System and method for detecting pressure loss rate and associated events of motor vehicle tires

Abstract

Disclosed herein are systems and methods for tire condition monitoring and more particularly for detecting slow leaks of inflation pressure. A data acquisition device (e.g., a tire pressure monitoring system sensor) is installed on a motor vehicle and collects data samples corresponding to at least tire inflation pressure. The collected data samples can be transmitted for analysis, for example, only when the motor vehicle is in a fleet yard or otherwise where the ambient air temperature effectively matches the environmental temperature. An elapsed time is calculated from a first data sample within a defined sampling period, and a statistical model is applied to data samples corresponding to at least inflation pressure versus elapsed time. A tire slow leak event is determined based on an evaluation of a decrease in inflation pressure from the statistical model, and an output signal corresponding to the determined slow leak event is selectively generated.

Description

TECHNICAL FIELD

[0001] The present disclosure relates generally to quantifying performance aspects of tires on wheeled motor vehicles. More specifically, the systems, methods, and related algorithms as disclosed herein relate to the detection of air pressure events, such as slow leaks of tires of wheeled motor vehicles, including but not limited to motorcycles, consumer vehicles (e.g., passenger cars and light trucks), commercial and off-the-road (OTR) vehicles. BACKGROUND

[0002] Undetected internal air pressure leaks are critical to the optimal performance of any tire on the market. This is true for the owner of a personal motor vehicle, but even more so for the owners and managers of heavy equipment and transportation fleets, who understandably worry about air loss in tires due to safety, performance, and environmental concerns, among others.

[0003] Air leaks can present different challenges to companies operating under tight schedules. Many of these issues can include downtime, potential delivery delays, and sometimes even serious damage to expensive equipment. There are many traditional solutions on the market, including various vehicle monitoring systems, including, for example, the ability to notify users of operational information such as vehicle speed, vehicle load, vehicle longitude and latitude coordinates. However, traditional solutions have failed to specifically and accurately address the issue of air pressure leaks at each wheel position on a motor vehicle.

[0004] For example, it is traditionally known to monitor the working pressure history of a tire obtained from a tire pressure monitoring system (TPMS) and to further calculate the internal pressure reduction of each tire by comparative analysis between each tire attached to the same motor vehicle to detect slow leaks. Typically, because the internal air pressure and the contained air temperature can be assumed to have a proportional relationship, changes in the internal air pressure can be considered to depend on changes in the external (ambient) temperature and the vehicle temperature. Due to temperature fluctuations, a “corrected” internal pressure value can be determined by applying the known relationship between the pressure, temperature, and volume of the subject.

[0005] However, the tire temperature during travel depends at least in part on the associated operating conditions. Tire temperature measurements from TPMS sensors during normal operation can vary, for example, from temperatures close to ambient air to approximately 70 degrees Celsius, depending on the operating and / or environmental conditions. Temperature-compensated internal pressures can still be derived, but even taking this into account, the amount of dispersion of the observation error is typically increased.

[0006] When the dispersion of the internal pressure measurements is relatively large, it is difficult to properly assess the reduction over time, and false positive or false negative slow leak detections can occur frequently as a result. SUMMARY

[0007] In view of the aforementioned deficiencies in conventional systems, machine learning methods as disclosed herein can be implemented to accurately detect slow leaks in tires using, for example, in-field monitoring units.

[0008] Generally speaking, various embodiments of the slow leak detection systems and methods as disclosed herein can implement detected operating and / or environmental conditions associated with motor vehicles, including, for example, internal air pressure measurements and contained air temperature (CAT) measurements associated with a given tire. Preferably, the responses can be measured directly using a data acquisition system, such as, for example, a tire pressure monitoring system (TPMS), which can be installed in, on, or otherwise associated with a tire. The data acquisition system can preferably transmit data, or otherwise obtain associated data collected therefrom, while the motor vehicle is located in a fleet yard, and wherein the dispersion of internal pressure measurements is reduced.

[0009] In embodiments as disclosed herein, a random forest machine learning method can preferably be implemented to enable detection of slow leaks, and more particularly to correlate slow leaks occurring while a respective motor vehicle is found in a monitored fleet yard environment, such that the systems and methods as disclosed herein can readily alert users of potential slow leaks occurring in their vehicles and intervene to bring the vehicle tires up to specification and provide optimal performance.

