Tire longitudinal speed estimation method, device, system and equipment

Abstract

The invention provides a tire longitudinal speed estimation method, device, system and equipment. The method comprises the following steps: determining key feature points of a strain response signal in a grounding print area according to tread strain information; inputting the key feature points into a preset tire longitudinal speed prediction model to obtain a first tire longitudinal speed; determining the longitudinal slip rate of the tire according to the tread strain information and a preset slip rate prediction model; calculating a second tire longitudinal speed of the tire according to the longitudinal slip rate; and carrying out weighted fusion on the first tire longitudinal speed and the second tire longitudinal speed to obtain the tire longitudinal speed. According to the method, the first tire longitudinal speed is obtained directly through the key feature points of the strain response signals in the grounding print area; indirectly calculating a second tire longitudinal speed of the tire through the longitudinal slip rate of the tire; and finally, carrying out weighted fusion on the first tire longitudinal speed and the second tire longitudinal speed so as to accurately estimate the tire longitudinal speed through two different modes.

Description

Technical Field

[0001] The present application relates to the field of vehicle technology, and in particular to a tire longitudinal velocity estimation method, device, system and equipment. Background Art

[0002] With the rapid development of the automobile industry and the continuous advancement of intelligent technology, the safety and stability of vehicles have become increasingly important. As one of the core indicators for measuring vehicle dynamic performance, accurate estimation of vehicle speed has become an indispensable part of the intelligent transportation system. It not only plays a key role in emergency accident warning, efficient traffic flow control and accurate route planning, but also is the cornerstone for improving the efficiency and safety of the overall transportation system. However, in the face of complex and changing road environments and the inherent limitations of vehicle sensors, achieving accurate estimation of vehicle speed has always been a technical challenge. In view of this, it is particularly urgent and far-reaching to optimize the vehicle speed estimation algorithm to enhance vehicle stability control, optimize driving trajectory planning and promote the innovation of active safety systems.

[0003] There are many methods for vehicle speed estimation on the market, such as GPS-based methods. However, low-cost GPS has a low update frequency, while high-precision, high-frequency GPS is too expensive. In order to improve the accuracy and stability of vehicle speed estimation, a method combining Kalman filtering with a long short-term memory network (LSTM) model is introduced. By fusing the vehicle's sensor data and combining it with a time series prediction model, the vehicle's speed can be predicted more accurately. However, although this method can estimate the vehicle speed to a certain extent, its accuracy is affected by multiple factors such as model parameter errors, tire nonlinear characteristics, and road conditions, making it difficult to accurately estimate the vehicle speed. Summary of the invention

[0004] The purpose of the present application is to provide a tire longitudinal velocity estimation method, device, system and equipment, aiming to solve the technical problem that it is difficult to accurately estimate the tire longitudinal velocity.

[0005] To achieve the above object, the present application proposes a tire longitudinal speed estimation method, the tire longitudinal speed estimation method comprising:

[0006] Collect tire tread strain information through PVDF sensors;

[0007] Determining key characteristic points of a strain response signal in a contact patch region according to the tread strain information;

[0008] Inputting the key feature points into a preset tire longitudinal speed prediction model to obtain a first tire longitudinal speed;

[0009] Determining the longitudinal slip rate of the tire according to the tread strain information and a preset slip rate prediction model;

[0010] calculating a second tire longitudinal velocity of the tire based on the longitudinal slip rate;

[0011] The first tire longitudinal velocity and the second tire longitudinal velocity are weightedly fused to obtain the tire longitudinal velocity.

[0012] Optionally, determining the key characteristic points of the strain response signal in the contact patch area according to the tread strain information includes:

[0013] determining a contact patch area of ​​the tire using the tread strain information;

[0014] Filtering the strain response signal in the contact patch area to obtain a filtered first strain curve;

[0015] Key feature points are determined according to the filtered first strain curve.

[0016] Optionally, determining key feature points according to the filtered first strain curve includes:

[0017] Performing first-order differentiation and second-order differentiation on the filtered first strain curve to obtain a first-order differential curve and a second-order differential curve;

[0018] The leading edge maximum value, trailing edge minimum value, and phase difference between the leading edge maximum value and the trailing edge minimum value of the first-order differential curve, and the leading edge maximum value, trailing edge maximum value, and phase difference between the leading edge maximum value and the trailing edge maximum value of the second-order differential curve are extracted as the key feature points.

[0019] Optionally, determining the contact patch area of ​​the tire by using the tread strain information comprises:

[0020] Extracting a second strain curve at the center of the tread from the tread strain information;

[0021] The second strain curve is analyzed to determine a contact patch area of ​​the tire.

[0022] Optionally, the step of performing weighted fusion of the first tire longitudinal velocity and the second tire longitudinal velocity to obtain the tire longitudinal velocity comprises:

[0023] Acquire a first probability distribution of the preset tire longitudinal speed prediction model and a second probability distribution of the preset slip rate prediction model;

[0024] Determining a weight coefficient of a weighted fusion process according to the first probability distribution and the second probability distribution;

[0025] The first tire longitudinal velocity and the second tire longitudinal velocity are weightedly fused according to the weight coefficient to obtain the tire longitudinal velocity.

[0026] Optionally, the determining the longitudinal slip rate of the tire according to the tread strain information and a preset slip rate prediction model comprises:

[0027] Acquiring tire test data, wherein the tire test data at least includes tread strain information and corresponding slip rate collected by a PVDF sensor;

[0028] Establishing a tire slip rate prediction model based on the tire test data by using a Gaussian process regression algorithm;

[0029] Accordingly, determining the longitudinal slip rate of the tire according to the tread strain information and a preset slip rate prediction model includes:

[0030] The tread strain information is input into the tire slip rate prediction model to determine the longitudinal slip rate of the tire.

