Indirect tire pressure monitoring method, device, vehicle and storage medium

Abstract

The present application relates to the field of tire technology, and more particularly to an indirect tire pressure monitoring method, device, vehicle, and storage medium. The method comprises: obtaining training data containing normal tire pressure labels and underpressure labels; performing signal processing on the wheel speed signal in the training data to obtain a processed wheel speed signal; detrending the processed wheel speed signal to obtain a wheel speed micro-signal; extracting wheel speed micro-vibration characteristics from the wheel speed micro-signal, using the wheel speed micro-vibration characteristics to train a pre-built neural network, obtaining a tire pressure probability identification model after training, and using the tire pressure probability identification model to identify the pressure state of each tire in the vehicle. This solves the problems in the related art of requiring manual construction of a tire rigid ring model, lacking the ability to generalize to uncalibrated tires, and high cost investment.

Description

Technical Field

[0001] The present application relates to the field of tire technology, and in particular to an indirect tire pressure monitoring method, device, vehicle, and storage medium. Background Art

[0002] Compared with direct tire pressure monitoring, the indirect tire pressure monitoring system does not require the installation of any additional equipment. It obtains the wheel speed signal through the car's wheel speed sensor and completes a series of algorithms to achieve tire pressure monitoring. Therefore, it has been more widely used.

[0003] There are two classic methods for indirect tire pressure monitoring: the radius method and the frequency method. The radius method relies on real-time monitoring of tire pressure changes based on the relationship between tire radius, relative tire radius, effective rolling radius, and the number of pulses emitted by the ABS wheel speed sensor. The other method directly estimates tire pressure by estimating the equivalent number of teeth per unit mileage, comparing the number of pulses per kilometer, and the cumulative value of wheel speed. The frequency method has a relatively straightforward process: wheel speed data from the wheel speed sensor is reconstructed in the time domain for a Fourier transform, and finally, the resonance frequency is estimated to determine under-pressure.

[0004] Indirect tire pressure monitoring systems (TPMs) in the prior art require excessively long initial calibration and matching times, typically requiring two months or more of road testing and calibration. Calibration must also be performed for each tire brand specified by the vehicle manufacturer. They lack the ability to generalize to uncalibrated tires, resulting in a high probability of false alarms (incorrectly illuminating the underpressure warning light when there is no actual underpressure), impacting the driver's driving experience. False alarms are a common problem with existing indirect TPM systems, subject to widespread consumer complaints and criticism. The system requires real-time calculation of the tire's circumferential vibration frequency, which is time-consuming, results in poor precision in manually constructed tire models, and results in low algorithm accuracy. The system is highly dependent on accumulated mileage during calibration. Summary of the Invention

[0005] The present application provides an indirect tire pressure monitoring method, device, vehicle and storage medium to solve the problems in related technologies such as the need to manually construct a tire rigid ring model, the lack of generalization capability for uncalibrated tires, and high cost investment.

[0006] The first aspect of the present application provides an indirect tire pressure monitoring method, comprising the following steps: obtaining training data carrying normal tire pressure labels and underpressure labels; performing signal processing on the wheel speed signal in the training data to obtain a processed wheel speed signal, and performing detrending processing on the processed wheel speed signal to obtain a wheel speed micro-signal; extracting wheel speed micro-vibration characteristics from the wheel speed micro-signal, and using the wheel speed micro-vibration characteristics to train a pre-constructed neural network, obtaining a tire pressure probability identification model after the training is completed, and using the tire pressure probability identification model to identify the pressure status of each tire in the vehicle.

[0007] Optionally, the detrending process of the processed wheel speed signal to obtain the wheel speed micro-signal includes: obtaining the longitudinal vehicle speed at the same moment as the processed wheel speed signal; and calculating the wheel speed micro-signal based on the wheel speed signal and the longitudinal vehicle speed at the same moment.

[0008] Optionally, the tire pressure probability identification model includes an input layer, an output layer and a hidden layer, wherein the input signal of the input layer is a wheel speed microscopic signal, the output layer outputs a normal tire pressure label and an underpressure label, and the hidden layer contains multiple neurons.

