Wheel looseness monitoring method based on wheel speed signal spectrum analysis

Abstract

The application provides a wheel loosening monitoring method based on wheel speed signal spectrum analysis, the method is based on wheel speed signal collected by original Hall sensor of a vehicle, and after the wheel speed signal is processed by using an interpolation algorithm, special filtering processing and a rough road surface compensation mode, power spectrum estimation of relevant frequency bands is carried out, and finally real-time power spectrum characteristic states of each wheel are obtained; in the special filtering processing process, according to the relationship between the frequency and the energy distribution in the amplitude spectrum diagram of the frequency band signal processed by filtering, 8-15Hz and 60-80Hz two frequency bands are determined as the frequency interval for judging the wheel loosening, and feature extraction is carried out for the frequency interval, and finally the accurate monitoring of the wheel loosening is realized.

Description

Technical Field

[0001] This invention relates to the field of wheel anomaly monitoring technology, and in particular to a wheel loosening monitoring method based on wheel speed signal spectrum analysis. Background Technology

[0002] During vehicle use, the wheels are subjected to pressure from the entire vehicle and friction from the underside. Over time, wheel nuts or bolts may loosen, causing the wheels to become loose and potentially leading to serious traffic accidents. Therefore, monitoring wheel looseness is crucial. Currently, there are two main methods for monitoring wheel looseness: The first method involves installing a pressure sensor between the wheel hub and the wheel core. This sensor detects the wheel's motion signal and inputs it to a signal processing circuit. By monitoring changes in the pressure signal, the pressure between the wheel hub and the wheel core can be monitored, thus identifying whether the wheel is loose. However, this method requires installing a pressure sensor and signal processing module in each wheel hub. Since the wheel hub and wheel core rotate relative to the vehicle, the sensor needs to be mounted on the wheel core and powered independently. Therefore, a battery and a wireless transmission module are also required. Furthermore, because the pressure between the wheel hub and the wheel core is relatively high, there is a risk of… The possibility of wear and tear places high demands on the pressure sensor, requiring a high-quality sensor with reliable performance. The second method identifies wheel loosening by analyzing the time-domain variance characteristics of the wheel speed signal. This software solution first corrects the wheel speed signal, removing manufacturing errors from the wheel speed sensor, and then calculates the variance of the wheel speed signal. The difference between the wheel speed signal variance and the reference signal is then integrated as the detection input. If the accumulated error exceeds a threshold, wheel loosening is considered to have occurred. This method primarily processes the time-domain information of the wheel speed signal; therefore, this solution has many requirements regarding the vehicle's operating conditions. The excitation from rough road surfaces can affect the system's accuracy.

[0003] In view of this, it is necessary to design an improved wheel loosening monitoring method based on wheel speed signal spectrum analysis to solve the above problems. Summary of the Invention

[0004] The purpose of this invention is to provide a wheel loosening monitoring method based on wheel speed signal spectrum analysis.

[0005] To achieve the above-mentioned objectives, this invention provides a method for monitoring wheel loosening based on wheel speed signal spectrum analysis, comprising the following steps:

[0006] S1. The vehicle speed signal during the motion process is collected by a Hall sensor installed inside the wheel hub, and then the discrete vehicle speed signal is converted into a digital signal based on a time uniform distribution by linear interpolation.

[0007] S2. After filtering and removing interference from the digital signal obtained in step S1, a spectrum analysis is performed. The interference removal process first calculates the harmonic components F(n) of the wheel speed digital signal, and then calculates the current angular velocity ω based on the number of wheel teeth K and the timestamps X(n) and X(n-1) of adjacent teeth, thereby obtaining the fundamental frequency and the frequency values ​​of its Nth harmonic. Then, X(n) is interpolated and the amplitude spectrum e(n) is calculated using a fast Fourier transform. Finally, the amplitude spectrum e(n) is locally smoothed and filtered to obtain the frequency band that has energy characteristics with the wheel frequency. This frequency band is used as the frequency range for judging wheel looseness. The frequency band is 8-15Hz and 60-80Hz.

[0008] S3. Calculate the average signal power of the 8-15Hz and 60-80Hz frequency bands respectively. Then, based on the average signal power of the two frequency bands, further calculate the wheel looseness judgment threshold Ps. When the alarm confirmation value Pr set by the system is greater than the alarm cumulative confirmation value Psum, the system issues a wheel looseness alarm and notifies the user through the vehicle instrument interaction interface.

