tire structural patterns are randomly excited during tire operation, and associated output signals generated by one or more accelerometers are captured.

[0045] In another exemplary aspect of the aforementioned fourth embodiment, a tire structural mode is induced by a controlled impact on the tire by an external object (such as, for example, a hammer).

[0046] In another exemplary aspect of the aforementioned fourth embodiment, tire structure patterns are stimulated by guiding the movement of the vehicle relative to one or more predetermined obstacles (such as, for example, anti-skid wedges or speed bumps) or a road with a sufficiently rough surface.

[0047] An exemplary system according to the fourth embodiment disclosed herein can achieve vehicle tire wear estimation via a server or server network functionally linked to a data storage network and a server for one or more sensors mounted on the tires and / or the vehicle, for example, according to any or more of the foregoing embodiments and aspects thereof.

[0048] Additional advantageous features are also implemented in exemplary variations of the first embodiment described above, wherein a fifth embodiment of a computer-implemented method for estimating vehicle tire wear is disclosed herein. The method of the fifth embodiment includes one or more sensors associated with a vehicle and / or at least one of a plurality of tires supporting the vehicle generating first data corresponding to the real-time dynamics of the vehicle and / or at least one tire. The first data is locally processed to generate second data as a reduced subset of the first data, wherein the second data represents the first data and includes any one or more predetermined features extracted therefrom. The second data is selectively transmitted via a communication network to a remote computing system, and the remote computing system processes the second data and any one or more extracted features to estimate the wear characteristics of at least one tire.

[0049] The second data may include multiple sequential data frames, each including a multidimensional histogram of forces associated with the vehicle and / or at least one tire.

[0050] In one exemplary aspect of the aforementioned fifth embodiment, the method further includes: selecting a subset of data frames at least between the first event and the second event; and summarizing the data frames within a specific time or a specific distance.

[0051] In another exemplary aspect of this fifth embodiment, the aggregation of data frames is performed via local processing before the aggregated data frames are transmitted to a remote computing system. Alternatively, a subset of data frames may be transmitted to the remote computing system, wherein the aggregation of data frames is performed via the remote computing system.

[0052] In another exemplary aspect of the fifth embodiment, the method further includes correcting missing data in the aggregated data frames by scaling the aggregated data frames to the expected number of data frames relative to the actual number of data frames collected.

[0053] The features extracted from the second data may include wear performance characteristics that represent vehicle driving behavior.

[0054] Processing the first data may include performing a Fourier transform on the first data and generating second data including the extracted relevant frequencies and associated amplitudes.

[0055] In another exemplary aspect of the fifth implementation, the second data includes aggregated low-frequency CAN data corresponding to the amount of time spent by the vehicle in each of one or more representative driving conditions.

[0056] In another aspect of the fifth embodiment, the first data includes a CAN bus signal. The second data is generated via an encoding neural network layer, the third data is generated via a decoding neural network layer, and a wear calculation layer is attached to the output of the decoding neural network layer and configured to transform the decoded CAN bus signal into an instantaneous estimated wear value for at least one tire.

[0057] In one exemplary aspect of the aforementioned fifth embodiment, the method further includes: comparing an estimated wear value and an actual wear value of at least one tire to generate an error value; and providing the error value as feedback to a neural network layer.

[0058] In another aspect of the fifth embodiment, the selective transmission of the second data is automated and event-based, rather than relying on manual selection of transmission. Alternatively, the selective transmission of the second data can be time-based.

[0059] In another aspect of the fifth embodiment, a method for estimating vehicle tire wear is implemented using one or more sensors associated with at least one of a plurality of tires supporting the vehicle, wherein first data corresponding to the real-time dynamics of the vehicle and / or at least one tire is generated. Low-frequency second data corresponding to the vehicle's position is generated via a Global Positioning System transceiver. The second data is selectively transmitted via a communication network to a remote computing system, wherein the second data is processed according to a vehicle model and one or more vehicle route characteristics to generate third data corresponding to the first data, and the third data is further processed to estimate the wear characteristics of at least one tire.

[0060] In one exemplary aspect of the aforementioned fifth embodiment, the second data further includes a plurality of sequential data frames, each data frame comprising a multidimensional histogram of forces associated with the vehicle and / or at least one tire telecomputing system. The telecomputing system reconstructs the vehicle route from the collected vehicle location data and provides vehicle route feedback to the corresponding multidimensional histogram.

[0061] Additional advantageous features are also implemented in an exemplary variant of the first embodiment described above, wherein a sixth embodiment of a computer-implemented method for estimating vehicle tire wear is disclosed herein. First data is generated via one or more sensors associated with the vehicle and / or at least one of a plurality of tires supporting the vehicle, the first data corresponding to the real-time dynamics of the vehicle and / or at least one tire. The first data is processed via a computing system on the vehicle to generate second data as a simplified subset of the first data, the second data representing the first data and including any one or more predetermined features extracted therefrom. The on-board computing system further processes the second data to estimate the wear characteristics of at least one tire and generates a notification associated with the estimated wear characteristics to a computing device associated with a vehicle user.

[0062] In one exemplary aspect of the aforementioned sixth embodiment, the step of processing the second data to estimate the wear characteristics of at least one tire includes: processing the second data to generate third data corresponding to the first data; and further processing the third data to estimate the wear characteristics of at least one tire.

[0063] In another exemplary aspect of the aforementioned sixth embodiment, the first data includes a CAN bus signal, the second data is generated via an encoding neural network layer, the third data is generated via a decoding neural network layer, and a wear calculation layer is attached to the output of the decoding neural network layer and configured to transform the decoded CAN bus signal into an instantaneous estimated wear value for at least one tire.

[0064] Another exemplary aspect of the aforementioned sixth embodiment also includes: comparing the estimated wear value and the actual wear value of at least one tire to generate an error value; and providing the error value as feedback to a neural network layer.

[0065] Additional advantageous features are further implemented in exemplary variations of any of the first to sixth embodiments described above, wherein a seventh embodiment of a computer-implemented method for estimating and applying vehicle tire traction states is disclosed herein. The method according to the seventh embodiment may include: collecting vehicle data associated with a first vehicle (e.g., including movement data and location data); and determining the tire wear state of at least one tire associated with the vehicle. One or more tire traction characteristics of at least one tire are predicted based at least on the transmitted vehicle data and the determined tire wear state, and one or more vehicle operating settings are selectively modified based at least on the predicted one or more tire traction characteristics.

[0066] In one exemplary aspect of the seventh embodiment described above, the maximum speed of the vehicle is determined at least based on the transmitted vehicle data and the determined tire wear condition of each tire associated with the vehicle.

[0067] In another exemplary aspect of the seventh embodiment described above, the maximum speed is provided to an autonomous vehicle control system associated with the vehicle. Alternatively, the maximum speed may be provided to a driver assistance interface associated with the vehicle.

[0068] In another exemplary aspect of the seventh embodiment described above, one or more tire wear input values ​​are received from the user via a user interface.

[0069] In another exemplary aspect of the seventh embodiment described above, the step of determining the tire wear condition includes receiving one or more tire wear input values ​​generated by one or more sensors mounted in or on the respective tire of at least one tire. Alternatively, the one or more tire wear input values ​​may be generated by sensors located outside the vehicle.

[0070] In another exemplary aspect of the seventh embodiment described above, the step of determining the tire wear condition includes predicting one or more tire wear input values ​​based at least on transmitted vehicle data and tire data generated by one or more sensors installed in or on the corresponding tire in at least one tire.