[0010] An exemplary embodiment of a tire monitoring method as disclosed herein includes collecting data samples corresponding to at least an inflation pressure of at least one tire of a plurality of tires via at least one data acquisition device installed on a motor vehicle having the plurality of tires. An elapsed time can be calculated from a first data sample within a defined sampling period, and a statistical model can be applied to data samples corresponding to at least the inflation pressure relative to the elapsed time. A slow leak event can be determined based on an evaluation of a decrease in the inflation pressure from the statistical model, wherein an output signal corresponding to the determined slow leak event can be selectively generated for at least one tire of the plurality of tires.

[0011] In one exemplary aspect according to the above-described embodiments, the statistical model can require at least a first threshold of data samples within a defined sampling period, and the elapsed time from the first data sample must exceed a second threshold.

[0012] In another exemplary aspect according to the above-described embodiments, data samples are collected for the statistical model only when a speed of the motor vehicle is determined to have been zero for a third time threshold.

[0013] In another exemplary aspect in accordance with the above-described embodiments, the data samples are collected only when the data acquisition device is within range of one or more data collection units in a fleet yard monitoring system.

[0014] The above-described embodiments can further include collecting, via the at least one data acquisition device installed on the motor vehicle having the plurality of tires, an internal air temperature associated with the data samples corresponding to at least an inflation pressure of at least one of the plurality of tires. For example, a temperature-compensated inflation pressure value can be generated for each of the data samples, where the statistical model implements the temperature-compensated inflation pressure value for determining the slow leak event. In another example, a determination can be made whether to selectively generate an output signal corresponding to the determined slow leak event based at least in part on the associated internal air temperature. In another example, a determination can be made whether to selectively generate the output signal corresponding to the determined slow leak event based on a rate of change per hour of the associated internal air temperature relative to a threshold value.

[0015] In another exemplary aspect in accordance with the above-described embodiments, the slow leak event can be further determined with respect to a second defined sampling period, taking into account a median value of a metric corresponding to a decrease in evaluation of the inflation pressure.

[0016] In another exemplary aspect in accordance with the above-described embodiments, the statistical model can include a linear regression model having a target variable including the inflation pressure and a descriptor variable including elapsed time. Additionally or alternatively, the statistical model can include a random forest model.

[0017] In another embodiment as disclosed herein, a tire monitoring system includes at least one data acquisition device installed on a motor vehicle having a plurality of tires and configured to collect data samples corresponding to at least an inflation pressure of at least one of the plurality of tires, and at least one data collection unit configured to receive the collected data samples from the on-board data acquisition device, for example, when the motor vehicle is in a yard and not operating on a road. A processing unit is linked to the at least one data collection unit and configured to direct performance of operations in accordance with the above-described method embodiments and, optionally, one or more of the above-associated exemplary aspects.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The application disclosed herein can be embodied in other specific forms without departing from the spirit or essential attributes of the application and it is therefore desired that the application be considered in all its aspects as illustrative and not restrictive. Any headings used herein are for convenience only and do not interpret the scope or limit the disclosure. Many of the objects, features and advantages of the embodiments described herein can become apparent from the following disclosure, when read in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0019] Embodiments of the application are illustrated in more detail below with reference to the accompanying drawings.

[0020] Figure 1 is a block diagram presenting an embodiment of a tire monitoring system as disclosed herein.

[0021] Figs. 2a and 2b are graphical diagrams presenting the presence of normal conditions and slow leak conditions, respectively, based on collected time series data.

[0022] Figures 3 to 7 is a graphical diagram presenting exemplary pattern recognition steps according to an embodiment of the system / method as disclosed herein.

[0023] Figure 8 is a graphical diagram presenting exemplary slopes of lines of best fit of internal air pressure measurements over time.

[0024] Figure 9 and Figure 10 is a graphical diagram presenting exemplary defined sampling periods according to the present disclosure.

[0025] Figs. 11a to 11d are graphical diagrams presenting exemplary time series data associated with true positive event detection, false negative event detection, false positive event detection and true negative event detection, respectively.