[0031] Optionally, calculating a second tire longitudinal velocity of the tire according to the longitudinal slip rate comprises:

[0032] Obtaining the rotation speed and effective rolling radius of the tire;

[0033] A second tire longitudinal speed of the tire is calculated according to the longitudinal slip rate, the rotation speed, and the effective rolling radius.

[0034] In addition, to achieve the above object, the present invention also provides a tire longitudinal speed estimation device, comprising:

[0035] An information acquisition module, used for collecting tire tread strain information through a PVDF sensor;

[0036] A feature point confirmation module, used to determine key feature points of a strain response signal in a contact patch area according to the tread strain information;

[0037] A speed determination module, used for inputting the key feature point into a preset tire longitudinal speed prediction model to obtain a first tire longitudinal speed;

[0038] A slip rate determination module, used to determine the longitudinal slip rate of the tire according to the tread strain information and a preset slip rate prediction model;

[0039] The speed determination module is further used to calculate a second tire longitudinal speed of the tire according to the longitudinal slip rate;

[0040] The speed determination module is further configured to perform weighted fusion of the first tire longitudinal speed and the second tire longitudinal speed to obtain the tire longitudinal speed.

[0041] In addition, to achieve the above-mentioned purpose, the present invention also proposes a tire longitudinal velocity estimation system, the system comprising: a strain sensor, a slip ring, a charge amplifier and a data acquisition module;

[0042] The strain sensor is arranged on the inner side of the tire, the strain sensor is connected to the charge amplifier via a slip ring, and the charge amplifier is connected to the data acquisition module.

[0043] In addition, to achieve the above-mentioned purpose, the present invention also proposes a tire longitudinal speed estimation device, which includes: a memory, a processor, a tire longitudinal speed estimation program stored in the memory and executable on the processor, and the tire longitudinal speed estimation system, wherein the tire longitudinal speed estimation program implements the steps of the tire longitudinal speed estimation method when executed by the processor.

[0044] The present application proposes a tire longitudinal velocity estimation method, device, system and equipment, which collects tire tread strain information through a PVDF sensor; determines the key characteristic points of the strain response signal in the contact patch area according to the tread strain information; inputs the key characteristic points into a preset tire longitudinal velocity prediction model to obtain a first tire longitudinal velocity; determines the longitudinal slip rate of the tire according to the tread strain information and a preset slip rate prediction model; calculates the second tire longitudinal velocity of the tire according to the longitudinal slip rate; and weightedly fuses the first tire longitudinal velocity with the second tire longitudinal velocity to obtain the tire longitudinal velocity. The present application directly obtains the first tire longitudinal velocity through the key characteristic points of the strain response signal in the contact patch area; indirectly calculates the second tire longitudinal velocity of the tire through the longitudinal slip rate of the tire; and finally weightedly fuses the first tire longitudinal velocity with the second tire longitudinal velocity, thereby accurately estimating the tire longitudinal velocity in two different ways. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a flow chart of Embodiment 1 of the tire longitudinal velocity estimation method of the present application;

[0046] Figure 2 This is a schematic diagram of the configuration of the PVDF sensor in Example 1 of the present application;

[0047] Figure 3 is a variation curve of tread strain information collected by the PVDF sensor in Example 1 of the present application;

[0048] Figure 4 This is a flow chart of Embodiment 2 of the tire longitudinal velocity estimation method of the present application;

[0049] Figure 5It is the variation curve of the strain curve, the first differential of the strain curve and the second differential of the strain curve in the second embodiment of the present application;

[0050] Figure 6 It is the characteristic change curve of the strain response of the tread center under the composite working condition of ±10% slip rate at a side slip angle of -8 degrees;

[0051] Figure 7 It is the characteristic change curve of the strain response of the tread center under the composite working condition of ±10% slip rate at +8 degrees side slip angle;

[0052] Figure 8 is the curve of the peak value of each area changing with the longitudinal velocity;

[0053] Fig. 9 is the variation curve of each contact patch length with the longitudinal speed;

[0054] Fig.10 This is a flow chart of Embodiment 3 of the tire longitudinal velocity estimation method of the present application;

[0055] Fig.11 This is a flow chart of a fourth embodiment of a tire longitudinal velocity estimation method of the present application;

[0056] Fig.12 It is the curve change of the weighted fusion estimation scheme of tire longitudinal velocity under pure sideslip condition;

[0057] Fig.13 It is the curve change of the weighted fusion estimation scheme of tire longitudinal velocity under pure longitudinal slip condition;

[0058] Fig.14 The curve variation of the weighted fusion estimation scheme of tire longitudinal velocity under complex working conditions;

[0059] Fig.15 This is a schematic diagram of the structure of the tire longitudinal speed estimation system of this application. DETAILED DESCRIPTION

[0060] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0061] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0062] Embodiment 1

[0063] Reference Figure 1 , Figure 1 This is a flow chart of the first embodiment of the tire longitudinal velocity estimation method of the present application. Figure 1 Embodiment 1 of the tire longitudinal velocity estimation method of the present application is proposed.

[0064] In this embodiment, the tire longitudinal speed estimation method includes:

[0065] Step S10: collecting tire tread strain information through a PVDF sensor;

[0066] It is understandable that existing tire speed estimation methods have certain limitations in practical applications. Although the estimation method based on the vehicle dynamics model can estimate the vehicle speed to a certain extent, its accuracy is affected by many factors such as model parameter errors, tire nonlinear characteristics and road conditions. Although sensor fusion technology can integrate information from multiple sensors to improve the estimation accuracy, these methods have some problems, such as errors caused by acceleration bias and noise, which will lead to inaccurate tire longitudinal speed. Tire longitudinal speed, as a core parameter in vehicle dynamics, is the speed component that describes the tire along its driving direction during driving, which can be regarded as the driving speed at the tire's rotation center. The longitudinal speed usually described is generally specified as the speed component along the vehicle's driving direction at the vehicle's center of mass, while the tire longitudinal speed of the non-steering wheel is equivalent to the longitudinal speed at the vehicle's center of mass.