[0071] A system may be provided for performing the method according to the seventh embodiment described above, and optionally also according to certain exemplary aspects, the system comprising a remote server functionally linked to a vehicle via a communication network, wherein vehicle data is transmitted from the vehicle to the remote server. The remote server is configured to provide one or more predicted tire traction characteristics to an active safety unit associated with the vehicle, and the active safety unit is configured to modify one or more vehicle operating settings based at least on the predicted one or more tire traction characteristics.

[0072] In one exemplary aspect of the system according to the seventh embodiment, the active safety unit may include an automated braking system associated with the vehicle, and a remote server is configured to provide the automated braking system with one or more parameters of a predicted μ-slip profile associated with the respective tire.

[0073] In another exemplary aspect of the system according to the seventh embodiment, the user interface is associated with a remote server and configured to receive one or more tire wear input values ​​from the user.

[0074] In another exemplary aspect of the system according to the seventh embodiment, the remote server is configured to determine the maximum speed of the vehicle based at least on the transmitted vehicle data and the determined tire wear state of each tire associated with the vehicle, and to provide the maximum speed to the driver assistance interface associated with the vehicle.

[0075] In other exemplary aspects of the system according to the seventh embodiment, the active safety unit may include a collision avoidance system and / or an autonomous vehicle control system.

[0076] Another example of the system can execute a method according to the seventh embodiment described above for each of a plurality of vehicles, and optionally also according to certain exemplary aspects associated therewith. The system includes a first remote server functionally linked to the vehicles via a communication network, a fleet management server functionally linked to the first remote server, and a vehicle control system associated with each of the plurality of vehicles. For each of the plurality of vehicles, vehicle data is transmitted from the respective vehicle to the remote server, the first remote server being configured to provide one or more predicted tire traction characteristics to the fleet management server, and the fleet management server being configured to interact with the respective vehicle control system to modify one or more vehicle operating settings at least based on the predicted one or more tire traction characteristics.

[0077] In one exemplary aspect of the system, the user interface is associated with a remote server and / or a fleet management server and / or a vehicle control system, and is configured to receive one or more tire wear input values ​​from the user.

[0078] In another exemplary aspect of the system, the fleet management server is configured to determine the maximum speed of a given vehicle based at least on the transmitted vehicle data and the determined tire wear state of each tire associated with the respective vehicle, and to provide the maximum speed to the vehicle control system associated with the vehicle.

[0079] In another exemplary aspect of the system, the fleet management server is configured to calculate the parking distance potential of a given vehicle based at least on the transmitted vehicle data and the determined tire wear state of each tire associated with the vehicle, and to provide the parking distance potential to the vehicle control system associated with the vehicle.

[0080] In another exemplary aspect of the system, the fleet management server is also configured to determine the optimal driving distance for each of a plurality of vehicles associated with a sequentially moving vehicle queue, and to transmit the determined optimal driving distance for each of the plurality of vehicles to the respective vehicle control system.

[0081] In another exemplary aspect of the system, the fleet management server is configured to determine, at least based on the transmitted vehicle data and the determined tire wear state of each tire associated with the respective vehicle, the maximum speed and / or stopping distance potential of a given vehicle, determine whether the vehicle meets threshold traction characteristics, and if the vehicle does not meet the threshold traction characteristics, interact with the vehicle control system to prevent deployment or otherwise exempt the respective vehicle from use.

[0082] The various implementation schemes described above can be readily combined with each other in the systems and / or methods disclosed herein.

[0083] For example, a technician may understand that the predicted tire wear according to the third or fourth embodiment can be provided as the output of the traction model according to the seventh embodiment, complementing each other without altering the range of the corresponding steps or features.

[0084] Furthermore, those skilled in the art will understand that the data extracted according to the fifth embodiment can be provided as input to a tire wear model according to one or more other embodiments disclosed herein. Attached Figure Description

[0085] In the following, embodiments of the invention are shown in more detail with reference to the accompanying drawings.

[0086] Figure 1 This is a block diagram illustrating exemplary embodiments of a system according to the various implementations disclosed herein.

[0087] Figure 2 This is a block diagram representing an exemplary traction estimation model.

[0088] Figure 3 It is a graphical representation of a set of exemplary traction characteristics generated by the model disclosed herein.

[0089] Figure 4 This is a graphical representation of another set of exemplary traction characteristics generated by the model disclosed herein.

[0090] Figure 5 It is a graphical representation of a set of exemplary tire wear (e.g., tread) status values ​​for an autonomous vehicle fleet.

[0091] Figure 6 It is a graphical representation of an exemplary application of tire wear (e.g., tread) status values ​​and predicted tire traction values ​​for a truck platoon.

[0092] Figure 7 This is a graphical representation of the effect of signal resolution on wear rate estimation when using signals collected during a mixed scenario of urban and highway vehicle routes.

[0093] Figure 8 This is a graphical representation of the impact of signal resolution on wear rate estimation when using signals collected primarily during urban routes.

[0094] Figure 9 This is a graphical representation of an exemplary method for aggregating and compressing vehicle dynamics data into histogram data frames.

[0095] Figure 10 It means according to Figure 9 A graphical representation of an exemplary bar chart data frame of the process.

[0096] Figure 11 This is a graphical representation of an exemplary process used for summarizing data in a bar chart.

[0097] Figure 12 It is a graphical representation of an exemplary process used for scaling bar chart data frames to correct for missing or incomplete data.

[0098] Figure 13 This is a graphical representation of an exemplary tire wear modeling flow.

[0099] Figure 14 It means according to Figure 13 A graphical representation of an exemplary real-time model integration for tire wear modeling flow.

[0100] Figure 15 This is a graphical representation of an exemplary neural network autoencoder application for tire wear.

[0101] Figure 16A It means according to Figure 15 A graphical representation of exemplary results from the compression and decompression of example x-axis acceleration data by a neural network autoencoder.

[0102] Figure 16B It means according to Figure 15 The example is a graphical representation of the exemplary results of compression and decompression of y-axis acceleration data by a neural network autoencoder.

[0103] Figure 16C It means according to Figure 15 The following is a graphical representation of exemplary results from the compression and decompression of vehicle speed data by a neural network autoencoder.

[0104] Figure 17 This is a block diagram representing a traditional method used for tire wear analysis (e.g., using vehicle alignment data).

[0105] Figure 18 This is a block diagram representing an exemplary Bayesian method for tire wear estimation.

[0106] Figure 19 This is a graphical representation of an exemplary tire wear model correction.

[0107] Figure 20 This is a graphical representation of an exemplary application of Monte Carlo simulation to establish a set of toe angle distributions.

[0108] Figure 21 This is a graphical representation of an exemplary application of Monte Carlo simulation to establish a set of outward tilt angle distributions.

[0109] Figure 22 It is a graphical representation of a set of exemplary wear progress curves for the front tire.

[0110] Figure 23 It is a graphical representation of a set of exemplary wear progress curves for the rear tire.

[0111] Figure 24 This is a graphical representation of an exemplary brush model used for wear output.

[0112] Figure 25 This is a graphical representation of an exemplary tire wear model prediction compared to measured data.

[0113] Figure 26 It is a graph representing the difference between the tire wear model predictions disclosed in this paper and the results of various indoor wear tests on the same control tire.

[0114] Figure 27 This is a graphical representation of exemplary results from a static natural frequency test of a given tire under new and wear conditions.