[0009] Optionally, the tire pressure probability identification model is also set with a training number. After using the tire pressure probability identification model to identify the actual state of the tire pressure, it also includes: obtaining an online wheel speed micro-signal during vehicle operation; using the wheel speed micro-signal to iteratively optimize the tire pressure probability identification model until the number of iterations reaches the set training number, thereby obtaining an optimized tire pressure probability identification model.

[0010] Optionally, the signal processing of the original wheel speed signal in the training data to obtain the processed wheel speed signal includes: extracting the angular physical error corresponding to each tooth in the wheel speed sensor ring gear; eliminating the angular physical error in the original wheel speed signal, and performing nonlinear interpolation to obtain the processed wheel speed signal.

[0011] A second aspect of the present application provides an indirect tire pressure monitoring method, comprising the following steps: collecting a wheel speed signal from each tire in a vehicle; performing signal processing on the wheel speed signal to obtain a processed wheel speed signal, and performing detrending processing on the processed wheel speed signal to obtain a wheel speed micro-signal; inputting the wheel speed micro-signal into a completed tire pressure probability identification model, and outputting the pressure status of each tire in the vehicle, wherein the tire pressure probability identification model is trained based on training data carrying normal tire pressure labels and underpressure labels.

[0012] An embodiment of the third aspect of the present application provides an indirect tire pressure detection device, including: an acquisition module for acquiring training data carrying normal tire pressure labels and underpressure labels; a first processing module for performing signal processing on the wheel speed signal in the training data to obtain a processed wheel speed signal, and performing detrending processing on the processed wheel speed signal to obtain a wheel speed microscopic signal; an identification module for extracting wheel speed microscopic vibration characteristics from the wheel speed microscopic signal, using the wheel speed microscopic vibration characteristics to train a pre-constructed neural network, and obtaining a tire pressure probability identification model after the training is completed, and using the tire pressure probability identification model to identify the pressure status of each tire in the vehicle.

[0013] The fourth aspect of the present application provides an indirect tire pressure monitoring device, including: an acquisition module for acquiring the wheel speed signal of each tire in a vehicle; a second processing module for performing signal processing on the wheel speed signal to obtain a processed wheel speed signal, and performing detrending processing on the processed wheel speed signal to obtain a wheel speed micro-signal; an output module for inputting the wheel speed micro-signal into a completed tire pressure probability identification model, and outputting the pressure status of each tire in the vehicle, wherein the tire pressure probability identification model is trained based on training data carrying normal tire pressure labels and underpressure labels.

[0014] The fifth aspect of the present application provides a vehicle, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the indirect tire pressure monitoring method as described in the above embodiment.

[0015] A sixth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement the indirect tire pressure monitoring method as described in the above embodiment.

[0016] Therefore, this application has at least the following beneficial effects:

[0017] Instead of using frequency-domain analysis methods like Fourier transforms and autoregressive analysis, which are common in existing technologies, the pre-processed wheel speed signal is directly input into the deep learning model in the time domain. This approach eliminates the excessive human and financial costs associated with existing technologies, and possesses self-learning capabilities, meaning it can learn both online and offline using data input during driving. Tire pressure status is no longer determined by the resonance frequency calculated from the tire's circumferential stiffness, thus breaking away from the limitations of the tire's rigid ring model assumption. The generalization capabilities of the deep learning model effectively reduce interference from both external environmental factors and the tire's own internal factors, significantly lowering the false alarm rate of current indirect tire pressure monitoring systems. This approach eliminates the need for extensive real-time mathematical operations such as trigonometric multiplications in Fourier transforms, and allows the trained deep learning model to be ported to the ABS (Anti-lock Braking System) / ESC (Electronic Stability Controller) system, effectively reducing the computational load of the ABS and ESC systems themselves, making it more practical than existing indirect tire pressure monitoring technologies.

[0018] This solves the technical problems in related technologies such as the need to manually construct a tire rigid ring model, the lack of generalization capability for uncalibrated tires, and high cost investment.