[0026] Figure 12 is a flowchart presenting an embodiment of a tire monitoring method as disclosed herein. DETAILED DESCRIPTION

[0027] OVERVIEW Figures 1 to 12 Various exemplary embodiments of the present application can now be described in detail in connection with the accompanying drawings. In various drawings, embodiments that can share common elements and characteristics can be described with like reference numerals, and redundant descriptions thereof can be omitted below.

[0028] INITIAL REFERENCES Figure 1An exemplary embodiment of the system 100 includes a data acquisition device 110 that is on-board the vehicle and configured to obtain at least data and transmit the data to one or more downstream computing devices (e.g., remote servers) to perform relevant computations as disclosed herein. The data acquisition device can be a standalone sensor unit that is suitably configured to collect raw measurement signals such as, for example, signals corresponding to the tire’s contained air temperature 112 and / or internal air pressure 114, and continuously or selectively transmit such signals. The data acquisition device can include an on-board computing device that is in communication with one or more distributed sensors and is portable or otherwise modular as part of a distributed vehicle data collection and control system, or can otherwise be integrally provided with respect to a central vehicle data collection control system. The data acquisition device can include a processor and memory (not shown) having program logic resident thereon, and in various embodiments can include a vehicle electronic control unit (ECU) or component thereof, or as otherwise can be discrete in nature, e.g., permanently or removably disposed with respect to a vehicle mount.

[0029] Generally speaking, the system 100 as disclosed herein can implement a number of components distributed across one or more vehicles, e.g., but not necessarily associated with a fleet management entity, and can further implement a central server network or event-driven serverless platform in functional communication with each of the vehicle motors via a communications network. Exemplary vehicle components can generally include one or more sensors such as, for example, a body accelerometer, a gyroscope, an inertial measurement unit (IMU), a location sensor such as a global positioning system (GPS) transponder, a tire pressure monitoring system (TPMS) sensor transmitter and associated on-board receiver, a gateway device, etc., linked to a controller area network (CAN) bus network and providing signals thereto to a local processing unit. For purposes of illustration, the illustrated embodiment can include a TPMS sensor unit mounted on a tire, an ambient temperature sensor, a speed sensor configured to collect, e.g., acceleration data associated with the vehicle, and a DC power source, without otherwise limiting the scope of the present disclosure.

[0030] One or more of the sensors as disclosed herein can be integrated or otherwise co-located in a given modular structure, rather than being discrete and distributed in the structure. For example, the TPMS sensor mounted on a tire as referred to herein can be configured to generate output signals corresponding to each of a plurality of tire-specific conditions, e.g., inflation pressure, contained air temperature. The TPMS sensor may, for example, be mounted inside a tire air chamber, slightly elevated and isolated from the metal wheel rim so as not to be adversely affected.

[0031] Various bus interfaces, protocols, and associated networks are well known in the art for communication between respective data sources and local computing devices, among other things, and those skilled in the art will recognize a wide range of such tools and apparatus for implementing these tools.

[0032] In various embodiments, the data collection devices and equivalent data sources 110 as disclosed herein are not necessarily limited to vehicle-specific sensors and / or gateway devices, and can also include third-party entities and associated networks, program applications resident on user computing devices such as driver interfaces, fleet management interfaces, and any enterprise devices or other providers of raw streams of recorded data that can be considered relevant to the algorithms and models as disclosed herein.

[0033] Referring again to Figure 1 A data pipeline stage 120 can be provided in which data collected from one or more data sources 110 is transmitted to a data processing stage 130. The data processing stage 130 also interacts with a data storage stage 140 including, for example, one or more database services, where the data processing stage 130 and / or the data storage stage 140 selectively interact with external devices 150 and / or networks via, for example, respective application program interfaces (API) requests.

[0034] In one embodiment, the exemplary data pipeline stage 120 can include an event-driven serverless architecture in which one or more event hubs are configured to facilitate capturing raw data from respective sources, generating normalized data streams therefrom, and further replicating ingested events in relevant time intervals to data storage resources. The normalized and enhanced data streams can be further submitted for analytics processing via, for example, a data lake platform known in the art. Non-limiting examples of data lakes known in the art can include Azure Data Lake, Amazon S3, and the like. and the like.