[0115] Figure 28A This is a graphical representation of exemplary results from a simulation of anti-skid wedge impacts from a tire in a new state.

[0116] Figure 28B This is a graphical representation of exemplary results from a simulation of anti-skid wedge impacts from a tire in a worn state.

[0117] Figure 29 This is a graphical representation of exemplary results from a transferability test for a tire in both its new and worn states. Detailed Implementation

[0118] General Reference Figures 1 to 29 Various exemplary embodiments of the invention will now be described in detail. Where various figures may depict embodiments that share various common elements and features with other embodiments, similar elements and features are given the same reference numerals, and their redundant descriptions may be omitted below.

[0119] Various implementations of the system disclosed herein may include a centralized computing node (e.g., a cloud server) that functionally communicates with multiple distributed data collectors and computing nodes (e.g., associated with individual vehicles) to efficiently implement the wear and traction model disclosed herein. Initial Reference Figure 1An exemplary embodiment of system 100 includes a computing device 102 located in a vehicle and configured to at least acquire data and transfer that data to a remote server 130 and / or perform relevant calculations as disclosed herein. The computing device may be portable or otherwise modular as part of a distributed vehicle data collection and control system (as shown), or may be provided integrally with respect to a central vehicle data collection and control system (not shown). The device may include a processor 104 and a memory 106 on which program logic 108 resides. Generally, systems as disclosed herein can implement numerous components distributed across one or more vehicles, for example, but not necessarily associated with a fleet management entity, and may also implement a central server or network of servers that communicates functionally with each vehicle via a communication network. Vehicle components typically include one or more sensors, such as vehicle accelerometers, gyroscopes, inertial measurement units (IMUs), position sensors (such as GPS transponders 112), tire pressure monitoring system (TPMS) sensor transmitters 118, and associated onboard receivers, which are linked to, for example, a controller local area network (CAN) bus network and thereby provide signals to a local processing unit. For illustrative purposes and without further limiting the scope of the invention, the illustrated embodiment includes an ambient temperature sensor 116, an engine sensor 114 configured to, for example, provide a sensed air pressure signal, and a DC power supply 110.

[0120] Based on the following discussion, other sensors used for collecting and transmitting vehicle data such as speed, acceleration, and braking characteristics will become sufficiently obvious to those skilled in the art and will not be discussed further herein. Various bus interfaces, protocols, and associated networks are well known in the art for transferring vehicle dynamics data between appropriate data sources and local computing devices, and those skilled in the art will recognize a wide range of such tools and apparatuses for implementing them.

[0121] The system may include additional distributed program logic (such as residing, for example, on a fleet management server or other computing device 140), or a user interface of a device residing in the vehicle or associated with its driver (not shown) for real-time notifications (e.g., via visual and / or audio indicators), wherein the fleet management device is linked to the onboard device via a communication network function in some embodiments. System programming information may be provided, for example, by the driver in the vehicle or from the fleet manager.

[0122] In implementations, vehicle and tire sensors are also provided with unique identifiers, wherein the on-board device processor 104 can distinguish between signals provided from corresponding sensors on the same vehicle, and further, in some implementations, wherein the central server 130 and / or fleet maintenance monitor client device 140 can distinguish between signals provided from tires and associated vehicle and / or tire sensors on multiple vehicles. In other words, in various implementations, sensor output values ​​may be associated with a specific tire, a specific vehicle, and / or a specific tire-vehicle system for the purpose of on-board or remote / downstream data storage and for specific implementations of calculations disclosed herein. The on-board device processor may communicate directly with a hosting server, such as... Figure 1 As shown, or alternatively, the driver's mobile device or computing device installed on the truck may be configured to receive data output from the onboard equipment and process / transmit it to a hosting server and / or a fleet management server / device.

[0123] Signals received from specific vehicle and / or tire sensors may be stored in the on-board device memory or in an equivalent data storage unit functionally linked to the on-board device processor for selective retrieval as needed for calculations according to the methods disclosed herein. In some embodiments, raw data signals from various signals may be transmitted from the vehicle to the server substantially in real time. Alternatively, particularly considering the inherent inefficiencies in continuous data transmission of high-frequency data, the data may be, for example, compiled, encoded, and / or aggregated for more efficient (e.g., based on periodic time or alternatively based on defined events) transmission from the vehicle to a remote server via a suitable communication network.

[0124] Once vehicle and / or tire data is transmitted to hosting server 130 via a communication network, it can be stored, for example, in a database 132 associated with it. The server may include a tire wear model and a tire traction model 134, or otherwise associate with these two models, for selectively retrieving and processing vehicle and / or tire data as appropriate input. The models may be implemented at least in part via an execution processor, thereby enabling selective retrieval of vehicle and / or tire data, and also enabling electronic communication to input any additional data or algorithms from databases, lookup tables, etc., stored in association with the server.

[0125] In one embodiment of the method disclosed herein, the system 100 described above can be implemented for modeling and predicting tire performance and providing feedback based on such modeling and prediction. The method may include collecting vehicle data, including movement data and / or location data, of a vehicle and / or at least one tire associated with the vehicle, and determining, in real time, the current tire wear condition of the at least one tire, at least in part based on the collected data. One or more tire performance characteristics are predicted, at least in part based on the determined tire wear condition and the collected data. Real-time feedback is selectively provided based on the predicted one or more tire performance characteristics and / or the determined current tire wear condition. In the various embodiments disclosed herein, some or all of these steps may be extended as discussed below to provide additional advantages.

[0126] For example, refer to the following Figure 2 The implementation scheme of the system and method disclosed herein enables a simplified tire model 134B, together with the tire wear condition 150, to predict the tire's traction capacity 160, which is relayed to the user to facilitate safe driving. The simplified model predicts the forces and moments on the tire under given friction, load, inflation pressure, speed, etc. For illustrative purposes, the terms "tire wear" and "tread wear" are used interchangeably herein.

[0127] For the traction model 134B to be accurate, especially under wet conditions, the tread depth 150 must be known / estimated. This can be achieved by any of the following exemplary techniques.

[0128] In one implementation, tire wear (tread) measurement 150 can be performed manually by a user and is provided as user input to an application or equivalent interface associated with the onboard computing device 102 or directly associated with the hosting server 130. This interface may, for example, allow the user to directly input wear values ​​for a selected tire from a plurality of tires mounted on an identified vehicle. Alternatively, the interface may be configured to prompt the user with a captured image associated with the tread profile or alternative input, wherein the wear value can be determined indirectly through user input.

[0129] In another embodiment, tire wear measurement 150 can be performed by sensors mounted on the tire and provided to a hosting server without input from, for example, a user. Such sensors can be mounted, for example, directly in the tire tread or on the tire liner.

[0130] In another embodiment, tire wear measurement 150 may be provided via one or more sensors external to the vehicle and then sent back to cloud server 130 without, for example, input from a user. As an example, the one or more sensors may include a driving optical sensor comprising: a laser emitter configured to capture tire tread information by projecting a laser beam onto or onto the surface of the tire passing through the sensor; and one or more laser receiving elements configured to capture reflected energy and thereby obtain the profile of the tire from which the tire tread can be determined.