[0019] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0021] Figure 1 Schematic diagram of a tire rigid ring model in the related art;

[0022] Figure 2 This is a flow chart of an indirect tire pressure monitoring method provided according to an embodiment of the present application;

[0023] Figure 3 A schematic diagram of a wheel speed sensor according to an embodiment of the present application;

[0024] Figure 4 A diagram showing a physical error extraction of a ring gear according to an embodiment of the present application;

[0025] Figure 5 A comparison diagram of wheel speed signal ring gear error elimination according to an embodiment of the present application;

[0026] Figure 6 A detrended wheel speed curve graph provided according to an embodiment of the present application;

[0027] Figure 7 A tire pressure state probability identification model provided according to an embodiment of the present application;

[0028] Figure 8 A flow chart of a model updating method according to an embodiment of the present application;

[0029] Figure 9 This is a flowchart for identifying the tire pressure status of a real vehicle provided in an embodiment of the present application;

[0030] Figure 10 This is a flow chart of an indirect tire pressure monitoring method according to one embodiment of the present application;

[0031] Figure 11 Schematic diagram of a carrier for an indirect tire pressure monitoring method according to an embodiment of the present application;

[0032] Figure 12 This is a flow chart of an indirect tire pressure monitoring method provided according to an embodiment of the present application;

[0033] Figure 13 This is an example diagram of an indirect tire pressure monitoring device provided according to an embodiment of the present application;

[0034] Figure 14 This is an example diagram of an indirect tire pressure monitoring device provided according to an embodiment of the present application;

[0035] Figure 15 A schematic structural diagram of a vehicle provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0036] The following describes in detail embodiments of the present application, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0037] At present, the theoretical basis of indirect tire pressure monitoring comes from the assumption of tire rigid ring model, which simplifies the tire model into Figure 1 The spring-damper system shown in Indicates the rotation angle of the tire around the central axis, k x 、 k t and k s Represents horizontal, circumferential, tread and suspension stiffness respectively. Among them, the physical quantity directly related to tire pressure is tire circumferential stiffness Theoretical analysis shows that changes in tire pressure will lead to The resonant frequency of the tire's circumferential motion is shifted, and the resonant frequency can be used to effectively monitor changes in tire pressure.

[0038] Rim Translation:

[0039]

[0040] Rim rotation:

[0041]

[0042] With cinch ring translation:

[0043]

[0044] With ring rotation:

[0045]

[0046] Where: x R and x B denote the displacements of the rim and belt ring respectively, and Represent the rotation angles of the rim and belt ring respectively.

[0047] choose is a state-space matrix, so the above formula can be rewritten as a state-space function, as shown below.

[0048]

[0049] M and K are the mass and stiffness matrices, respectively, as shown in the following equations.

[0050]

[0051]

[0052] Then the vibration frequency f related to the tire pressure is egi for:

[0053]

[0054] Frequency f egi This is the basis for the existing method to identify whether the tire is under-inflated. The tire rigid ring model is obtained by artificially simplifying the tire model. It cannot truly and comprehensively reflect the physical characteristics of the actual tire. Therefore, the method based on the tire rigid ring model itself has its own limitations in principle. In addition, other physical characteristics such as road roughness, ambient temperature, vertical load, etc. will still affect the tire's performance. This causes different degrees of impact, which is also the limitation of the artificially constructed tire rigid ring model and the fundamental reason why the existing indirect tire pressure monitoring technology is prone to false alarms.

[0055] Based on the above theoretical analysis, it is clear that the relevant technology cannot distinguish the difference between the normal tire pressure wheel speed signal and the underpressure wheel speed signal in the time domain. It can only extract the frequencies related to circumferential vibration in the wheel speed signal in the frequency domain through Fourier transform or autoregressive analysis for analysis.

[0056] The following describes the indirect tire pressure monitoring method, device, vehicle and storage medium of the embodiment of the present application with reference to the accompanying drawings. In view of the problem that the indirect tire pressure monitoring system in the related art mentioned in the above background technology does not have the generalization capability for uncalibrated tires and has high cost investment, the present application provides an indirect tire pressure monitoring method. In this method, there is no need to manually construct a tire rigid ring model, but instead uses a neural network algorithm to directly analyze the time domain wheel speed signal and the tire pressure. The system uses learning, modeling, and feature recognition to obtain a generalized model of the wheel speed signal, which can then be used to determine the tire pressure status. This solves the problems of related technologies that require manual construction of a tire rigid ring model, lack generalization capabilities for uncalibrated tires, and are costly.