[0067] It should be understood that PVDF sensor is a piezoelectric polymer material sensor. Its unique piezoelectric effect and excellent mechanical properties have been widely used in the sensor field. PVDF sensor has become an ideal sensing element for tires under extreme working conditions due to its high sensitivity, fast response and good weather resistance.

[0068] It should be noted that in this embodiment, the PVDF sensor is set on the tire of the vehicle, and the data presented by the tire can be collected during the operation of the vehicle. In order to collect tire information extensively and accurately, multiple PVDF sensors can be set; each PVDF sensor can be set on the inner surface of the tire. Figure 2 ,exist Figure 2 Three PVDF sensors are set in the tire, and the PVDF sensors can be evenly set on the inner surface of the tire to collect tread strain information at different positions. Among them, the tread strain information is the information about the change of tire state caused by the vehicle adapting to the driving state during driving. The tread strain information can include the pressure information, friction information, deformation information, etc. of the tire. The tread strain information can reflect the state parameters of the vehicle during operation, such as the vehicle's driving speed, the tire's slip rate, the vehicle's acceleration, etc.

[0069] In a specific implementation, when the vehicle is driving, the PVDF sensor disposed inside the tire can be used to collect the tire tread strain information. Figure 2In the figure, three PVDF sensors are set up. The PVDF sensor on the left is used to collect the tread strain information on the left side of the tire, the PVDF sensor on the right is used to collect the tread strain information on the right side of the tire, and the middle PVDF sensor is used to collect the tread strain information at the center line of the tire.

[0070] Step S20: determining key characteristic points of a strain response signal in a contact patch region according to the tread strain information.

[0071] It should be noted that the contact patch area is the area where the tire contacts the ground. During the rotation of the tire, the contact patch area is constantly changing; the tread strain information collected by the PVDF sensor in the contact patch area is completely different from the tread strain information collected when it is not in the contact patch area. Figure 3 ,exist Figure 3 As the tire rotation angle changes, the tread strain information collected by the PVDF sensor will also change. The strain response signal is the information that the tire responds to the change in the vehicle's driving state. The strain response signal is a manifestation of the tread strain information. The key feature points are points that can reflect the characteristic information during the vehicle's driving process, such as points that can determine the longitudinal speed of the tire, points that can determine the tire slip rate, etc. Since the key feature points of the strain response signal are mainly generated by the operating state of the tire, the operating state of the tire can be calculated using the key feature points of the strain response signal.

[0072] It can be understood that when the area where the PVDF sensor is set is not in the contact patch area, the area where the PVDF sensor is located will not contact the ground. At this time, the tread strain information collected by the PVDF sensor is only the information of the tire in the air, and the movement state of the tire mainly establishes friction with the ground to make the vehicle run. Therefore, the strain response signal of the contact patch area can show the running state of the tire.

[0073] Therefore, in a specific implementation, the strain response signal in the contact patch area can be extracted from the tread strain information, and then key feature points related to the longitudinal speed of the tire, such as the peak point of the strain response signal, the waveform inflection point, etc., can be extracted from the strain response signal.

[0074] Step S30: inputting the key feature points into a preset tire longitudinal speed prediction model to obtain a first tire longitudinal speed.

[0075] It is understandable that if the tire longitudinal speed prediction model uses Gaussian process regression as the core algorithm and is based on a model trained in a machine learning framework, the model can effectively predict the tire longitudinal speed. The core idea of ​​the model is to use machine learning algorithms to conduct in-depth learning of the tire tread strain response characteristics, and then directly infer the tire longitudinal speed based on the collected tread strain information. Gaussian process regression is mainly aimed at the dynamic distribution of each feature point, and then a preset tire longitudinal speed prediction model is constructed based on the dynamic distribution.

[0076] It should be noted that the first tire longitudinal speed is the tire longitudinal speed directly predicted by the preset tire longitudinal speed prediction model based on the key feature points. In a specific implementation, the extracted key feature points can be input into the preset tire longitudinal speed prediction model, and the preset tire longitudinal speed prediction model predicts the first tire longitudinal speed of the tire based on the information of each key feature point.

[0077] Step S40: determining the longitudinal slip rate of the tire according to the tread strain information and a preset slip rate prediction model.

[0078] It should be understood that slip refers to the relative slip between the tire and the ground, and the longitudinal slip rate refers to the degree of slip in the rolling direction of the tire. During the operation of the tire, the longitudinal slip rate of the tire can be reflected by the characteristic information collected by the PVDF sensor. The preset slip rate prediction model also uses Gaussian process regression as the core algorithm based on the model trained by the machine learning framework. The preset slip rate prediction model can effectively predict the longitudinal slip rate of the tire.

[0079] It is understandable that, when the tread strain information is collected, the features related to the longitudinal changes in the tread strain information can be input into the preset slip rate prediction model, and the preset slip rate prediction model can determine the longitudinal slip rate of the tire according to the input features. For example, the features corresponding to the tire strain are input into the preset slip rate prediction model, and the preset slip rate prediction model determines the longitudinal slip rate that the tire will generate according to the feature changes reflected in the longitudinal direction of the tire.

[0080] Step S50: Calculating a second tire longitudinal speed of the tire according to the longitudinal slip rate.