[0035] It should be noted that Figure 1 The embodiments presented in the foregoing do not limit the scope of the systems or methods as disclosed herein, and in alternative embodiments, one or more of the pressure loss rate models can be implemented locally at an on-board computing device (e.g., an electronic control unit) rather than at a downstream computing stage. For example, the models can be generated and trained over time at a server level and downloaded to an on-board computing device for local execution of one or more steps or operations as disclosed herein.

[0036] In other alternative embodiments, one or more of the various sensors 112, 114 can be configured to communicate with a downstream platform without a local on-board device or gateway component, such as, for example, via a cellular communication network or via a mobile computing device (not shown) carried by a user of the vehicle.

[0037] The term "user interface" as used herein, unless otherwise indicated, can include any input-output module through which a user device facilitates user interaction with respect to processing units, servers, devices, etc. as disclosed herein, including but not limited to downloaded or otherwise resident program applications; web browsers; web portals such as individual web pages or those collectively defining a hosted website; etc. The user interface can also be described in the context of buttons and display portions with respect to a personal mobile computing device, which can be independently arranged or otherwise interrelated with respect to, for example, a touch screen, and can also include audio and / or visual input / output functionality, even without explicit user interaction.

[0038] In one embodiment, the vehicle and tire sensors 112, 114, etc. can also be provided with unique identifiers, wherein the on-board device processor can distinguish between signals provided from corresponding sensors on the same vehicle, and further in certain embodiments, wherein the central processing unit and / or fleet maintenance supervisor client device can distinguish between signals provided from tire and associated vehicle and / or tire sensors on multiple vehicles. In other words, in various embodiments, the sensor output values can be associated with a particular tire, a particular vehicle, and / or a particular tire-vehicle system for the purposes of on-board or remote / downstream data storage and implementation for calculations as disclosed herein. The on-board data acquisition device can communicate directly with the downstream processing stage 130 as shown, or alternatively, the driver's mobile device or computing device installed on the truck can be configured to receive the on-board device output data and process / transmit it to one or more downstream processing units. Figure 1

[0039] ​The raw signals received from the particular vehicle and / or tire sensors 112, 114, etc. can be stored in the on-board device memory, or in an equivalent local data storage network functionally linked to the on-board device processor, for selective retrieval and transmission via the data pipeline stage 120 for computation as needed according to the methods disclosed herein. As used herein, a local or downstream "data storage network" can generally refer to individual, centralized or distributed logical and / or physical entities configured to store data and enable selective retrieval of data therefrom, and can include, for example and without limitation, memories, lookup tables, files, registers, databases, database services, etc. In some embodiments, the raw data signals from the various sensors 112, 114, etc. can be transmitted from the vehicle to a downstream processing unit in substantially real-time. Alternatively, particularly in light of the inefficiencies inherent in continuous data transmission of high frequency data, the data can be compiled, encoded, and / or aggregated, for example, for more efficient (e.g., periodic time-based or alternatively defined event-based) transmission from the vehicle to the processing unit via an appropriate (e.g., cellular) communication network.

[0040] The vehicle data and / or tire data 112, 114, etc. once transmitted to the downstream processing unit via the communication network can be stored in a database associated therewith, for example, and further processed or otherwise retrievable as input for processing via one or more algorithmic models as disclosed herein. The models can be implemented at least in part via execution of a processor, enabling selective retrieval of the vehicle data and / or tire data, and also enabling electronic communication for input of any additional data or algorithms from databases, lookup tables, etc. stored in association with the processing unit.

[0041] As used herein, the term "processor" or "processing unit" or "processing stage" 130 can refer at least to a general or special purpose processing device and / or logic as can be appreciated by those skilled in the art, including but not limited to a microprocessor, microcontroller, state machine, etc. The processor can also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration.

[0042] The various illustrative logical blocks, modules, and algorithm steps described in connection with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. The described functionality can be implemented in varying ways for each particular application, but such implementation decisions can not be interpreted as causing a departure from the scope of the present disclosure.