[0057] Specifically, Figure 2 A flow chart of an indirect tire pressure monitoring method provided in an embodiment of the present application.

[0058] like Figure 2 As shown, the indirect tire pressure monitoring method includes the following steps:

[0059] In step S101 , training data carrying a normal tire pressure label and an under-pressure label is obtained.

[0060] In step S102 , the wheel speed signals in the training data are processed to obtain processed wheel speed signals, and the processed wheel speed signals are detrended to obtain wheel speed micro signals.

[0061] In an embodiment of the present application, signal processing is performed on the original wheel speed signal in the training data to obtain a processed wheel speed signal, including: extracting the angular physical error corresponding to each tooth in the wheel speed sensor ring gear; eliminating the angular physical error in the original wheel speed signal, and performing nonlinear interpolation to obtain a processed wheel speed signal.

[0062] It can be understood that the processing of the wheel speed signal in the embodiment of the present application mainly includes: (1) extracting the angle error corresponding to each tooth in the wheel speed sensor ring; (2) eliminating the angle physical error and performing nonlinear interpolation.

[0063] Step (1) is as follows: The wheel speed sensor schematic is as follows Figure 3 As shown, assuming that the gear ring has n teeth, theoretically, each tooth is evenly distributed on the gear ring at the same angle θ0, and the arc of two adjacent teeth can be calculated by the following formula.

[0064]

[0065] However, due to manufacturing errors, there is a Δθ in the arc between two adjacent teeth. i (k) The actual arc of the two adjacent teeth can be calculated by the following formula.

[0066] θ i (k) = 0 - Δθ i (k)

[0067] Figure 4 This is the result of identifying the actual physical error of the ring gear using recursive least squares with a forgetting factor (there are many identification methods, and other methods besides RLS can also be used). Each point on the horizontal axis represents each tooth on the ring gear (this wheel speed sensor has a total of 48 teeth per circle), and the vertical axis represents the angular error between the actual angle and the theoretical angle between two adjacent teeth on the ring gear.

[0068] The wheel speed signal WSP(t) after eliminating the physical error of the ring gear is calculated as follows, where Δt(k) is the time interval between the kth tooth and the k-1th tooth.

[0069]

[0070] The results are as follows Figure 5 As shown, the blue curve is the original wheel speed signal, and the red curve is the processed wheel speed signal. It can be seen that after the wheel speed preprocessing, the noise and variance of the wheel speed signal are significantly improved, which can improve the signal-to-noise ratio of the tire pressure feature in the wheel speed curve.

[0071] Step (2) is as follows: The original wheel speed signal is normalized using a nonlinear interpolation method to uniformly sample the data in the time domain. The interpolation frequency is 200 Hz, and the time interval between adjacent data points after interpolation is 5 ms. This step aims to eliminate interference from non-barometric factors and ensure the validity of the data input to the model. Nonlinear interpolation methods are relatively general and can be used with Bessel Kernel or other kernel functions, without limitation.

[0072] In an embodiment of the present application, the processed wheel speed signal is detrended to obtain a wheel speed micro-signal, including: obtaining the longitudinal vehicle speed at the same time as the processed wheel speed signal; and calculating the wheel speed micro-signal based on the wheel speed signal and the longitudinal vehicle speed at the same time.

[0073] It should be noted that the embodiment of the present application needs to perform detrending processing on the processed wheel speed signal to obtain the wheel speed micro-signal curve, the purpose of which is to remove the influence of low-frequency vehicle speed fluctuations and only retain the wheel speed micro-vibration characteristics, such as Figure 6 shown.

[0074] The specific steps of detrending are as follows: subtract the longitudinal speed V of the current vehicle from the wheel speed value WSP(t) obtained by nonlinear interpolation x (t) can get the detrended wheel speed size WSP FNL (t).

[0075] WSP FNL (t)=WSP(t)-V x (t).

[0076] In step S103, the wheel speed micro-vibration characteristics in the wheel speed micro-signal are extracted, and the pre-built neural network is trained using the wheel speed micro-vibration characteristics. After the training is completed, a tire pressure probability identification model is obtained, and the tire pressure probability identification model is used to identify the pressure status of each tire in the vehicle.