[0081] It should be understood that the longitudinal slip rate of the tire is a manifestation of the longitudinal running characteristics of the tire, and is directly related to the longitudinal motion parameters of the tire. For example, when the longitudinal acceleration of the tire is greater, the strain characteristics of the tire tread are greater, and the possibility of the tire slipping is greater, and when slipping occurs, the tire slip rate is greater. When the longitudinal slip rate of the tire is determined, the longitudinal speed of the tire can be analyzed based on the longitudinal slip rate of the tire.

[0082] It should be noted that, when all parameters of the tire are determined, there is a certain mapping relationship between the longitudinal slip rate of the tire and the longitudinal speed of the tire. Therefore, when the longitudinal slip rate of the tire is determined, the longitudinal speed of the tire can be directly calculated based on the longitudinal slip rate. Among them, the second tire longitudinal speed is the tire longitudinal speed obtained after calculating the tire slip rate. The method for obtaining the second tire longitudinal speed is completely different from the method for obtaining the first tire longitudinal speed.

[0083] Step S60: performing weighted fusion of the first tire longitudinal velocity and the second tire longitudinal velocity to obtain a tire longitudinal velocity.

[0084] It is understandable that when the longitudinal speeds of the two tires are determined by two methods, the accuracy of the two acquisition methods can be determined according to the difference between the two acquisition methods. In the weighted fusion process, the proportion of the longitudinal speeds of the two tires can be determined according to the accuracy of the tire longitudinal speed acquisition.

[0085] The preset tire longitudinal speed prediction model mainly focuses on the dynamic distribution of feature points during the actual prediction of the tire longitudinal speed and the preset slip rate prediction model during the determination of the tire longitudinal slip rate. The more concentrated the feature points of the dynamic distribution are, the more accurate the estimated tire longitudinal speed is and the greater the corresponding weighted proportion is.

[0086] In a specific implementation, the first tire longitudinal speed and the second tire longitudinal speed are directly weighted and fused to obtain the tire longitudinal speed. Further, in the weighting process, the weighted proportion can be determined according to the dynamic distribution preset by the model, and then the first tire longitudinal speed and the second tire longitudinal speed are weighted and fused to obtain the tire longitudinal speed according to the weighted proportion of each tire longitudinal speed.

[0087] It can be seen from the above embodiments that the present application directly obtains the first tire longitudinal velocity through the key characteristic points of the strain response signal in the contact patch area; indirectly calculates the second tire longitudinal velocity of the tire through the longitudinal slip rate of the tire; and finally weightedly fuses the first tire longitudinal velocity and the second tire longitudinal velocity, thereby accurately estimating the tire longitudinal velocity in two different ways.

[0088] Embodiment 2

[0089] Reference Figure 4 , Figure 4 The flowchart of the second embodiment of the method for estimating the longitudinal speed of a tire of the present application is shown in FIG. Based on the first embodiment of the method for estimating the longitudinal speed of a tire of the present application, the second embodiment of the method for estimating the longitudinal speed of a tire of the present application is proposed.

[0090] In the second embodiment, the step S20 includes:

[0091] Step S201: determining the contact patch area of ​​the tire through the tread strain information.

[0092] It should be understood that in the process of determining the key feature points, in order to avoid feature interference of the non-contact footprint area, it is necessary to first identify the contact footprint area, so as to eliminate the influence of tread strain information in the non-contact footprint area on the determination of the key feature points.

[0093] It can be understood that the tread strain information collected when the area where the PVDF sensor is located is the contact patch area is different from the tread strain information collected when the area where the PVDF sensor is located is not the contact patch area. Therefore, the contact patch area of ​​the tire can be determined based on the tread strain information collected by the PVDF sensor during a complete cycle.

[0094] The step S201 specifically includes:

[0095] Step S2011: extracting a second strain curve at the center of the tread from the tread strain information.

[0096] It should be understood that the center of the tread is the center of the tire surface, that is, Figure 2 The position of the PVDF sensor in the middle of the tire. Due to different working conditions, the strain characteristics of the tire on both sides and the center are not the same. For example, in the case of turning left, the steering force received by the left side is completely different from the steering force received by the right side. Under different working conditions, the characteristics at the center of the tread will not undergo significant changes due to different working conditions. Therefore, in this embodiment, the tread strain information at the center of the tread can be used to more accurately extract key feature points. The second strain curve is a curve corresponding to the tread strain information at the center of the tread. The second strain curve can reflect the changes in characteristic variables during the rotation of the tire.

[0097] Step S2012: Analyze the second strain curve to determine the contact patch area of ​​the tire.

[0098] It is understandable that, when the second strain curve is determined, the second strain curve may be analyzed to determine the change of characteristic variables during the rotation of the tire, and then the contact patch area may be determined according to the change of characteristic variables.

[0099] In the specific identification process, the tread strain information mainly comes from the strain signal (second strain curve) generated when the sensor installation area enters the contact patch area. Effective extraction of the second strain curve of the contact patch area is a necessary process for determining the key feature points.

[0100] It should be understood that the tire contact patch area can be obtained using the rotation angle output by the tire encoder, but the contact patch area needs to be calibrated in advance, and the boundary points and boundary points of the tire contact patch area can be determined by the footprint recognition method. Then the center point of the contact patch area is determined based on the two boundary points, and the center of the first contact patch is calibrated as 0 degrees. The subsequent data collection work is based on the center point A, and the tread strain information is collected within the range of ±180 degrees front and back.

[0101] When determining the contact patch area, it can be determined based on information such as the peak value or slope of the strain response signal corresponding to the tread strain information in the entire cycle (within the tire rolling range of ±180 degrees from the contact patch center). Figure 5 , Figure 5 Three methods for determining the ground contact patch area are recorded in the paper. Method 1 is to determine the boundary values ​​of the peak curve on both sides of the center point A (the trough value of the second response curve), and the area between the boundary values ​​on both sides is the ground contact patch area. The second method is to identify the ground contact patch based on the curve obtained by the first differential of the strain curve, and determine the ground contact patch area by identifying the area between the maximum value of the leading edge and the minimum value of the trailing edge; Method 3 shows the determination of the ground contact patch area based on the area between the two peaks of the second differential curve of the strain curve.