[0043] The various illustrative logical blocks and modules described in connection with the embodiments disclosed herein can be implemented by machines such as general purpose processors, Digital Signal Processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general purpose processor can be a microprocessor, but in the alternative, the processor can be a controller, a microcontroller, or a state machine, combinations of the

[0044] The steps of a method, process, or algorithm described in connection with the embodiments disclosed herein can be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module can reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of computer-readable medium known in the art. An exemplary computer-readable medium can be coupled to the processor such that the processor can read information from, and write information to, the computer-readable medium. In the alternative, the medium can be integral to the processor. The processor and the medium can reside in an ASIC. The ASIC can reside in a user terminal. In the alternative, the processor and the medium can reside as discrete components in a user terminal.

[0045] Reference will now be made to FIGS. 2 to Figure 12An exemplary method 200 for quantifying performance aspects of motor vehicle tires, and more particularly for detecting slow leak phenomena of internal air pressure at an early stage, can now be described. A slow leak event, one exemplary leak event of which is presented in FIG. 2b, can be defined as a phenomenon of rapid internal pressure reduction as compared to a natural reduction, an example of which is presented in FIG. 2a. If a motor vehicle continues to travel with one or more tires having an internal pressure that is extremely low, the risk of permanent damage to the tire typically increases. On the other hand, if a slow leak event can be identified at an early stage, intervention such as repairing and replacing the tire can greatly reduce such risks.

[0046] Method 200 begins with collecting signals at a data collection stage (step 210), which as previously described can implement conventional on-board data collection equipment such as a tire pressure monitoring system (TPMS) installed in or on a tire that can generate signals corresponding to one or more of internal air temperature, ambient temperature, inflation pressure, tire identifier, vertical load, speed, etc. In some embodiments, the data collection equipment can be configured to collect data as the tire rolls over different roads and surfaces, but the system can be configured such that the only data considered in subsequent steps is collected while the motor vehicle is stopped or otherwise resident in a fleet yard.

[0047] In other words, a data collection unit that provides tire data to a processing unit can geographically limit its ability to communicate with data collection equipment, and can also be fixed relative to a fleet yard location such that the only time a data sample is collected from a given motor vehicle is when the motor vehicle is in the fleet yard. In one embodiment, a wireless communication network associated with the in-yard monitoring unit can include one or more wireless routers that receive signals from data collection equipment (e.g., either directly from TPMS sensors or indirectly via an on-board computing device), decode the signals, and forward the decoded signals for downstream processing. An exemplary data transmission frequency can be about 2.5 readings per minute when the vehicle is resident in the fleet yard.

[0048] As previously described, conventional methods of utilizing tire inflation pressure data during vehicle operation can result in false positives or false negatives due to variability of associated temperature compensation. It can also be appreciated that due to the periodic nature of data collection (corresponding to times when the vehicle is in the fleet yard), the variable frequency of readings can introduce noise into subsequent calculations and at least potentially result in false positives or false negatives. Accordingly, the data models as disclosed herein can implement various techniques to filter data and flag and identify time series data patterns for accurate identification of relevant pressure loss events.

[0049] Reference is next made toFigure 3 In the example data set, the cold inflation pressure value 310 for a given tire is shown to gradually decrease from a first time 340 until a second time 350 at which a significant inflation event occurs. At the first time 340, the measured value 310 is comfortably between the recommended threshold 302 and the warning threshold 304. Over time, the measured value 310 decreases below the warning threshold 304 and even the critical threshold 306, just prior to the second time 350, and as an indication of a gradual decrease over several hours, and substantially out of sync with fluctuations in the associated temperature value 320. In other words, while the measured temperature value 320 continuously fluctuates within the band 330, the gradual decrease in tire inflation pressure is clearly not a function of temperature.

[0050] Referring next to Figure 4 , another example data set shows another gradual multi-day decrease in inflation pressure value 310, which is interrupted by a downward spike and returns to the general trend. Despite the presence of the downward spike, which can correspond to, for example, an inflation pressure check, the overall trend indicates a slow leak event, as preferred by the models disclosed herein. It should also be noted that time series data corresponding to slow leak events are often bracketed on the first and second ends by tire inflation events (as shown herein), but not all such slow leak events are. As Figure 5 shown, for example, two successive gradual multi-day decreases in tire inflation pressure measurements 310 can preferably be flagged as two separate events.