[0077] The tire pressure probability identification model includes an input layer, an output layer, and a hidden layer. The tire pressure probability identification model is also set with a training number, such as Figure 7 As shown in the figure, the input signal of the input layer is the wheel speed micro-signal, the output layer outputs normal tire pressure labels and underpressure labels, and the hidden layer contains multiple neurons.

[0078] It should be noted that the tire pressure probability identification model is essentially a binary classification neural network. The model outputs two tire pressure states: normal pressure and underpressure. The model structure and parameter settings are as follows:

[0079] (1) The input layer is the time domain wheel speed vibration signal (t, WSP FNL (t));

[0080] (2) The output layer is labeled [0, 1], where 0 is normal tire pressure and 1 is under pressure;

[0081] (3) The activation function is set to sigmoid;

[0082] (4) The model sets a hidden layer, which contains n ner neurons, the activation function is set to ReLU function;

[0083] (5) The loss function uses binary_crossentropy;

[0084] (6) The model uses the Adam optimization algorithm;

[0085] (7) The number of model training times is set to n train The number of training times will be adjusted in time according to the offline data analysis results. The more training data sets, the more accurate the recognition results.

[0086] In summary, the present application first collects the original wheel speed signals of normal tire pressure and under-pressure sent by the wheel speed sensor through the ABS / ESC system of the vehicle, and then calculates the wheel speed signal of the WSP after wheel speed preprocessing. FNL The input is presented to the neural network, which then performs model self-learning based on the labeled database. The network learns filters in its hidden layer that can be used to distinguish between tire pressure status categories, and ultimately trains a probabilistic identification model for tire pressure status.

[0087] In an embodiment of the present application, after using the tire pressure probability identification model to identify the actual state of the tire pressure, it also includes: obtaining the online wheel speed micro-signal during the vehicle operation process; using the wheel speed micro-signal to iteratively optimize the tire pressure probability identification model until the number of iterations reaches the set training number, and the optimized tire pressure probability identification model is obtained.

[0088] It is understandable that the embodiment of the present application can continuously acquire new wheel speed signal features online to perform iterative parameter optimization, or can manually inject a large amount of offline data into the model to update the neural network feature parameters, such as Figure 8 shown.

[0089] It should be noted that in actual engineering applications, the processed four-wheel speed time domain signals are input into the learned tire pressure state probability identification model, and the current four-wheel tire pressure state can be fed back, such as Figure 9 As shown in the figure, from the perspective of deep learning's generalization mechanism, this method is not limited by the specific parameters of the tire rigid ring model (external factors such as tire type, brand, and process). As long as the vehicle is operating normally, the model's characteristic parameters can be continuously iterated and optimized throughout the vehicle's life cycle. There is no need to manually modify the tire rigid ring model's related parameters; the algorithm automatically forms an accurate and extensive database.

[0090] Specifically, the indirect tire pressure monitoring method of the present application mainly includes three parts: wheel speed signal preprocessing, model self-learning and tire pressure status recognition, such as Figure 10 Among them, the model self-learning part is the core of the indirect tire pressure monitoring method.

[0091] It should be noted that the indirect tire pressure monitoring method described in this application is an indirect tire pressure monitoring method that does not rely on a pressure sensor. It is usually installed in the ABS or ESC of a passenger car and prompts the driver with tire pressure information in a visual or acoustic way when a wheel is under-pressured, such as Figure 11shown.

[0092] In summary, the embodiment of the present application does not require artificial construction of a tire rigid ring model, but uses a neural network algorithm to directly analyze the time domain wheel speed signal. This method uses learning, modeling, and feature recognition to obtain a generalized model of the wheel speed signal, thereby determining the tire pressure status. This process does not require the construction of a physical tire model or complex Fourier transform or autoregressive analysis to extract the tire's circumferential vibration frequency. Instead, it uses a neural network algorithm from deep learning as the core for overall modeling. This significantly differs from existing methods that directly calculate the tire's circumferential vibration frequency, replacing the existing indirect tire pressure monitoring system's method of extracting the tire's resonance frequency using Fourier transform or autoregressive algorithms on the wheel speed signal.