[0009] S4. Considering the impact of road surface roughness on the road excitation caused by the moving wheels on the spectral energy of 8-15Hz and 60-80Hz, compensation calculations and disabling / exclusion are performed. The slip ratio variance Var is obtained by using the average speed Vfront of the front axle wheels and the average speed Vrear of the two rear axle wheels. The road surface roughness compensation value Pcomp is further obtained by using the slip ratio variance Var. An alarm can be triggered when the alarm confirmation value Pr(n) of each wheel is greater than the alarm cumulative confirmation value Psum.

[0010] Preferably, in step S4, the expression for Var is: Where is the average relative slip rate S of a set of samples, Sn is the slip rate of that sample, and the expression for the relative slip rate S is S=(Vfront-Vrear) / Vfront, where Vfront represents the average speed of the front axle and Vrear represents the average speed of the rear axle.

[0011] Preferably, to determine whether to perform road surface compensation, a set value Vset is given. The set value Vset is the average value of Var1 when the vehicle is driving on a smooth road surface and Var2 when it is driving on a rough road surface. When Var is less than the set value Vset, it is considered that the vehicle is driving on a relatively smooth road surface, and the road surface roughness compensation value is 0.

[0012] Preferably, when Var is greater than the set value Vset, within a certain period of time, since the road surface provides relatively consistent road surface excitation to the four wheels in motion, the wheel loosening alarm estimates P(n) of the four wheels are sorted. If the differences between the P(n) of the four wheels are large, the smallest road surface excitation is used as the road surface roughness compensation value. If the differences between the wheel loosening alarm estimates P(n) of the four wheels are not large, identification is required after the vehicle has traveled to a good road surface.

[0013] Preferably, in step S3, the wheel loosening judgment threshold Ps = P stable +(P loss -P stable )*β, where P loss P represents the energy value when the wheel becomes loose. stable The P value represents the energy level when the wheel is not loose. loss and the P stable Both are calculated using P = ε * P1 + (1 - ε) * P2, where P1 is the signal power in the 8-15Hz frequency band and P2 is the signal power in the 60-80Hz frequency band.

[0014] Preferably, in step S4, the expression for the alarm confirmation value Pr(n) is Pr(n)=∑(P(n)-Pcomp-Ps(n)); where Pcomp is the road surface roughness compensation value, and its expression is Pcomp=Psmin*α*Var, where α is an adjustment factor determined by calibration data, and Psmin is the minimum wheel energy value.

[0015] Preferably, in step S3, the formula for calculating the signal power in the 8-15Hz frequency band is as follows: N = 8, n = 8 - 15; the formula for calculating the signal power in the 60-80Hz frequency band is: N = 21, n = 60 - 80.

[0016] Preferably, in step S2, the expression for the harmonic component F(n) is F(n) = N * f0, where N is a positive integer greater than 0, and F(n)max < 80.

[0017] Preferably, the formula for calculating the angular velocity ω is ω=2π / ((X(n)-X(n-1))*k).

[0018] Preferably, in step S2, the expression for e(n) is e(n) = (e(n-1) + e(n+1)), where e(n) is the amplitude of the fundamental frequency and its Nth harmonic spectrum, e(n-1) is the amplitude of the frequency preceding the harmonic frequency point, e(n+1) is the amplitude of the frequency following the harmonic frequency point, and the frequency point n < 80Hz.

[0019] The beneficial effects of this invention are:

[0020] This invention proposes a wheel loosening monitoring method based on wheel speed signal spectrum analysis. This method uses the vehicle's existing Hall effect sensors to collect wheel speed signals. The wheel speed signals are then processed using interpolation algorithms, special filtering, and road roughness compensation before power spectrum estimation of relevant frequency bands is performed. This yields the real-time power spectrum characteristics of each wheel. During the special filtering process, based on the relationship between frequency and energy distribution in the amplitude spectrum of the filtered signal, the 8-15Hz and 60-80Hz frequency bands are identified as the frequency ranges for judging wheel loosening. Feature extraction is then performed on these frequency ranges, ultimately achieving accurate monitoring of wheel loosening. This invention's detection method, by employing spectral feature analysis and comprehensively considering the impact of various error factors such as road excitation and gear ring manufacturing errors on wheel loosening assessment, achieves higher monitoring accuracy than existing technologies through compensation. Attached Figure Description

[0021] Figure 1 This is a diagram showing the positional relationship between the wheel teeth and the Hall sensor involved in the wheel loosening monitoring method based on wheel speed signal spectrum analysis of the present invention.

[0022] Figure 2 This is the amplitude spectrum of the frequency band signal monitored during the monitoring and testing process of this invention;

[0023] Figure 3 for Figure 2 The amplitude spectrum of the frequency band signal obtained after filtering. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0025] It should also be noted that, in order to avoid obscuring the present invention with unnecessary details, only the structures and / or processing steps closely related to the present invention are shown in the accompanying drawings, while other details that are not closely related to the present invention are omitted.