[0051] As Figure 6 shown, another example data set illustrates the importance of focusing on the overall trend in inflation pressure measurements 310, rather than determining a false positive for a slow leak event based on an intermittent downward spike (e.g., corresponding to a pressure check). In this case, the overall trend in pressure measurements is flat. On the other hand, in Figure 7 , a sudden (i.e., non-gradual) decrease in internal pressure measurements 310 is presented without a return to the trend, which can indicate that the tire has suffered a major failure, such as a spike or other obstruction. In some embodiments, such an event can be characterized as equivalent to a slow leak event, in order to generate a warning or the like for intervention, but in other embodiments, such as Figure 7 shown in the data set shown in

[0052] Returning to Figure 12, the method 200 can continue by compiling data samples within a defined sampling time period (step 220). In one embodiment, this involves creating a subset of internal pressure measurements collected within the previous twenty hours. Other thresholds are set to define a data acquisition gate, such that the amount of data on which the pressure loss rate estimate relies is sufficient, and thus the results of the algorithm are less susceptible to noise. In step 230, the number of samples in the created subset is compared to a predetermined first threshold. If the first threshold is satisfied, in step 240 the method continues to calculate the difference between the time of the first observation in the created subset and the time of the most recent observation in the created subset, and compare the difference to a predetermined second threshold. In one example, the necessary number of samples is five and the necessary time period between the first sample and the most recent sample is thirty minutes, although various alternative values and thresholds can be implemented within the scope of the present disclosure.

[0053] Where the first threshold and the second threshold (or their equivalents) have been satisfied, the method 200 can continue by applying a statistical model to determine a metric associated with the pressure loss rate as disclosed herein (step 250). In the following example, the hourly pressure loss rate (hPLR) can be estimated by constructing a linear regression model using least squares, where the target variable is the tire internal pressure in pounds per square inch (psi) (or a temperature-compensated internal pressure as described further below), and the descriptor variable is elapsed time:

[0054] Pressure (psi) = a + b x (Elapsed Time)

[0055] In this case, and as shown in Figure 8 , the regression coefficient b is estimated as the hourly pressure loss rate (hPLR) over a defined time window (e.g., twenty hours), while the pressure loss rate itself can be the slope of the best-fit line 360 of the internal pressure over time calculated using the linear regression model.

[0056] In various embodiments (not shown in FIG. 2), “corrected” internal pressure values (i.e., temperature-compensated) can be further used in the above equation, and obtained as, for example, by applying the ideal gas law to the sampled temperature and internal pressure values.

[0057] In step 260, the method 200 continues by comparing the estimated metric (e.g., hPLR) to a predetermined third threshold, where if the threshold is exceeded, a slow leak event is determined.

[0058] In the event that a slow leak event is determined, the method 200 can continue by determining a leak rate associated with the slow leak event (step 270). In one example, the leak rate can be determined by using the linear regression model described above, and calculating the slope of the best-fit line 360 of the internal pressure over time. In another example, the leak rate can be determined by using the linear regression model described above, and calculating the difference between the internal pressure at the time of the first observation in the created subset and the internal pressure at the time of the most recent observation in the created subset. Figure 9In the illustrated embodiment, from the current time Tl back, a separate hPLR estimate is determined for each new measurement within a defined sampling period. For each new estimate, the median hPLR of each data block (e.g., 45 minutes) is determined, in this example defining three different median hplr1, hplr2, hplr3 for three consecutive 45 minute blocks. If the median hPLR value of the respective block of each estimate is determined to be, for example, less than -1.0 psi, a slow leak event for the corresponding tire can be determined. As Figure 10 As further illustrated in the middle, for each new hPLR estimate, the individual historical hPLR values can be regrouped as long as they fall within a defined sampling window (e.g., 20 hours), and the median is calculated and compared to a threshold (e.g., -1.0 psi) again.