[0131] In another implementation, such as, for example, Figure 2 The wear model 134A can be used to estimate tire wear values ​​150 in various implementations, as illustrated and exemplified herein. The wear model may include a “digital twin” virtual representation of various physical components, processes, or systems, where digital and physical data are paired and combined with a learning system such as, for example, a neural network. For instance, real-world data 136 from the vehicle and associated location / route information may be provided to generate a digital representation of the vehicle tires for tire wear estimation, where subsequent comparisons of the estimated tire wear with the determined actual tire wear may be implemented as feedback to a machine learning algorithm. The wear model 134A may be implemented at the vehicle for processing via an onboard system 102, or tire data 138 and / or vehicle data 136 may be processed to provide representative data to a hosting server 130 for remote wear estimation.

[0132] like Figure 2 The tire wear condition (e.g., tread depth) 150 shown may be provided, for example, along with certain vehicle data 136 as input to a traction model 134B, which may be configured to provide an estimated traction condition 160 or one or more traction characteristics 160 for the corresponding tire. Similar to the aforementioned wear model, the traction model may include a “digital twin” virtual representation of a physical part, process, or system, wherein digital and physical data are paired and combined with a learning system such as, for example, an artificial neural network. Real vehicle data 136 and / or tire data 138 from a specific tire, vehicle, or tire-vehicle system may be provided throughout the lifecycle of the corresponding asset to generate a virtual representation of the vehicle tire for estimating tire traction, wherein subsequent comparisons of the estimated tire traction with the corresponding measured or determined actual tire traction may preferably be implemented as feedback to a machine learning algorithm executed at the server level.

[0133] In various implementations, the traction model 134B can utilize an associated combination of values ​​of previous tests (including, for example, stopping distance test results, tire traction test results, etc.) collected for many tire-vehicle systems and input parameters (e.g., tire tread, inflation pressure, road surface characteristics, vehicle speed and acceleration, slip rate and angle, normal force, braking pressure and load), wherein tire traction output can be effectively predicted for a given set of current vehicle data and tire data inputs.

[0134] In one implementation, the output 160 from the traction model 134B can be incorporated into an active safety system. As previously described, data is collected from sensors on the vehicle to feed into a tire wear model 134A, which predicts the tread depth 150, and this is then fed into the traction model 134B. The term "active safety system" as used herein preferably encompasses systems commonly known to those skilled in the art, including but not limited to examples such as collision avoidance systems, advanced driver assistance systems (ADAS), anti-lock braking systems (ABS), etc., which can be configured to utilize the traction model output information 160 to achieve optimal performance. For example, collision avoidance systems are typically configured to take evasive action, such as automatically engaging the brakes of the primary vehicle to avoid or mitigate a potential collision with a target vehicle, and enhanced information regarding tire traction and, consequently, the braking capability of the tire-vehicle system is highly desirable.

[0135] For reference Figure 3 and Figure 4 The exemplary model shown includes two curves for each graph representing the same hypothetical tire at different wear levels. As can be seen, wet traction performance deteriorates accordingly with tire wear. During inclement weather, there is a critical speed at which worn tires reach a point where the user is at risk of slipping. Using a traction model remotely linked to an in-vehicle display or equivalent user interface, the maximum speed can be transmitted to the user to provide safer driving conditions.

[0136] Traction output information determined based on the corresponding wear condition, such as, for example, μ-slip curves (see, for example...) Figure 4 This can also be fed into active safety systems for vehicle control implementation and thereby performance optimization. Slip ratio is expressed as ((vehicle speed – tire rotation speed) / vehicle speed), where 0% slip ratio corresponds to a freely rolling tire, and 100% slip ratio corresponds to a locked wheel. When the tire μ-slip curve is shaped like... Figure 4As the curve changes from a "new tire" profile to a "worn tire" profile over time, as represented, the active safety system can preferably be configured to determine what changes (if any) can be made to improve tire-vehicle performance characteristics. Different μ-slip profiles can be considered to have relevant shape and positional characteristics that affect the ability of active safety systems (e.g., ABS) to optimize performance, where, for example, the corresponding peak amplitude "μ" is generally understood to affect stopping distance (the higher the better). Other relevant characteristics of the μ-slip profile shape may include, for example, the slip ratio at the y-axis (μ) peak of the curve, the curvature at or near that peak, the initial slope of the curve, etc.

[0137] In another implementation, the ride-sharing autonomous fleet can use output data 160 from the traction model 134B to disable or otherwise selectively exempt vehicles with low tread depth during inclement weather, or potentially limit the maximum speed of such vehicles. (See reference...) Figure 3 The illustrated exemplary model shows that, compared to tires in a "new" state that can exceed 100 mph without their peak coefficient of friction dropping below a threshold of 0.25, tires in a "worn" state are identified as having a slippage critical speed of approximately 55 mph, at which point the peak coefficient of friction drops below the same threshold. Therefore, the system can limit the speed of vehicles including those with one or more tires worn to this state. If the vehicle is part of a carpooling autonomous fleet and the user is seeking to travel along routes requiring a minimum (e.g., highway) speed during inclement weather conditions, the system can be configured to disable the deployment of vehicles with tread depths below a certain level or otherwise insufficient traction. Figure 5 As shown, an exemplary autonomous vehicle fleet may include a number of vehicles with different minimum tread state values, wherein the fleet management system may be configured to disable the deployment of vehicles that fall below a minimum threshold. The system may be configured to operate based on the minimum tread value of each of a plurality of tires associated with a vehicle, or, in one embodiment, the aggregated tread state of the plurality of tires may be calculated for comparison with a minimum threshold.

[0138] In another embodiment, the fleet management system can implement output data 160 from the traction model 134B for a defined vehicle platoon, such as to better optimize its spacing by better understanding the stopping distance potential of each tire to achieve maximum fuel savings. Those skilled in the art will understand that minimizing spacing results in reduced aerodynamic drag on all vehicles in the platoon, and thereby improves the corresponding fuel economy, particularly when the platoon includes more than two trucks, and the disclosed improvements to the vehicle platooning method can advantageously contribute to reducing spacing beyond more conventional "one-size-fits-all" methods. Most fuel savings are typically achieved at spacings of less than about 20 meters, a distance that may be difficult or impossible to maintain using conventional techniques for determining traction / braking capacity during adverse weather conditions. By more effectively determining safe spacing, the percentage of time spent platooning can be increased, even in adverse weather conditions.

[0139] In one implementation, active safety or platooning spacing information may be provided to the vehicle braking control system or vehicle platooning driving control system 120 associated with each corresponding vehicle. In the context of a vehicle platoon, an individual vehicle associated with the platoon may receive spacing information and / or certain vehicle control information, and transmit this information to other vehicles in the platoon via other conventional vehicle-to-vehicle communication systems and protocols. The spacing information provided by the system as disclosed herein can be considered, for example, a nominal or minimum effective spacing setting based on the corresponding traction state of the vehicles in the platoon. It should be understood that the vehicle platooning driving control system for a given vehicle or vehicle platoon may further modify the spacing setting based on monitored traffic events, road conditions, and other environmental conditions that may be outside the range determined by the traction state in a given implementation. For example, an acceptable first spacing for a given vehicle under normal driving conditions may necessarily increase based on monitored real-time events such as changes in the grade of the road to be traversed, or an increased risk of braking events for any one or more vehicles in the platoon.

[0140] The components of the vehicle platooning driving control system 120 are generally known in the art and may include, for example, a vehicle braking control system, a collision mitigation system, vehicle-to-vehicle communication, and one or more sensors configured to monitor vehicle data such as the current driving distance of the main vehicle (relative to another vehicle in the platoon or non-platoon target vehicle), the appropriate type of the target vehicle, the relative acceleration or deceleration value of the main vehicle, the pressure value of the brake actuator relative to the main vehicle, etc.