[0026] Additionally, it should be noted that the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0027] Please see Figure 1 As shown, this invention provides a wheel loosening monitoring method based on wheel speed signal spectrum analysis, comprising the following steps:

[0028] S1. Wheel speed signal interpolation: The vehicle speed signal during the motion process is collected by a Hall sensor installed inside the wheel hub, and then the discrete vehicle speed signal is converted into a digital signal based on a time uniform distribution by linear interpolation.

[0029] The principle of using a Hall sensor to collect wheel speed signals in the above process is as follows: During wheel movement, the toothed ring on the wheel core rotates with the wheel. When the protruding tooth of the toothed ring approaches the Hall sensor, the Hall sensor outputs a high-level signal, and whenever the protruding tooth of the toothed ring moves away from the Hall sensor, the Hall sensor outputs a low-level signal. The above level signals are transmitted to the control system through the circuit to form a digital signal sequence. Since the above level signals are generated according to the protruding tooth of the toothed ring approaching or moving away from the Hall sensor, and the actual vehicle speed is not stable and the road surface is not perfectly flat, the generation of each digital signal is based on events rather than equal time intervals. That is, the digital signal is not standard time-uniformly sampled data, so it needs to be converted into a time-uniformly distributed digital signal before it can be converted from the time domain to the frequency domain and processed by an algorithm. This invention uses interpolation to solve the above problem. After interpolation, the sampled data is distributed based on equal time intervals.

[0030] S2. Fundamental and Harmonic Elimination: Due to manufacturing errors, the spacing between toothed rings is not perfectly uniform. Therefore, when the wheel rotates once, a periodic (number of rotations) signal caused by toothed ring errors is superimposed on the ideal wheel speed signal. This signal is the fundamental signal related to the wheel speed, with a frequency of f0 = ω / 2π. The value of frequency f0 changes with the angular velocity ω. Since this fundamental signal is physically a signal with obvious periodic characteristics, when the wheel speed digital signal is converted to the frequency domain, it will generate a very large interference signal at the corresponding frequency point. Its signal amplitude is much larger than the signal amplitude changes caused by tire vibration, ground excitation, and wheel looseness. Therefore, this signal must be filtered out first before subsequent spectrum analysis can be performed.

[0031] Since the wheel speed digital signal is a non-sinusoidal periodic signal, it has N harmonic components. The expression for the harmonic components is F(n) = N * f0, where N is a positive integer greater than 0, and F(n)max < 80. It should be noted that F(n)max here is obtained based on multiple actual tests. During the monitoring process, the current angular velocity ω is calculated based on the number of wheel teeth K and the timestamps X(n) and X(n-1) of adjacent teeth. The formula for calculating the angular velocity ω is ω = 2π / ((X(n) - X(n-1)) * k). Based on this, the fundamental frequency and the frequency values ​​of its Nth harmonic are obtained. The amplitude spectrum e(n) of the interpolated signal sequence X(n) is calculated using a fast Fourier transform. To ensure the accuracy of the results, a large amount of data is collected during the monitoring process for verification. The amplitude spectrum of the frequency band signal is as follows: Figure 2 As shown in the figure, the signal values ​​of the fundamental wave and harmonics are much larger than the amplitudes of adjacent frequency points. The changes in signal energy caused by noise signal energy, road surface excitation, and wheel hub vibration are usually range-bound, and their amplitude envelopes are relatively flat. Therefore, this invention uses a smoothing filtering algorithm to perform local smoothing filtering on the fundamental wave and its Nth harmonic. That is, the spectral amplitude of the fundamental wave and its Nth harmonic is obtained by averaging the amplitudes of adjacent frequency points. The specific calculation method is e(n) = (e(n-1) + e(n+1)), where e(n) is the spectral amplitude of the fundamental wave and its Nth harmonic, e(n-1) represents the amplitude of the frequency before the harmonic frequency point, and e(n+1) represents the amplitude of the frequency after the harmonic frequency point, with frequency point n < 80Hz. To more intuitively see the amplitude distribution in the monitoring frequency band, spike filtering is used. Figure 2 The amplitude spectrum of the frequency band signal is as follows: Figure 3 As shown in the figure, the fundamental wave and its Nth harmonic are smoothly filtered out, and there are local energy characteristics in 8-15Hz and 60-80Hz. Therefore, the 8-15Hz and 60-80Hz frequency bands are used as the frequency ranges for judging wheel looseness.