[0059] The method 200 can optionally include a step 270 in which the determined slow leak event can be suppressed or eliminated based on further determining that a temperature-based metric, such as, for example, a rate of hourly change in temperature values corresponding to the same internal pressure values that are the basis for the slow leak event determination, exceeds a predetermined fourth threshold. As one example, a rate of hourly temperature change (hTCR) that exceeds twenty degrees Fahrenheit over a corresponding time period can result in the determined slow leak event being suppressed or eliminated.

[0060] Referring next to FIGS. 11a-11d, sample slow leak events and non-events from a random forest approach are illustrated. The highlighted portion of FIG. 11a represents a “true positive” (TP) system output, which corresponds to a correctly detected actual slow leak event. The highlighted portion of FIG. 11b represents a “false negative” (FN) system output, which corresponds to an actual slow leak event that was not detected. The highlighted portion (arrow) in FIG. 11c represents a “false positive” (FP) system output, which corresponds to a slow leak event that was incorrectly detected, where the slope downward from a short-lived upward spike is associated with a slow leak event, but there is no significant change in the tire internal pressure trend. FIG. 11d presents a “true negative” (TN) system output, which corresponds to a correctly detected non-event.

[0061] It should be understood that the various thresholds provided herein for the method 200 are merely exemplary, as these values can be optimized over time for a given application.

[0062] The evaluation methods of the above statistical models can be considered as classification problems in the context of machine learning, which can be used to maximize the precision, recall, and accuracy of the evaluation over time in the following examples. In this case, the "precision" can relate to the probability, for example, that if a warning is triggered, the warning corresponds to a true condition, and can be quantified as {precision = TP / (TP+FP)}. In this case, the "recall" can relate to the probability, for example, that if there is a true slow leak condition, the corresponding warning is triggered, and can be quantified as {recall = TP / (TP+FN)}. In this case, the "accuracy" can consider both precision and recall to account for the frequency of correct warning triggering. For example, if a single FP warning is triggered for one (1) wheel position on a motor vehicle, the precision and accuracy are equal to 0. If the same single FP warning occurs in a population of one hundred (100) wheel positions, the precision is still equal to 0, but the accuracy is 99% (because there are 99 TN readings). This can be quantified as {accuracy = (TP+TN) / population}. Since the precision and recall metrics are effectively trade-offs between each other, an additional score (F1) can be implemented to balance the precision and recall metrics by taking their average, such that {F1 = (precision+recall) / 2}. In order to compute the index, the correct and incorrect answers must be defined in the classification problem. In an exemplary evaluation, the answer is correct for a "slow leak" sample if even one warning is issued in the six days of the extracted partial time series, and the answer is correct for a "normal" tire sample if no warnings are issued in the partial time series.

[0063] In various embodiments, the method 200 can further include generating an output signal corresponding to the detected slow leak event (step 280), such as, for example, in the form of a warning or message issued to a user interface or display unit. The output signal can be programmatically generated in response to the detected slow leak event. The output signal can be generated responsively to a received user request, such as, for example, by logging events over time and delivering reports in batch format. The output signal can also be generated to automatically trigger or otherwise facilitate a control response or intervention with respect to motor vehicle control or fleet management control.

[0064] In one embodiment, the slow leak event output from the system 100 and method 200 can be further implemented to predict future timing of tire intervention, such as recommended or required inflation or replacement scheduling.

[0065] Throughout the specification and claims, the following terms take at least the meanings explicitly associated herein, unless the context response otherwise. The meanings identified below do not necessarily limit the terms, but provide illustrative examples for the terms. The meaning of "a," "an," and "the" includes plural references, and the meaning of "in" includes "in" and "on." As used herein, the phrase "in an embodiment" does not necessarily refer to the same embodiment, although it can.

[0066] Unless specifically stated otherwise, as apparent from the following discussions, it is appreciated that throughout the specification discussions utilizing terms such as "can," "could," "might," "may," "e.g.," or the like, refer to potentialities that the actual implementation can or can not utilize one or more of the potential features, elements, and / or states. Accordingly, no inference should be drawn regarding an inherent requirement that any particular features, elements, and / or states are or are not more or less essential to one or more implementations.