[0141] As previously mentioned, various implementations of the method can estimate tire wear values ​​150 based on wear model 134A. Current wear models require several inputs to the system to accurately predict tire wear life and are developed using very high-frequency data. However, transferring high-frequency data from distributed data collectors (e.g., associated with individual vehicles) to centralized computing nodes (e.g., cloud servers) is prohibitively expensive on scale.

[0142] Reference for illustrative purposes Figure 7 and Figure 8 The data presented illustrates the impact of signal resolution on wear rate estimation. To construct these plots, the source data was downsampled to reduce its resolution. The downsampling in these examples was performed by simply sampling the source data. The source data has a resolution of one meter per sample in the distance domain, or approximately 20 Hz at a speed of 45 mph. The x-axis represents the range from one meter per sample to one kilometer per sample. The y-axis shows the relative error in wear estimation.

[0143] In these two graphs, the datasets correspond to all four tires of a front-wheel-drive Toyota Camry equipped with Turanza EL400 all-season tires. Figure 7 In this context, the data represents the "average North American driver" across a mix of urban and highway roads, where lower predicted wear rates typically correspond to lower wear prediction accuracy. Figure 8 The data represents a fleet of urban taxis, with the vast majority of mileage occurring in urban driving environments, and requires a significantly higher sampling rate compared to previous datasets.

[0144] The results shown in the figure demonstrate that simple downsampling of data is not a reliable, robust, and efficient method for reducing data storage and transmission requirements. The minimum resolution required to achieve good predictions depends heavily on the route traveled (e.g., predominantly urban, or a mix of urban and highway driving) and driving style. Furthermore, the required minimum resolution also depends on the tire position on the vehicle (e.g., front left, front right, etc.).

[0145] Therefore, those skilled in the art will understand that more sophisticated strategies are expected to maximize the efficiency of vehicle dynamics data storage and transmission for tire wear estimation.

[0146] The exemplary tire wear model 134A disclosed herein can aggregate data from high-frequency sources or alternatively low-frequency sources into low-frequency data, such as route data, which can be transmitted to the cloud at that lower frequency in a cost-effective manner, thereby enabling direct wear modeling. In some embodiments, adaptive solutions can be utilized to achieve improved efficiency, for example by encoding wear estimation features into a compressed / reduced dataset, making the method more robust and adaptive to field conditions.

[0147] In one implementation, real-time vehicle dynamics data can be collected from sensors on the vehicle and then filtered and downsampled into aggregated buckets to create a histogram of relevant forces. For example, raw accelerometer data can be downsampled and aggregated into a histogram representing the raw data at a coarse level.

[0148] For example in Figure 9 As shown, real-time vehicle dynamics data 310 can be compiled into windows 320 of time and / or distance. The compiled data can also be aggregated into a histogram data frame 330. In the illustrated embodiment, the data frame 330 is multi-dimensional and contains vehicle acceleration and vehicle velocity. Each point in the histogram represents the time or distance spent under that condition. The bins of the histogram can be optimized to maximize wear calculation accuracy and further minimize data storage and transmission costs, thereby enabling, for example, a simple, equally spaced or non-linear bin layout.

[0149] Figure 10 An example of a histogram data frame is shown, which has a first dimension associated with lateral vehicle acceleration and a second dimension associated with front-to-rear vehicle acceleration. The individual points in this example are color-coded to represent the time or distance spent in the corresponding condition.

[0150] Next reference Figure 11 Since wear is a cumulative process, it is useful to summarize data between specific events in terms of time and / or distance. Examples of relevant events may include, but are not limited to: vehicle trips, tire tread depth measurement events, tire rotation events, tire installation events, vehicle maintenance events, daily / monthly / yearly summaries, mileage summaries (5k, 10k, 20k miles, etc.). The histogram data frame 330 allows for flexible and efficient summarization, which can be used on static data in the cloud (after transmission) or transient data on the vehicle (before transmission).

[0151] Unfortunately, data from vehicle and communication systems is often, or even inherently, unreliable. Those skilled in the art will understand that, in the event of missing or corrupted data, it is desirable to design software systems that are predictable and robust. Since wear is a cumulative process, missing data poses a problem for wear calculations. The histogram data frame 330 disclosed according to this embodiment allows for effective compensation for missing data.

[0152] Next reference Figure 12 Multiple histogram data frames 330 containing missing data subsets can be aggregated to generate partial data frames 430, which can be further corrected by scaling the data frames relative to the number of data frames collected by a desired number of data frames. The result (corrected data frames 440) will be the average of the driver's behavior.

[0153] As previously stated, and now further referenced as follows Figure 13 The represented tire wear modeling flow and such Figure 14 The illustrated exemplary real-time model integration uses one or more sensors on or associated with the vehicle to acquire vehicle dynamics sequence data 710. A real-time vehicle-tire model 720 can then be used to simulate tire forces on each tire. Furthermore, the tire model can be used to generate wear rate simulations 730. Both models can be implemented in real-time on time / distance sequence data or on aggregated data frames. The simulation results can be stored or transmitted in data frame format.

[0154] Figure 14 The example illustrates a real-time simulation of tire forces and the transmission of tire force data frame 830. The scope of this embodiment is not necessarily limited thereto, and those skilled in the art will understand alternative strategies for various use cases.

[0155] It should be noted that while many embodiments disclosed herein are based on vehicle dynamics data to simulate forces on each tire, the scope of the invention is not limited thereto unless otherwise specifically stated. In other words, providing raw data corresponding to one or more forces applied to at least one tire is within the scope of the invention if such data is available in a given application.

[0156] In another embodiment of the method disclosed herein, vehicle dynamics data may be filtered, downsampled, and aggregated into a subset representing behavior or “driver severity” values ​​that indicate how the vehicle is driven. These values ​​are extracted from the raw data to specifically capture predetermined wear performance characteristics of the driver’s behavior. The extracted behavioral features are further processed by a downstream (e.g., host server-based) wear model. According to other embodiments disclosed herein, behavioral values ​​extracted from the raw data as features before being transmitted to the cloud may optionally supplement or otherwise complement other forms of aggregated or compressed data.

[0157] In another implementation, low-frequency GPS data from the vehicle can be transmitted to a cloud server, where routes are reconstructed using a reverse mapping algorithm and fed into a time-series histogram to understand the time spent under various driving conditions (highway, turning, braking, etc.). Similar to the aforementioned implementations, according to other implementations disclosed herein, vehicle location data collected or extracted prior to transmission to the cloud may optionally supplement or otherwise enhance other forms of aggregated or compressed data.

[0158] In another implementation, low-frequency CAN data can be aggregated to count the time spent under various driving conditions used to calculate wear status. Similar to the first two implementations, feature extraction in the form of event-based driving detection prior to transmission to the cloud may optionally supplement or otherwise complement other forms of aggregated or compressed data, according to one or more other implementations disclosed herein.