[0032] S3. Spectral Power Estimation: Average signal power is calculated for the 8-15Hz and 60-80Hz frequency bands respectively. The formula for calculating the signal power in the 8-15Hz band is as follows: N = 8, n = 8 - 15; the formula for calculating the signal power in the 60-80Hz frequency band is: N = 21, n = 60 - 80;

[0033] Since the energy of P1 and P2 will increase when the wheel is loose, the power of the two signals is weighted and summed to comprehensively evaluate the impact of the wheel loosening. The impact evaluation value P = ε * P1 + (1 - ε) * P2, where ε is an adjustment factor that can be adjusted according to the vehicle model.

[0034] Furthermore, to better assess the looseness of the wheel, the energy values ​​P for both loose and non-loose wheel conditions are calculated using the aforementioned formula for calculating the impact assessment value P. stable and P loss Based on this, the threshold for judging the looseness of a single wheel, Ps = P, is calculated. stable +(P loss -P stable )*β, where β is generally around 0.8. When the system detects that the system's energy value exceeds the threshold, it issues a wheel loosening alarm. In particular, during the monitoring process, to ensure the accuracy of the monitoring, an alarm confirmation value is determined for the system. The alarm confirmation value Pr is calculated as Pr=∑(P-Ps). When the alarm confirmation value Pr>Psum, the system issues a wheel loosening alarm and notifies the user through the vehicle's instrument panel interface. Here, Psum represents the cumulative alarm confirmation value, the value of which can be set according to the vehicle model.

[0035] S4. Road Surface Roughness Compensation and Disable: To ensure the accuracy of wheel loosening monitoring, the road surface excitation of the wheel also affects the spectral energy of 8-15Hz and 60-80Hz. The higher the road surface roughness, the greater the spectral noise energy. This is because the greater the road surface roughness, the greater the spectral noise energy, causing frequent fluctuations in wheel speed. Therefore, it is necessary to perform certain disable / exclusion operations to determine the alarm confirmation value Pr(n) for each wheel, which is expressed as Pr(n)=∑(P(n)-Pcomp-Ps(n)); where Pcomp is the road surface roughness compensation value, which is expressed as Pcomp=Psmin*α*Var, α is an adjustment factor determined by calibration data, Psmin is the minimum wheel energy value, Pcomp and Var are positively correlated, and Var represents the variance of slip ratio, expressed as: S′ is the average of the relative slip ratio S of a set of samples, Sn is the sum of the relative slip ratios of the four wheels, and the expression for the relative slip ratio S is S=(Vfront-Vrear) / Vfront, where Vfront represents the average speed of the two wheels on the front axle and Vrear represents the average speed of the two wheels on the rear axle.

[0036] To determine whether road surface compensation should be performed, a set value Vset is given. The set value Vset is the average of Var1 when the vehicle is driving on a smooth road surface and Var2 when it is driving on a rough road surface. When Var is less than the set value Vset, it is considered that the vehicle is driving on a relatively smooth road surface. At this time, the road surface roughness compensation value is 0, and an alarm can be triggered when the alarm confirmation value Pr(n) of each wheel is greater than Psum.

[0037] When Var is greater than the set value Vset, within a certain period of time, the road surface provides relatively consistent road surface excitation to the four wheels in motion. The wheel loosening alarm estimates P(n) of the four wheels are sorted. If the differences between the P(n) of the four wheels are large, the road surface excitation with the smallest value is used as the road surface roughness compensation value. If the differences between the wheel loosening alarm estimates P(n) of the four wheels are not large, it is impossible to determine whether the current spectral energy change is caused by the road surface or by wheel loosening. Therefore, it is impossible to determine whether the four wheels are loose. The identification is required after the vehicle travels to a good road surface.

[0038] The wheel loosening monitoring method based on wheel speed signal spectrum analysis of the present invention will be further defined below with reference to specific embodiments:

[0039] Example 1

[0040] This embodiment proposes a wheel loosening monitoring method based on wheel speed signal spectrum analysis, which includes the following steps:

[0041] S1. Wheel speed signal interpolation: The vehicle speed signal during the motion process is collected by a Hall sensor installed inside the wheel hub, and then the discrete vehicle speed signal is converted into a digital signal based on a time uniform distribution by linear interpolation.