[0102] In addition, in order to avoid the possible noise interference in the tread strain information that will seriously affect the quality of the differential curve, the collected strain signal needs to be low-pass filtered before differentiation, and then each contact patch area signal is differentiated separately to more accurately determine the contact patch area. During the filtering process, signals with a frequency greater than 200Hz can be filtered out.

[0103] Step S202: filtering the strain response signal in the contact patch area to obtain a filtered first strain curve.

[0104] It should be understood that, when the contact patch area is determined, the strain response signal in the contact patch area can be extracted from the tread strain information. Considering that the possible interference of the strain response signal has a certain impact on the accuracy of the identification of the key feature points, in this embodiment, the strain response signal in the contact patch area can also be filtered first to eliminate other signal interference.

[0105] The first strain curve is a strain response signal obtained by filtering the strain response signal in the contact patch area, and the first strain curve can intuitively reflect the characteristic variables in the contact patch area.

[0106] Step S203: determining key feature points according to the filtered first strain curve.

[0107] It is understandable that the first strain curve is a characteristic variable change curve in the contact patch area, and each point on the first strain curve related to longitudinal operation can be used as a key characteristic point, such as the peak point, trough point and other points related to longitudinal operation of the first strain curve.

[0108] The step S203 specifically includes:

[0109] Step S2031: performing first-order differentiation and second-order differentiation on the filtered first strain curve to obtain a first-order differential curve and a second-order differential curve;

[0110] Step S2032: extract the leading edge maximum value, trailing edge minimum value, and phase difference between the leading edge maximum value and the trailing edge minimum value of the first-order differential curve, and the leading edge maximum value, trailing edge maximum value, and phase difference between the leading edge maximum value and the trailing edge maximum value of the second-order differential curve as the key feature points.

[0111] It can be understood that the first-order differential curve and the second-order differential curve are curves obtained after the first-order differential and the second-order differential of the filtered first strain curve, respectively. The first-order differential curve can reflect the slope change of the first strain curve, and the second-order differential curve can reflect the slope change of the first-order differential curve. The slope change and the peak area are both affected by the longitudinal running state of the tire. The maximum value of the leading edge is the maximum value in front of the center point of the tire ground contact area, the minimum value of the trailing edge is the minimum value behind the center point of the tire ground contact area, and the maximum value of the trailing edge is the maximum value behind the center point of the tire ground contact area. The phase difference between the maximum value of the leading edge and the minimum value of the trailing edge refers to the corresponding ground contact length in the first-order differential curve, that is, the second ground contact length CP2; the maximum value of the trailing edge and the phase difference between the maximum value of the leading edge and the maximum value of the trailing edge refer to the corresponding ground contact length in the second-order differential curve, that is, the third ground contact length CP3.

[0112] Reference Figure 6 and Figure 7 , Figure 6 The characteristic curve of the tread center strain response under the combined working condition of ±10% slip rate at -8 degrees side slip angle. Figure 7 The characteristic curve of the tread center strain response under the combined working condition of ±10% slip rate at +8 degrees side slip angle. Figure 6 and Figure 7In the figure, the first strain curve, the first order differential curve of the first strain curve and the second order differential curve of the first strain curve are included. There are 6 regional peaks: the leading edge minimum value R1 of the first strain curve, the trailing edge minimum value R2 of the first strain curve, the leading edge maximum value R3 of the first order differential curve, the trailing edge minimum value R4 of the first order differential curve, the leading edge maximum value R5 of the second order differential curve, and the trailing edge maximum value R6 of the second order differential curve; and the first ground contact patch length CP1, the second ground contact patch length CP2 and the third ground contact patch length CP3 corresponding to the three ground contact patch areas are also included.

[0113] Reference Figure 8 and Fig. 9 , Figure 8 is the curve of the peak value of each region changing with the longitudinal velocity, Fig. 9 is the curve of the change of each contact patch length with the longitudinal speed. Figure 8 In the case of , the peak value and longitudinal velocity of each region maintain an approximately linear response law. Figure 7 In the figure, the second contact patch length CP2 is approximately linearly positively correlated with the speed, while the third contact patch length CP3 is approximately linearly negatively correlated with the speed, and the first contact patch length CP1 is not stable with the speed; therefore, the second contact patch length CP2 and the third contact patch length CP3 are more stable than the first contact patch length CP1.

[0114] For the first contact patch length CP1, from Figure 6 and Figure 7 It can be observed that the rotation angle corresponding to the leading edge minimum value R1 of the first strain curve and the trailing edge minimum value R2 of the first strain curve in the free rolling condition is symmetrical. Due to the influence of the lateral deviation and longitudinal slip, the leading edge minimum value R1 of the first strain curve changes, which affects the calculation of the first contact patch length CP1. By observing the regional characteristics of the leading edge maximum value R3 of the first-order differential curve, the trailing edge minimum value R4 of the first-order differential curve, the leading edge maximum value R5 of the second-order differential curve, and the trailing edge maximum value R6 of the second-order differential curve of the center strain first-order differential curve and the second-order differential curve, it can be seen that the central strain first-order differential curve and the second-order differential curve have the following characteristics: Figure 6 and Figure 7The composite working conditions all maintain distribution characteristics similar to free rolling. Therefore, the first contact patch length CP1, the leading edge minimum value R1 of the first strain curve, and the trailing edge minimum value R2 of the first strain curve will be affected by tire lateral deviation and longitudinal slip. The calculation of the tire longitudinal speed based on the first contact patch length CP1, the leading edge minimum value R1 of the first strain curve, and the trailing edge minimum value R2 of the first strain curve will also be affected by tire lateral deviation and longitudinal slip, thereby causing the first tire longitudinal speed to be inaccurate. Therefore, in this embodiment, the leading edge maximum value R3 of the first-order differential curve, the trailing edge minimum value R4 of the first-order differential curve, the leading edge maximum value R5 of the second-order differential curve, the trailing edge maximum value R6 of the second-order differential curve, the second contact patch length CP2, and the third contact patch length CP3 can be selected as key feature points.