[0159] In another implementation scheme, further reference is now made. Figure 15 The neural network autoencoder 900 can be implemented to transform and compress the input CAN bus signal 910 in the first (i.e., encoder) layer 920, and further reconstruct the data in the second (i.e., decoder) layer 940 after transmitting the compressed data to the cloud for use by a tire wear model to predict tire performance. As further illustrated in the three graphical diagrams, the first vehicle acceleration data stream (such as...) Figure 16A The x-axis acceleration shown), and the second vehicle acceleration data stream (such as...) Figure 16B The y-axis acceleration and vehicle speed data streams shown (e.g.) Figure 16C (As shown) can be compressed and reconstructed to its corresponding original signal with very high accuracy. In each figure, the original and reconstructed data are overlaid to highlight this accuracy.

[0160] Neural network autoencoders 900 are well known in the art for reducing data dimensionality and typically comprise multiple pairs of layers. An input layer 910 has a first size, which decreases via encoding layer 920 and subsequent layers until an intermediate layer 930 is reached. Thereafter, the layer size increases via decoding 940 until an output layer 950 has the first size. Exemplary uses of the autoencoder disclosed herein may differ from conventional arrangements because it also includes a dedicated third (i.e., wear estimation) layer 960, which is designed and attached to the second layer 950. The dedicated third layer 960 is configured to perform wear rate calculations to transform the raw CAN bus signal into an instantaneous (actual) wear rate 970. For example, the wear-specific layer may include proprietary formulas containing vehicle and tire-specific information related to the physical system. Since the raw vehicle dynamics data signal can be reconstructed with very high accuracy via the first and second layers of the neural network, the attached third (wear-specific) layer can similarly be highly accurate.

[0161] The third layer 960 also enables the first (encoding) layer 920 and the second (decoding) layer 940 to be specifically trained over time for wear estimation. During training, the encoding and decoding layers learn to capture and store the most basic information for wear calculation. For example, the estimated instantaneous wear rate or predicted wear rate can be compared with the actual wear rate to generate a model error value 980. The feedback loop 990 provides the model error value back to the autoencoder to update the model weights and biases in the first layer 920 and / or the second layer 940. The third layer 960 will propagate through weights specific to the estimated or predicted tire wear.

[0162] In other words, attaching the third layer 960 to the end of a conventional automatic encoder (i.e., after the second layer 940) allows the neural network to learn how to optimally transform the representation of the CAN bus signal to be used for predicting tire wear, whereas the conventional automatic encoder would simply learn the optimal representation for the direct feedback of the original signal. This improved encoding layer, learned over time via, for example, the aforementioned feedback system, enables the decoding layer to encode the data in a way that produces the optimal signal for estimating or predicting tire wear.

[0163] This network architecture enables the network to learn the most physically important signal characteristics and patterns (peaks, valleys, cross signal relationships, etc.) about wear and efficiently propagate those characteristics through the network.

[0164] In another implementation, the system can be configured to perform a Fourier transform on the raw data stream and extract the most relevant frequencies. These frequencies and their associated amplitudes can then be used to reconstruct the complete original data state after transmission to the cloud.

[0165] Further reference Figures 17 to 23Another exemplary embodiment disclosed herein involves the use of Bayesian methods in the characterization and prediction of tire wear. The basis of this method is to represent factors contributing to wear, such as driving style, vehicle alignment settings, route, road surface, environmental conditions, tire manufacturing variations, etc., as probability distributions. The fundamental principle behind representing these as probability distributions is that the variation observed in each of these factors is not noise, but a true representation of the natural variations observed for wear. For example, the same tire used by an aggressive first driver (who accelerates and brakes forcefully) will experience a very different tire wear lifespan compared to a more careful second driver. When used with conventional prediction models, the average representation of the two drivers would produce predictions that are insufficient when applied individually to either situation.

[0166] The effect of this probabilistic representation of contributing factors is that predictions made by wear algorithms will also be probabilistic; that is, predictions are also distributions. Using distributions when reporting predictions has several benefits. First, predictions can provide a measure of uncertainty, i.e., tread wear of 4.1 mm + / - 0.05 mm, or a wear prediction of 55,000 miles + / - 3,000 miles (these ranges can correspond to specific confidence levels, such as 95% or 98%). Second, these distributions can be updated using Bayesian inference based on observations. Such observations can be, for example, for predictor variables (e.g., measurements of tread depth) or input variables (acceleration characterizing driving style). The value of this inference can be because a model or associated system, as further described below, can continue to update predictions over time with respect to, for example, a specified distance traveled or the time spent traveling with the associated tires, and the confidence level of such predictions.

[0167] refer to Figure 18 The schematic diagram in the figure illustrates an exemplary process flow that can be connected with, for example, Figure 17 The traditional method shown is illustrated by comparison. A probability distribution of factors related to tire wear (such as vehicle wheels and suspension settings) can be generated and fed into the vehicle model, which is the opposite of a specific target value or measurement for the same factor. An example of such a factor is camber angle, also known as the angle from the road surface normal passing through the center of the corresponding wheel (on which the tire is mounted) to the centerline of the wheel. Another example of such a factor is toe angle, also known as the angle of the tire relative to the longitudinal axis of the corresponding vehicle.

[0168] From these initial ranges, additional probability distributions can be generated for each of several relevant forces (e.g., traction or longitudinal force Fx, lateral force Fy, vertical or normal force Fz) and / or moments (e.g., rollover torque My, alignment torque Mz) on the associated tire, which are also the opposite of the individual values ​​of the same force. These force distributions can be fed into a tire wear model, where the tread depth is estimated based on a baseline value (e.g., 5.8 mm) over a calculated uncertainty range (e.g., + / - 0.3 mm) for a given distance traveled (e.g., 15,000 km), which is the opposite of a baseline value alone.

[0169] Subsequently, the probability distribution of tread depth can be updated based on the observations, as illustrated in the diagram above. This update can be implemented using the representation of Bayes' theorem, which is shown here:

[0170]

[0171] Bayesian filtering methods are known in the art for determining the probability of a given measurement based on, for example, all previous corresponding measurements in a sensor data stream. Here, the term "model" refers to the parameters of the model, and the term "observation" refers to a measurement performed on any / all variables involved in the model. According to the above formula, information related to tire wear prediction can be updated over time using actual measurements. In other words, using this method, model predictions can be "corrected" with each measurement performed on a particular tire element and / or vehicle tire system. For example, such measurements can be implemented to reduce uncertainty and achieve better predictions over time if tread depth measurements are periodically collected and transmitted or otherwise compiled for application according to the systems and methods disclosed herein.

[0172] Next reference Figure 19 This method can periodically provide tire wear prediction corrections along with tread depth measurements, thereby reducing the uncertainty of wear predictions. When tread depth (or equivalent tire wear-related factors) measurements are collected over time, potential alternative models or time series curves can be effectively eliminated or minimized for a given tire, vehicle driver-tire system, etc., and subsequent tire wear estimates can be provided more accurately with less uncertainty in the results.

[0173] As shown in the figure, the wear prediction curve progresses from a first point (along the y-axis) with an overall wear prediction uncertainty U0. Following a subsequent tread depth measurement, a corrected wear prediction curve is generated, along with an uncertainty U1 representing a reduced level of wear prediction. In this example, the second envelope of uncertainty U1 falls entirely within the first envelope. After another tread depth measurement, a third and further corrected wear prediction curve is generated, along with an uncertainty U2 representing a further reduced level of wear prediction.