[0067] While certain preferred embodiments of the present application can be generally described herein with respect to methods performed by or on behalf of a fleet management system and more particularly for autonomous vehicle fleets or commercial truck applications, the present application is expressly not limited thereto and unless otherwise noted, the term "vehicle" as used herein can refer to an automobile, truck or any equivalent thereof whether self-propelled or otherwise, as can include one or more tires and thus require accurate estimation or prediction of tire internal air pressure loss and potential disabling, replacement or intervention.

[0068] Unless otherwise noted, the term "user" as used herein can refer to a driver, passenger, mechanic, technician, fleet manager or any other person or entity that can be associated with a device having a user interface for providing features and steps as disclosed herein, for example.

[0069] The foregoing detailed description has been provided for purposes of illustration and description. Thus, although specific embodiments of a new and useful application have been described, it is not intended to limit the scope of the application to the specific embodiments described. Rather, it is intended to cover any and all modifications, variations, and equivalents that fall within the scope of the present application as defined by the following claims.

Claims

1. A computer-implemented tire monitoring method (200), the method comprising the steps of: collecting, via at least one data collection device (110, 112, 114) installed on a motor vehicle having a plurality of tires, data samples (210) corresponding to at least an inflation pressure of at least one of the plurality of tires; computing a time elapsed from a first data sample over a defined sampling period (220); applying a statistical model (250) to at least the data samples corresponding to the inflation pressure relative to the time elapsed when a number of data samples over the defined sampling period exceeds a first threshold (230) and the time elapsed from the first data sample exceeds a second threshold (240); determining a slow leak event (260) based on an assessed decrease in the inflation pressure from the statistical model; and selectively generating an output signal (280) corresponding to the determined slow leak event for the at least one of the plurality of tires.

2. The tire monitoring method of claim 1, wherein the data samples are collected for the statistical model only when a speed of the motor vehicle is determined to have been zero for a third time threshold.

3. The tire monitoring method of claim 1, wherein the data samples are collected only when the data collection device is within a range of one or more data collection units in a fleet yard monitoring system.

4. The tire monitoring method of claim 3, further comprising: collecting, via the at least one data collection device installed on a motor vehicle having a plurality of tires, an inclusion air temperature associated with the data samples corresponding to the at least inflation pressure of at least one of the plurality of tires; and generating a temperature compensated inflation pressure value for each of the data samples, wherein the statistical model implements the temperature compensated inflation pressure value for determining the slow leak event.

5. The tire monitoring method of claim 3, further comprising: collecting, via the at least one data collection device installed on a motor vehicle having a plurality of tires, an inclusion air temperature associated with the data samples corresponding to the at least inflation pressure of at least one of the plurality of tires; and determining whether to selectively generate the output signal corresponding to the determined slow leak event based at least in part on the associated inclusion air temperature.

6. The tire monitoring method of claim 5, comprising determining whether to selectively generate the output signal corresponding to the determined slow leak event based on a rate of change per hour of the associated inclusion air temperature relative to a threshold.

7. The tire monitoring method of claim 1, wherein the slow leak event is further determined in view of a median of a measure corresponding to the assessed decrease in the inflation pressure over a second defined sampling period.

8. The tire monitoring method of claim 1, wherein the output signal corresponding to the determined slow leak event is implemented to predict timing of a tire intervention event, the tire intervention event including one or more of a recommended inflation, a required inflation, and a replacement schedule.

9. The tire monitoring method of claim 1, wherein the statistical model includes a linear regression model having a target variable including the inflation pressure and a descriptor variable including elapsed time.

10. The tire monitoring method of claim 1, wherein the statistical model includes a random forest model.

11. A tire monitoring system (100), the system comprising: at least one data acquisition device (110, 112, 114) mounted on a motor vehicle having a plurality of tires and configured to collect data samples corresponding to at least an inflation pressure of at least one of the plurality of tires; at least one data collection unit configured to receive the collected data samples from the on-board data acquisition device (120, 140); and a processing unit (130) linked to the at least one data collection unit, wherein the processing unit and / or the at least one data acquisition device is configured to direct performance of each step of the method of any one of claims 1 to 8.

12. The tire monitoring system of claim 11, wherein the statistical model includes a linear regression model having a target variable including the inflation pressure and a descriptor variable including elapsed time.

13. The tire monitoring system of claim 11, wherein the statistical model includes a random forest model.

Images