[0115] Embodiment 3

[0116] Reference Fig.10 , Fig.10 The flowchart of the third embodiment of the method for estimating the longitudinal speed of a tire of the present application is shown in FIG. Based on the first embodiment or the second embodiment of the method for estimating the longitudinal speed of a tire of the present application, the third embodiment of the method for estimating the longitudinal speed of a tire of the present application is proposed.

[0117] In the third embodiment, the step S60 includes:

[0118] Step S601: obtaining a first probability distribution of prediction points in the preset tire longitudinal speed prediction model and a second probability distribution of prediction points in the preset slip rate prediction model.

[0119] It should be understood that the preset tire longitudinal speed prediction model and the preset slip rate prediction model are both constructed using Gaussian process regression, the principle of which is to obtain the relationship between the input source feature and the output feature based on the overall probability distribution of each prediction point or feature point. In the process of constructing the model using Gaussian process regression, there may be some prediction points or feature points that are not in the overall probability distribution, such as prediction points with noise.

[0120] The first probability distribution is the probability distribution of the key feature points in the preset tire longitudinal speed prediction model, that is, the confidence interval of the first tire longitudinal speed in the method of directly estimating the vehicle longitudinal speed. The second probability distribution is the probability distribution of the output feature of the preset slip rate prediction model, that is, the confidence interval of the slip rate estimation value in the method of indirectly estimating the second tire longitudinal speed through the slip rate.

[0121] In the specific implementation, since the Gaussian process regression algorithm is a non-parametric regression method based on Bayesian theory, the algorithm is good at handling complex nonlinear relationships and exhibits strong generalization capabilities. In addition, the algorithm can also calculate the probability distribution of the prediction point while predicting the output characteristics. Therefore, in the case where the preset tire longitudinal speed prediction model and the preset slip rate prediction model are adopted in the present application, the first probability distribution of the prediction point in the preset tire longitudinal speed prediction model and the second probability distribution of the prediction point in the preset slip rate prediction model can be directly calculated by the Gaussian process regression algorithm.

[0122] Step S602: Determine a weight coefficient of a weighted fusion process according to the first probability distribution and the second probability distribution.

[0123] It should be noted that the weight coefficient reflects the numerical value of the weight of the first tire longitudinal speed and the second tire longitudinal speed. In the weighted fusion process, it is necessary to allocate the weight ratio of the first tire longitudinal speed and the second tire longitudinal speed according to the weight coefficient, and then obtain the accurate tire longitudinal speed. For example, when the first probability distribution is 95% and the second distribution probability is also 95%, it is determined that the accuracy of the first tire longitudinal speed and the second tire longitudinal speed is the same, and the two contribute the same to the accuracy of the tire longitudinal speed. At this time, the weight coefficient can be 0.5.

[0124] In the specific calculation process, the calculation of the weight coefficient follows the minimum mean square error criterion, which can be calculated using the formula: G i =R1 i / (R1 i +R2 i ), where G i is the weight coefficient, and its value range is between [0,1], R1 i is the first probability distribution, R2 i is the second probability distribution.

[0125] Step S603: performing weighted fusion of the first tire longitudinal velocity and the second tire longitudinal velocity according to the weight coefficient to obtain the tire longitudinal velocity.

[0126] It is understandable that, when the weight coefficient is determined, the weight ratio of the first tire longitudinal speed and the second tire longitudinal speed in the weighted fusion process can be directly determined. In a specific implementation, the formula can be directly used: Where V xi is the actual tire longitudinal speed, is the longitudinal speed of the first tire, is the longitudinal speed of the second tire. i =1 means that the system fully trusts the longitudinal speed of the second tire; when G i= 0, the longitudinal speed of the first tire is fully trusted; in other cases, the weight G i Dynamic weighting is done between the two.

[0127] Embodiment 4

[0128] Reference Fig.11 , Fig.11 The flowchart of the fourth embodiment of the method for estimating the longitudinal speed of a tire of the present application is shown in FIG. The fourth embodiment of the method for estimating the longitudinal speed of a tire of the present application is proposed based on any one of the first to third embodiments of the method for estimating the longitudinal speed of a tire of the present application.

[0129] Before step S40, the following steps are also included:

[0130] Step S41: Acquire tire test data, where the tire test data at least includes tread strain information and corresponding slip rate collected by a PVDF sensor.

[0131] It should be understood that before using the preset slip rate prediction model to determine the longitudinal slip rate of the tire, it is also necessary to establish the preset slip rate prediction model. In the process of establishing the model, a large amount of tire test data is needed as support. The tire test data can be the data corresponding to the tread strain information collected by the tire during the experiment, or it can be the data corresponding to the tread strain information recorded during the actual use of the tire. In the process of determining the tread strain information, it is necessary to select a fixed position on the tire; when the position is fixed, select the tread strain information of the tire within a period of time and the slip rate corresponding to each tread strain information. The specific data of the tread strain information at different times on the tread, as well as the specific values ​​at different angles of the tread.

[0132] In the process of acquiring tire test data, the data can be acquired by extracting previous tire data, or the tire can be tested and the tire test data can be recorded in real time.

[0133] Step S42: establishing a tire slip rate prediction model based on the tire test data by using a Gaussian process regression algorithm.