[0174] Next reference Figure 20 and Figure 21 This illustrates an exemplary application of the Monte Carlo method to construct probability distributions and use these distributions to generate wear progress curve distributions (see, for example, ...). Figure 22 The wear progress curve of the exemplary front tire shown and as... Figure 23 (Example rear tire wear progression curve shown). In other words, for a given change in vehicle alignment settings, this method attempts to determine the corresponding change in wear progression. In the particular case shown, it is assumed that the input is a normal distribution independent only with respect to toe angle and camber angle, where all other factors are referred to as single points. Although toe angle and camber angle have been chosen for illustration herein, it should be understood that, unless otherwise specifically indicated, alternative or additional vehicle and / or tire settings may be applied to the tire wear model and thus implemented by the systems and methods disclosed herein.

[0175] Especially for reference Figure 23 The rear tire wear progress curves are shown in the diagram, with the center curve representing the nominal toe / camber setting, and the surrounding area representing ten thousand individual wear progress curves corresponding to the respective initial wear rate Ew. It can be observed that the change in wear progress increases with mileage. By implementing periodic measurements of the values ​​of the fundamental factors, appropriate subsets of the individual wear progress curves can be identified with increasing certainty over time, allowing for accurate prediction of tire wear conditions using only a relatively small number of actual measurements.

[0176] Therefore, even periodic measurements of tread depth or other relevant factors provide real-time feedback to users (e.g., fleet managers, end users) and enhance the ability to predict the remaining wear life of tires and further maximize the remaining value of tires.

[0177] Periodic measurements used to supplement the probability distribution associated with tire wear (e.g., tire tread depth) can be performed directly (by the user manually and / or via one or more sensors) and / or estimated based on tire wear models and techniques as further described herein.

[0178] Further reference Figures 24 to 26Another exemplary embodiment of the method disclosed herein involves the use of brush pattern analysis in characterizing and predicting tire wear. A brush pattern model is a simplified tire model with a logical-physical background that models tread elements as independent “bristles” extending outward from the tire’s base material (e.g., carcass). The brush pattern model significantly reduces the complexity of modeling the contact interface between the road surface and the base material, where the modeled tread elements can deform in a variety of measurable directions (e.g., longitudinal, lateral, vertical) and can capture the first-order effects (tread block hardening and increased contact area) that occur in a real tire during actual tire wear. In alternative embodiments, the characterization and prediction of tire wear can be achieved using other physics-based tire wear models, such as, for example, finite element analysis (FEA).

[0179] One embodiment of the method disclosed herein also advantageously predicts the absolute wear rate of a tire under given conditions, rather than just predicting how the wear rate changes as the tread depth decreases. This is achieved, at least in part, by normalizing the currently modeled wear rate (e.g., based on periodically or otherwise updated measurements) relative to the wear rate at the original tread depth (i.e., the initial wear rate).

[0180] See, for example Figure 24 The graph in the image shows an exemplary output of the model, where the normalized wear ratio of two different tires is on the y-axis, and the tread loss of the two different tires is on the x-axis. The initial wear rate can be provided as input to the system, for example, but not limited to, from the FEA stage, machine learning models, etc., to predict the tread depth progress of a given tire throughout its lifespan.

[0181] Next reference Figure 25 This paper presents a set of exemplary results when using this predictive method to simulate the wear of a reference tire, compared to measurement data from the same tire / vehicle / vehicle-tire system obtained through outdoor wear testing. Circular markers indicate the average tread depth of the control tire test results at each inspection mileage, while the solid line below represents the predicted tread depth relative to the initial tread depth and further normalized via a brush-type model.

[0182] like Figure 26 The validation data further illustrates an acceptable model fit for exemplary tire wear models, such as the hybrid brush type model disclosed herein. In this case, for each mileage at which the tread depth was examined, the difference between the predicted result and the indoor wear test result for a control tire was less than 0.25 mm.

[0183] The hybrid brush pattern model disclosed in this paper is extremely fast and efficient, and can be executed virtually in real time. Test results to date demonstrate that the model accurately predicts wear progression for very different tire designs. Only a relatively small subset of input is required, such as, for example, the original tread depth and the contact / void area at various tread depths. This information can be taken from, for example, a 3D model of the tread pattern, or from measurements taken by the circumferential tread wear imaging system (CTWIST), which is typically provided for each tire for indoor or outdoor wear testing.

[0184] In one implementation, within the scope of this disclosure, other tire-related threshold events can be predicted and implemented for alerts and / or interventions. For example, the system can identify additional services recommended for a given vehicle based on time-series inputs received and processed as described above, predicted tire wear, etc. Examples of such services may include, but are not limited to, tire rotation, alignment, inflation, etc. The system can generate alert and / or intervention recommendations based on comparisons of various thresholds, threshold groups, and / or non-threshold algorithms with respect to predetermined parameters.

[0185] In one implementation, within the scope of this disclosure, optimal types of tires and / or tire parameters can be predicted and implemented for alerts and / or interventions. For example, the system can identify vehicle application (more urban driving instances, more highway driving instances, etc.) and / or driving style based at least in part on time-series inputs received and processed as described above, predicted tire wear, etc. The system can determine that certain tires are more suitable for a given vehicle not only based on the vehicle type but also on the identified vehicle application and / or driving style, and also generate alerts and / or intervention recommendations based at least in part on this.

[0186] As previously mentioned, tire information can be obtained from one or more sensors mounted on a given tire or associated vehicle. These sensors may be accelerometers directly mounted, for example, on the tire liner or the vehicle's main shaft. Output signals from the sensors can be provided to a hosting server, eliminating the need for user input.

[0187] Now for more specific reference Figures 27 to 29 This document discloses another exemplary technique for estimating the tread depth of a tire. Those skilled in the art will understand that as the tire wears and loses mass, the modal frequencies change in a way that is directly related to or can be related to the mass loss. This principle is clear when considering a single-degree-of-freedom mass-spring system, where the natural frequency is equal to the square root of the spring stiffness divided by the mass. As the mass decreases, the natural frequency increases. Applying the same principle to the structural modes of the tire, the mass loss can be determined based on the modal frequency shift as follows:

[0188]

[0189] Where Δm is the mass change, m is the mass when the tire is new, and ωn is the natural frequency.

[0190] Modal frequencies can be identified through several methods, including (as previously described) attaching an accelerometer to the tire or to the vehicle's axle. Tire configuration patterns can also be excited in various ways, including, for example, controlled impacts to the tire by an object (such as a hammer, kicking the tire, etc.), electrical excitation, traversing obstacles (such as anti-skid wedges or speed bumps), and / or traveling on rough surfaces with the vehicle-tire assembly. In some embodiments, random excitation events can occur during operation of the vehicle-tire assembly, where output signals from sensors can be collected, stored, and / or processed to estimate tire wear.

[0191] Figure 27 An example from a static natural frequency test is shown, in which a given tire is impacted by a hammer, and an accelerometer is attached to the tire's inner liner. The vibration associated with the impact produces an output signal from the accelerometer with a power spectral density (PSD) waveform as shown. The PSD waveform for a given impact represents the frequency distribution of the associated output signal. The accelerometer can be configured to provide an output voltage, which can be converted by signal processing circuitry into an equivalent acceleration signal. These time-domain signals themselves can be further transformed into the frequency domain using, for example, a fast Fourier transform. The frequency response function in the power spectrum typically contains magnitude information expressed in decibels (dB).

[0192] The corresponding peaks in the spectra of the respective waveforms from the new state and wear state of a given tire are highlighted to show the frequency shift caused by tread loss between the two. In this example, the mass loss calculated according to the above formula is 0.474 kg, which is substantially the same as the actual measured value of 0.467 kg. In various embodiments, additional steps can be implemented to correlate the mass loss with the tread loss, or alternatively, the correlation of the modal frequency shift with respect to tread depth can be performed more reliably for a given tire.