[0093] According to the indirect tire pressure monitoring method proposed in the embodiment of the present application, the frequency domain analysis means used in the prior art, such as Fourier transform and autoregressive analysis, are not used, and the pre-processed wheel speed signal is directly input into the deep learning model in the time domain; it does not require excessive manpower, expenses and other cost investments in the prior art, and the model has self-learning capabilities, that is, it can learn online during driving or input data for learning offline; it no longer relies on the resonance frequency obtained by calculating the tire circumferential stiffness to judge the tire pressure status, and can get rid of the limitations of the tire rigid ring model assumption. The generalization ability of the deep learning model can effectively reduce the interference of external environmental factors and tire factors itself, and greatly reduce the false alarm rate of the current indirect tire pressure monitoring system; it does not require a large number of real-time mathematical operations such as trigonometric function multiplication in Fourier transform, and the trained deep learning model can be transplanted into the ABS / ESC system, which can effectively reduce the computing load of the ABS and ESC systems themselves, and is more practical than the existing indirect tire pressure monitoring technology.

[0094] Figure 12 A flowchart of an indirect tire pressure monitoring method provided in an embodiment of the present application.

[0095] like Figure 12 As shown, the indirect tire pressure monitoring method includes the following steps:

[0096] In step S201 , the wheel speed signal of each tire in the vehicle is collected.

[0097] In step S202 , the wheel speed signal is processed to obtain a processed wheel speed signal, and the processed wheel speed signal is detrended to obtain a wheel speed micro signal.

[0098] In step S203, the wheel speed microscopic signal is input into the completed tire pressure probability identification model to output the pressure status of each tire in the vehicle, wherein the tire pressure probability identification model is trained based on training data carrying normal tire pressure labels and underpressure labels.

[0099] The methods for processing the original wheel speed signal, obtaining the wheel speed microscopic signal, and obtaining the tire pressure probability identification model have been described in the above embodiments and will not be repeated here.

[0100] In summary, the specific steps of the indirect tire pressure monitoring method are as follows: (1) collecting the wheel speed signal of each tire in the vehicle; (2) processing the original wheel speed signal and detrending it to obtain the wheel speed micro-signal; (3) inputting the wheel speed micro-signal into the tire pressure probability identification model and outputting the pressure status of each tire in the vehicle.

[0101] Next, an indirect tire pressure detection device according to an embodiment of the present application will be described with reference to the accompanying drawings.

[0102] Figure 13 、 14 Schematic diagram of an indirect tire pressure detection device according to an embodiment of the present application.

[0103] like Figure 13 As shown, the indirect tire pressure detection device 10 includes: an acquisition module 101 , a first processing module 102 and an identification module 103 .

[0104] Among them, the acquisition module 101 is used to obtain training data carrying normal tire pressure labels and underpressure labels; the first processing module 102 is used to perform signal processing on the wheel speed signal in the training data to obtain a processed wheel speed signal, and detrend the processed wheel speed signal to obtain a wheel speed micro-signal; the identification module 103 is used to extract the wheel speed micro-vibration characteristics in the wheel speed micro-signal, use the wheel speed micro-vibration characteristics to train a pre-constructed neural network, obtain a tire pressure probability identification model after the training is completed, and use the tire pressure probability identification model to identify the pressure status of each tire in the vehicle.

[0105] like Figure 14 As shown, the indirect tire pressure detection device 20 includes: an acquisition module 201 , a second processing module 202 and an output module 203 .

[0106] Among them, the acquisition module 201 is used to collect the wheel speed signal of each tire in the vehicle; the second processing module 202 is used to process the wheel speed signal to obtain a processed wheel speed signal, and detrend the processed wheel speed signal to obtain a wheel speed micro signal; the output module 203 is used to input the wheel speed micro signal into the completed tire pressure probability identification model, and output the pressure status of each tire in the vehicle, wherein the tire pressure probability identification model is trained based on training data carrying normal tire pressure labels and underpressure labels.

[0107] It should be noted that the aforementioned explanation of the embodiment of the indirect tire pressure monitoring method is also applicable to the indirect tire pressure monitoring device of this embodiment, and will not be repeated here.