[0134] It is understandable that, in the case of determining the tire experimental data for training, the model training can be carried out based on each tire tread strain information and the corresponding slip rate with intelligent machine learning as the framework and Gaussian regression process as the core. During the training process, the tire tread strain information is used as input and the corresponding slip rate is used as output for model training. When the training is completed, the accuracy of the tire slip rate prediction model needs to be tested. When the accuracy reaches the pre-set standard, the tire slip rate prediction model training is considered to be completed, otherwise the above training process is repeated.

[0135] Correspondingly, the step S40 is step S40': inputting the tread strain information into the tire slip rate prediction model to determine the longitudinal slip rate of the tire.

[0136] It should be understood that, when the tread strain information is collected, the tread strain information can be input into the preset slip rate prediction model, and the preset slip rate prediction model can determine the longitudinal slip rate of the tire according to the input tread strain information.

[0137] The step S50 comprises:

[0138] Step S501: Obtain the rotation speed and effective rolling radius of the tire.

[0139] It should be noted that although there is a certain mapping relationship between the longitudinal speed of the tire and the longitudinal slip rate, the corresponding mapping relationship is not the same for different tires. For example, the effective rolling radius of tires of different sizes is not the same. When the tire rotation speed is the same, the larger the effective rolling radius, the greater the longitudinal speed of the tire, and the smaller the longitudinal acceleration of the tire, the smaller the longitudinal slip rate of the corresponding tire.

[0140] Therefore, before determining the second tire longitudinal speed, it is also necessary to obtain the tire's rotation speed and effective rolling radius. The rotation speed refers to the angular velocity of the tire during rotation. In the specific acquisition process, the tire's rotation speed can be directly obtained from the angular velocity sensor, and of course it can also be calculated by other parameters, which are not specifically limited here. The effective rolling radius can be determined based on the size of the tire and the deformation generated during the tire operation.

[0141] Step S502: Calculate a second tire longitudinal speed of the tire according to the longitudinal slip rate, the rotation speed and the effective rolling radius.

[0142] It should be understood that when the longitudinal slip rate, the rotation speed and the effective rolling radius are all determined, the formula: V x =ΩR e (1-S x ) can be used to directly calculate the longitudinal speed of the tire. x is the longitudinal speed of the tire, Ω is the rotation speed, R e is the effective rolling radius, S x is the longitudinal slip rate.

[0143] In addition, in the embodiments of the present application, refer to Figure 12 to Figure 14 , the estimation results of the tire longitudinal velocity fusion estimation scheme under pure sideslip conditions, pure longitudinal slip conditions and combined conditions. Among them, Figure 12 to Figure 14 The load value changes are the same.

[0144] By observing the data in the figure, we can see that the trained algorithm has shown excellent recognition performance in different sideslip angles and slip ratios. At the same time, we can observe the probability distribution of the estimated value of the tire longitudinal velocity, that is, the 95% confidence interval. Although the area with a larger probability distribution is mainly concentrated in the large slip area, the estimated value in this area is very close to the actual measured value. Even under extreme conditions where the tire is in extreme working conditions (side slip angle reaches 8 degrees and longitudinal slip rate reaches 30%), actual measurements show that the algorithm's estimation error is as low as 0.47%, which can accurately identify the tire longitudinal velocity.

[0145] In order to achieve the above object, the present application also proposes a tire longitudinal speed estimation device, comprising:

[0146] The information acquisition module 10 is used to collect tire tread strain information through a PVDF sensor; the feature point confirmation module 20 is used to determine the key feature points of the strain response signal in the contact patch area according to the tread strain information; the speed determination module 30 is used to input the key feature points into a preset tire longitudinal speed prediction model to obtain a first tire longitudinal speed; the slip rate determination module 40 is used to determine the longitudinal slip rate of the tire according to the tread strain information and the preset slip rate prediction model; the speed determination module 30 is also used to calculate the second tire longitudinal speed of the tire according to the longitudinal slip rate; the speed determination module 30 is also used to perform weighted fusion of the first tire longitudinal speed and the second tire longitudinal speed to obtain the tire longitudinal speed.

[0147] The present application directly obtains the first tire longitudinal velocity through the key characteristic points of the strain response signal in the contact patch area; indirectly calculates the second tire longitudinal velocity of the tire through the longitudinal slip rate of the tire; and finally weightedly fuses the first tire longitudinal velocity and the second tire longitudinal velocity, thereby accurately estimating the tire longitudinal velocity through two different methods.

[0148] Other embodiments of the tire longitudinal velocity estimation device of the present application may refer to the embodiments of the tire longitudinal velocity estimation method, which will not be described in detail here.

[0149] In order to achieve the above-mentioned purpose, the present application also proposes a tire longitudinal velocity estimation system, the system comprising: a strain sensor A, a slip ring B, a charge amplifier C and a data acquisition module D;

[0150] The strain sensor A is arranged on the inner side of the tire. The strain sensor A is connected to the charge amplifier C via a slip ring B. The charge amplifier C is connected to the data acquisition module D.

[0151] The tire longitudinal velocity estimation system integrates three strain sensors, slip rings, charge amplifiers and data acquisition modules to accurately capture the tread strain information of the tire tread area. Fig.15 ,exist Fig.15 During the data analysis process, the correlation between the strain response characteristics and the longitudinal speed of the tire is mainly reflected in the center area of ​​the tread. Slip ring B is used to solve the physical interference problem of rotating parts during wired signal transmission. The charge amplifier C efficiently converts the high-impedance charge signal generated by the PVDF piezoelectric film into a low-impedance voltage signal, so that the original difficult-to-process charge signal is converted into a more easily processed voltage signal, significantly improving the signal processability and accuracy. The data acquisition module D is highly adjustable and can flexibly adjust the signal channel, sampling frequency and sampling method. A sampling frequency of 10kHz can be used to ensure accurate data collection. Fig.15 Notebook E is used as a carrier of the model, and a preset tire longitudinal speed prediction model and a preset slip rate prediction model are set in the notebook.