[0193] Finite element analysis (FEA) simulations were also performed, which showed similar frequency shifts from both, for example, a transmissivity test (where the matrix is ​​excited by random input) and a skid wedge impact (where the tire rolls on the skid wedge).

[0194] Figure 28A The results are from a simulation of anti-skid wedge impact on a new tire, where the first graph shows the change in vertical force over time, and the second graph shows the magnitude of the Fast Fourier Transform (FFT) over a frequency range (in Hz).

[0195] Figure 28BThis represents the corresponding results from a simulation of anti-skid wedge impact on the same tire under wear conditions, where modal shifts can be easily observed between the new state and the wear state.

[0196] Figure 29 The results of the transmissivity simulations from the new state and wear state of a given tire are shown, with the transmissivity (in dB) relative to the spectrum. Modal shifts are readily observable between the new and wear states, and these shifts can be used to estimate changes in mass and, consequently, changes in tire wear / tread depth.

[0197] In each of the foregoing exemplary cases, the results shown are for the same tire, wherein the same frequency shift is observed between the worn tire model and the new tire model, and the frequency shift is realized in the disclosed tire wear model.

[0198] Throughout the specification and claims, unless the context otherwise requires, the following terms have at least the meaning explicitly associated herein. The meanings indicated below are not necessarily limiting of the terms, but rather provide illustrative examples only. The meanings of “an,” “a,” and “the” may include plural references, and the meaning of “in” may include “in” and “on”. As used herein, the phrase “in one embodiment” does not necessarily refer to the same embodiment, although it may be the same embodiment.

[0199] The various exemplary logic blocks, modules, and algorithmic steps described in conjunction with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or a combination of both. To clearly illustrate this interchangeability between hardware and software, the various exemplary components, blocks, modules, and steps have been generally described above in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the system as a whole. The described functionality may be implemented differently for each specific application, but such implementation decisions should not be construed as departing from the scope of this disclosure.

[0200] The various exemplary logic blocks and modules described in conjunction with the embodiments disclosed herein can be implemented or executed by a machine such as a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general-purpose processor may be a microprocessor, but in alternative embodiments, the processor may be a controller, a microcontroller, or a state machine, a combination thereof, etc. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, a combination of multiple microprocessors, a combination of one or more microprocessors combined with a DSP core, or any other combination of such configurations.

[0201] The steps of the methods, processes, or algorithms described in conjunction with the embodiments disclosed herein may be directly embodied in hardware, in a software module executed by a processor, or a combination of both. The software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM, or any other form of computer-readable medium known in the art. An exemplary computer-readable medium may be coupled to a processor such that the processor can read information from and write information to the memory / storage medium. Alternatively, the medium may be integrated into the processor. The processor and medium may reside in an ASIC. The ASIC may reside in a user terminal. Alternatively, the processor and medium may reside as discrete components in the user terminal.

[0202] Unless otherwise specifically stated or otherwise understood in the context in which they are used, the conditional language used herein (such as "may," "may," "can," "for example," etc.) is generally intended to convey that certain embodiments include certain features, elements, and / or states, while other embodiments do not include certain features, elements, and / or states. Therefore, such conditional language is not generally intended to imply that features, elements, and / or states are necessary in any way for one or more embodiments, or that one or more embodiments need to include logic for determining, with or without author input or prompting, whether such features, elements, and / or states are included in any particular embodiment or whether they will be performed in any particular embodiment.

[0203] While this document generally describes certain preferred embodiments of the invention in relation to tire wear and / or tire traction estimation for fleet management systems and more specifically autonomous vehicle fleets or commercial truck applications, the invention is explicitly not limited thereto, and unless otherwise stated, the term "vehicle" as used herein can refer to an automobile, truck, or any equivalent of the like that which may include one or more tires and thus require accurate estimation or prediction of tire wear and / or tire traction, as well as potential disabling, replacement, or intervention in the form of, for example, direct vehicle control adjustments (whether self-propelled or otherwise).

[0204] Unless otherwise stated, the term “user” as used herein may refer to a driver, passenger, mechanic, technician, fleet manager, or any other person or entity that may be associated, for example, with a device having a user interface for providing the features and steps disclosed herein.

[0205] The preceding detailed description has been provided for purposes of illustration and description. Therefore, although specific embodiments of the new and useful invention have been described, these references are not intended to be construed as limiting the scope of the invention, except as set forth in the following claims.

Claims

1. A computer-implemented method for estimating tire wear condition, comprising: Accumulate information in the data storage device about the probability distribution corresponding to each of the multiple tire wear factors; Collect vehicle data of the vehicle and / or tire data of at least one tire associated with the vehicle; The collected vehicle data and / or tire data are transmitted from the vehicle to a remote server that is communicatively linked to the data storage device; Based on the transmitted vehicle data and / or tire data, generate at least one observation corresponding to one or more of the plurality of tire wear factors; A Bayesian estimate of the current tire wear state of at least one tire associated with the vehicle is determined in real time, based at least in part on at least one observation generated and stored information about the probability distribution. One or more tire performance characteristics are predicted, at least in part, based on the determined current tire wear condition and the collected data; Selectively provide real-time feedback based on one or more predicted tire performance characteristics and / or determined current tire wear condition, and Information on updated probability distributions corresponding to a plurality of tire wear factors that contribute to tire wear of the at least one tire associated with the vehicle is stored, based at least one of the generated observations.

2. The method of claim 1, wherein the predicted one or more tire performance characteristics include predicted tire wear conditions at one or more future times for the at least one tire associated with the vehicle.

3. The method of claim 2, wherein the predicted one or more tire performance characteristics include the replacement time of the at least one tire associated with the vehicle, the replacement time being based on the current tire wear condition or the predicted tire wear condition relative to a tire wear threshold associated with the at least one tire associated with the vehicle.

4. The method of claim 1, wherein the information regarding the probability distribution reflects a time series relation array.

5. The method according to claim 1, further comprising: Receive one or more tire wear input values ​​from the user via a user interface associated with the remote server; and At least one observation of one or more of the plurality of tire wear factors is generated based on the one or more tire wear input values.

6. The method according to claim 1, further comprising: Receive one or more tire wear input values ​​generated by one or more sensors installed in the respective tires of the at least one tire; and At least one observation of one or more of the plurality of tire wear factors is generated based on the one or more tire wear input values.

7. The method according to claim 1, further comprising: Receive one or more tire wear input values ​​generated by sensors located outside the vehicle; as well as At least one observation of one or more of the plurality of tire wear factors is generated based on the one or more tire wear input values.

8. The method according to claim 7, wherein: At least one of the tire wear input values ​​generated by the sensors outside the vehicle includes a tread depth measurement result.

9. The method according to claim 1, further comprising: The estimated current tire wear condition is generated using the baseline value and the range corresponding to the confidence level of the Bayesian estimate.

10. A system for estimating tire wear condition, comprising: A data storage network that stores information about the probability distributions corresponding to each of the multiple tire wear factors; For each of the multiple vehicles, a distributed computing node is linked to one or more onboard sensors that are respectively configured to collect vehicle data. A server-based computing network, including a computer-readable medium having instructions residing thereon and executable by one or more processors, the server-based computing network being configured to direct the execution of the steps of the method according to any one of claims 1 to 9.

Images