[0108] According to the indirect tire pressure monitoring device proposed in the embodiment of the present application, the frequency domain analysis methods used in the prior art, such as Fourier transform and autoregressive analysis, are not used, and the pre-processed wheel speed signal is directly input into the deep learning model in the time domain; it does not require excessive manpower, expenses and other cost investments in the prior art, and the model has self-learning capabilities, that is, it can learn online during driving or input data for learning offline; it no longer relies on the resonance frequency obtained by calculating the tire circumferential stiffness to judge the tire pressure status, and can get rid of the limitations of the tire rigid ring model assumption. The generalization ability of the deep learning model can effectively reduce the interference of external environmental factors and tire factors itself, and greatly reduce the false alarm rate of the current indirect tire pressure monitoring system; it does not require a large number of real-time mathematical operations such as trigonometric function multiplication in Fourier transform, and the trained deep learning model can be transplanted into the ABS / ESC system, which can effectively reduce the computing load of the ABS and ESC systems themselves, and is more practical than the existing indirect tire pressure monitoring technology.

[0109] Figure 15 A schematic diagram of the structure of a vehicle provided in an embodiment of the present application. The vehicle may include:

[0110] Memory 1501 , processor 1502 , and computer programs stored in the memory 1501 and executable on the processor 1502 .

[0111] When the processor 1502 executes the program, the indirect tire pressure monitoring method provided in the above embodiment is implemented.

[0112] Furthermore, the vehicle further comprises:

[0113] The communication interface 1503 is used for communication between the memory 1501 and the processor 1502 .

[0114] The memory 1501 is used to store computer programs that can be run on the processor 1502 .

[0115] The memory 1501 may include a high-speed RAM (Random Access Memory) memory, and may also include a non-volatile memory, such as at least one disk memory.

[0116] If the memory 1501, processor 1502, and communication interface 1503 are implemented independently, the communication interface 1503, memory 1501, and processor 1502 can be connected to each other via a bus and communicate with each other. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 15 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0117] Optionally, in a specific implementation, if the memory 1501, the processor 1502 and the communication interface 1503 are integrated on a chip, the memory 1501, the processor 1502 and the communication interface 1503 can communicate with each other through an internal interface.

[0118] The processor 1502 may be a CPU (Central Processing Unit), or an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application.

[0119] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, which implements the above-mentioned indirect tire pressure monitoring method when executed by a processor.

[0120] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0121] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of technical features indicated. Thus, a feature specified as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of this application, "N" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0122] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present application belong.

[0123] It should be understood that various parts of the present application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiment, the N steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array, a field programmable gate array, etc.

[0124] Those skilled in the art will understand that all or part of the steps in the method of the above embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0125] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. An indirect tire pressure monitoring method, characterized in that: The following steps are involved: Obtain training data with normal tire pressure labels and underpressure labels; The wheel speed signal in the training data is processed to obtain a processed wheel speed signal, and the processed wheel speed signal is detrended to obtain a wheel speed micro-signal; the detrending process of the processed wheel speed signal to obtain the wheel speed micro-signal includes: obtaining the longitudinal vehicle speed at the same time as the processed wheel speed signal; calculating the wheel speed micro-signal based on the wheel speed signal and the longitudinal vehicle speed at the same time, wherein the detrending formula of the wheel speed micro-signal is: WSP FNL (t)=WSP(t)-V x (t), where WSP FNL (t) is the wheel speed micro signal, WSP(t) is the wheel speed signal, V x (t) is the longitudinal vehicle speed; A wheel speed micro-vibration feature in the time domain is extracted from the wheel speed micro-signal, and a pre-built neural network is trained using the wheel speed micro-vibration feature. After the training is completed, a tire pressure probability identification model is obtained, and the tire pressure probability identification model is used to identify the pressure state of each tire in the vehicle.

2. The method according to claim 1, characterized in that The tire pressure probability identification model includes an input layer, an output layer and a hidden layer, wherein the input signal of the input layer is a wheel speed microscopic signal, the output layer outputs a normal tire pressure label and an underpressure label, and the hidden layer contains multiple neurons.