[0152] In order to achieve the above-mentioned purpose, the present application also proposes a tire longitudinal speed estimation device, which includes: a memory, a processor, a tire longitudinal speed estimation program stored in the memory and executable on the processor, and the tire longitudinal speed estimation system. When the tire longitudinal speed estimation program is executed by the processor, the steps of the tire longitudinal speed estimation method described in any of the above-mentioned embodiments are implemented.

[0153] The above are only some embodiments of the present application, and the scope of implementation of the present application is not limited thereto. Any equivalent structure or equivalent process transformation made using the contents of the specification and drawings of this application, or directly or indirectly applied in other related technical fields, are also included in the protection scope of the present application.

Claims

1. A tire longitudinal velocity estimation method, characterized in that: The tire longitudinal speed estimation method comprises: Collect tire tread strain information through PVDF sensors; Determining key characteristic points of a strain response signal in a contact patch region according to the tread strain information; Inputting the key feature points into a preset tire longitudinal speed prediction model to obtain a first tire longitudinal speed; Determining the longitudinal slip rate of the tire according to the tread strain information and a preset slip rate prediction model; calculating a second tire longitudinal velocity of the tire based on the longitudinal slip rate; The first tire longitudinal velocity and the second tire longitudinal velocity are weightedly fused to obtain the tire longitudinal velocity.

2. The tire longitudinal velocity estimation method according to claim 1, characterized in that: Determining the key characteristic points of the strain response signal in the contact patch area according to the tread strain information includes: determining a contact patch area of ​​the tire using the tread strain information; Filtering the strain response signal in the contact patch area to obtain a filtered first strain curve; Key feature points are determined according to the filtered first strain curve.

3. The tire longitudinal velocity estimation method according to claim 2, characterized in that: The determining of key feature points according to the filtered first strain curve comprises: Performing first-order differentiation and second-order differentiation on the filtered first strain curve to obtain a first-order differential curve and a second-order differential curve; The leading edge maximum value, trailing edge minimum value, and phase difference between the leading edge maximum value and the trailing edge minimum value of the first-order differential curve, and the leading edge maximum value, trailing edge maximum value, and phase difference between the leading edge maximum value and the trailing edge maximum value of the second-order differential curve are extracted as the key feature points.

4. The tire longitudinal velocity estimation method according to claim 2, characterized in that: Determining the contact patch area of ​​the tire using the tread strain information includes: Extracting a second strain curve at the center of the tread from the tread strain information; The second strain curve is analyzed to determine a contact patch area of ​​the tire.

5. The tire longitudinal velocity estimation method according to claim 1, characterized in that: The step of performing weighted fusion of the first tire longitudinal velocity and the second tire longitudinal velocity to obtain the tire longitudinal velocity comprises: Acquire a first probability distribution of the preset tire longitudinal speed prediction model and a second probability distribution of the preset slip rate prediction model; Determining a weight coefficient of a weighted fusion process according to the first probability distribution and the second probability distribution; The first tire longitudinal velocity and the second tire longitudinal velocity are weightedly fused according to the weight coefficient to obtain the tire longitudinal velocity.

6. The tire longitudinal velocity estimation method according to claim 1, characterized in that: The determining of the longitudinal slip rate of the tire according to the tread strain information and a preset slip rate prediction model comprises: Acquiring tire test data, wherein the tire test data at least includes tread strain information and corresponding slip rate collected by a PVDF sensor; Establishing a tire slip rate prediction model based on the tire test data by using a Gaussian process regression algorithm; Accordingly, determining the longitudinal slip rate of the tire according to the tread strain information and a preset slip rate prediction model includes: The tread strain information is input into the tire slip rate prediction model to determine the longitudinal slip rate of the tire.

7. The tire longitudinal velocity estimation method according to claim 6, characterized in that: The step of calculating a second tire longitudinal velocity of the tire according to the longitudinal slip rate comprises: Obtaining the rotation speed and effective rolling radius of the tire; A second tire longitudinal speed of the tire is calculated according to the longitudinal slip rate, the rotation speed, and the effective rolling radius.

8. A tire longitudinal velocity estimation device, characterized in that: include: An information acquisition module, used for collecting tire tread strain information through a PVDF sensor; A feature point confirmation module, used to determine key feature points of a strain response signal in a contact patch area according to the tread strain information; A speed determination module, used for inputting the key feature point into a preset tire longitudinal speed prediction model to obtain a first tire longitudinal speed; A slip rate determination module, used to determine the longitudinal slip rate of the tire according to the tread strain information and a preset slip rate prediction model; The speed determination module is further used to calculate a second tire longitudinal speed of the tire according to the longitudinal slip rate; The speed determination module is further configured to perform weighted fusion of the first tire longitudinal speed and the second tire longitudinal speed to obtain the tire longitudinal speed.

9. A tire longitudinal velocity estimation system, characterized in that: The system comprises: a strain sensor, a slip ring, a charge amplifier and a data acquisition module; The strain sensor is arranged on the inner side of the tire, the strain sensor is connected to the charge amplifier via a slip ring, and the charge amplifier is connected to the data acquisition module.

10. A tire longitudinal velocity estimation device, characterized in that: The device includes: a memory, a processor, a tire longitudinal speed estimation program stored in the memory and executable on the processor, and the tire longitudinal speed estimation system according to claim 9. When the tire longitudinal speed estimation program is executed by the processor, the steps of the tire longitudinal speed estimation method according to any one of claims 1 to 7 are implemented.

Images