3. The method according to claim 2, characterized in that The tire pressure probability identification model is further provided with a training number. After the tire pressure probability identification model is used to identify the actual state of the tire pressure, the method further includes: Obtain online wheel speed microscopic signals during vehicle operation; The tire pressure probability identification model is iteratively optimized using the wheel speed microscopic signal until the number of iterations reaches a set number of training times, thereby obtaining an optimized tire pressure probability identification model.

4. The method according to claim 1, wherein The processing of the original wheel speed signal in the training data to obtain a processed wheel speed signal includes: Extract the angular physical error corresponding to each tooth in the wheel speed sensor ring gear; The angular physical error in the original wheel speed signal is eliminated, and nonlinear interpolation is performed to obtain the processed wheel speed signal.

5. An indirect tire pressure monitoring method, characterized in that: The following steps are involved: Collect wheel speed signals of each tire in the vehicle; The wheel speed signal is processed to obtain a processed wheel speed signal, and the processed wheel speed signal is detrended to obtain a wheel speed micro-signal; the detrending process of the processed wheel speed signal to obtain the wheel speed micro-signal includes: obtaining the longitudinal vehicle speed at the same time as the processed wheel speed signal; calculating the wheel speed micro-signal based on the wheel speed signal and the longitudinal vehicle speed at the same time, wherein the detrending formula of the wheel speed micro-signal is: WSP FNL (t)=WSP(t)-V x (t), where WSP FNL (t) is the wheel speed micro signal, WSP(t) is the wheel speed signal, V x (t) is the longitudinal vehicle speed; The wheel speed microscopic signal is input into a completed tire pressure probability identification model to output the pressure status of each tire in the vehicle, wherein the tire pressure probability identification model is trained based on training data carrying normal tire pressure labels and underpressure labels.

6. An indirect tire pressure detection device, characterized in that: include: An acquisition module is used to obtain training data with normal tire pressure labels and underpressure labels; The first processing module is configured to perform signal processing on the wheel speed signal in the training data to obtain a processed wheel speed signal, and perform detrending processing on the processed wheel speed signal to obtain a wheel speed micro-signal; the detrending processing on the processed wheel speed signal to obtain the wheel speed micro-signal includes: obtaining the longitudinal vehicle speed at the same time as the processed wheel speed signal; calculating the wheel speed micro-signal based on the wheel speed signal and the longitudinal vehicle speed at the same time, wherein the detrending formula of the wheel speed micro-signal is: WSP FNL (t)=WSP(t)-V x (t), where WSP FNL (t) is the wheel speed micro signal, WSP(t) is the wheel speed signal, V x (t) is the longitudinal vehicle speed; The identification module is used to extract the wheel speed micro-vibration characteristics in the time domain from the wheel speed micro-signal, use the wheel speed micro-vibration characteristics to train a pre-built neural network, obtain a tire pressure probability identification model after the training is completed, and use the tire pressure probability identification model to identify the pressure state of each tire in the vehicle.

7. An indirect tire pressure monitoring device, characterized in that: include: An acquisition module, used to collect the wheel speed signal of each tire in the vehicle; The second processing module is configured to perform signal processing on the wheel speed signal to obtain a processed wheel speed signal, and perform detrending processing on the processed wheel speed signal to obtain a wheel speed micro-signal; the detrending processing on the processed wheel speed signal to obtain the wheel speed micro-signal includes: obtaining the longitudinal vehicle speed at the same time as the processed wheel speed signal; calculating the wheel speed micro-signal based on the wheel speed signal and the longitudinal vehicle speed at the same time, wherein the detrending formula of the wheel speed micro-signal is: WSP FNL (t)=WSP(t)-V x (t), where WSP FNL (t) is the wheel speed micro signal, WSP(t) is the wheel speed signal, V x (t) is the longitudinal vehicle speed; An output module is configured to input the wheel speed microscopic signal into a completed tire pressure probability identification model and output the pressure status of each tire in the vehicle, wherein the tire pressure probability identification model is trained based on training data carrying normal tire pressure labels and underpressure labels.

8. A vehicle, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the indirect tire pressure monitoring method according to any one of claims 1 to 5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the indirect tire pressure monitoring method according to any one of claims 1 to 5.

Images