[0042] S2. Fundamental and Harmonic Elimination: Due to manufacturing errors, the spacing between toothed rings is not perfectly uniform. Therefore, when the wheel rotates once, a periodic (number of rotations) signal caused by toothed ring errors is superimposed on the ideal wheel speed signal. This signal is the fundamental signal related to the wheel speed, with a frequency of f0 = ω / 2π. The value of frequency f0 changes with the angular velocity ω. Since this fundamental signal is physically a signal with obvious periodic characteristics, when the wheel speed digital signal is converted to the frequency domain, it will generate a very large interference signal at the corresponding frequency point. Its signal amplitude is much larger than the signal amplitude changes caused by tire vibration, ground excitation, and wheel looseness. Therefore, this signal must be filtered out first before subsequent spectrum analysis can be performed.

[0043] Since the wheel speed digital signal is a non-sinusoidal periodic signal, it has N harmonic components. The expression for the harmonic components is F(n) = N * f0, where N is a positive integer greater than 0, and F(n)max < 80. During monitoring, the current angular velocity ω is calculated based on the number of wheel teeth K and the timestamps X(n) and X(n-1) of adjacent teeth. The formula for calculating the angular velocity ω is ω = 2π / ((X(n) - X(n-1)) * k). Based on this, the fundamental frequency and the frequency values ​​of its Nth harmonic are obtained. The amplitude spectrum e(n) of the interpolated signal sequence X(n) is calculated using a fast Fourier transform. To ensure the accuracy of the results, a large amount of data is collected during the monitoring process for verification. The amplitude spectrum of the frequency band signal is as follows: Figure 2 As shown in the figure, the signal values ​​of the fundamental wave and harmonics are much larger than the amplitudes of adjacent frequency points. The changes in signal energy caused by noise signal energy, road surface excitation, and wheel hub vibration are usually range-bound, and their amplitude envelopes are relatively flat. Therefore, this invention uses a smoothing filtering algorithm to perform local smoothing filtering on the fundamental wave and its Nth harmonic. That is, the spectral amplitude of the fundamental wave and its Nth harmonic is obtained by averaging the amplitudes of adjacent frequency points. The specific calculation method is e(n) = (e(n-1) + e(n+1)), where e(n) is the spectral amplitude of the fundamental wave and its Nth harmonic, and the frequency point n < 80Hz. To more intuitively see the amplitude distribution in the monitoring frequency band, spike filtering is used. Figure 2 The amplitude spectrum of the frequency band signal is as follows: Figure 3 As shown in the figure, the fundamental wave and its Nth harmonic are smoothly filtered out, and there are local energy characteristics in 8-15Hz and 60-80Hz. Therefore, the 8-15Hz and 60-80Hz frequency bands are used as the frequency ranges for judging wheel looseness.

[0044] S3. Spectral Power Estimation: Average signal power is calculated for the 8-15Hz and 60-80Hz frequency bands respectively. The formula for calculating the signal power in the 8-15Hz band is as follows: N = 8, n = 8 - 15; the formula for calculating the signal power in the 60-80Hz frequency band is: N = 21, n = 60 - 80;

[0045] Since the energy of P1 and P2 will increase when the wheel is loose, the power of the two signals is weighted and summed to comprehensively evaluate the impact of the wheel loosening. The impact evaluation value P = ε * P1 + (1 - ε) * P2, where ε is an adjustment factor that can be adjusted according to the vehicle model.

[0046] Furthermore, to better assess the looseness of the wheel, the energy values ​​P for both loose and non-loose wheel conditions are calculated using the aforementioned formula for calculating the impact assessment value P. stable and P loss Taking ε = 0.7, and without loosening P1 = 4500 and P2 = 6500, we calculate P. stable =5100; when loosened, P1=7000, P2=8600. Calculated P loss =7480, and based on this, the threshold for judging the looseness of a single wheel, Ps = P stable +(P loss -P stable)*β, where β=0.8 in this embodiment, when the system detects that the energy value of the system exceeds the threshold, an alarm for wheel loosening is issued; in particular, during the monitoring process, in order to ensure the accuracy of the monitoring, an alarm determination value is determined for the system. The alarm determination value Pr is calculated as Pr=∑(P-Ps). When the alarm determination value Pr>Psum, the system issues a wheel loosening alarm and notifies the user through the vehicle instrument interaction interface. In this embodiment, Psum=1000;

[0047] S4. Road Surface Roughness Compensation and Disable: To ensure the accuracy of wheel loosening detection, the road surface excitation of the wheel also affects the spectral energy of 8-15Hz and 60-80Hz. The higher the road surface roughness, the greater the spectral noise energy. This is because the greater the road surface roughness, the greater the spectral noise energy, causing frequent fluctuations in wheel speed. Therefore, it is necessary to perform certain disable / exclusion operations to determine the alarm confirmation value Pr(n) for each wheel. The expression is Pr(n)=∑(P(n)-Pcomp-Ps(n)); where Pcomp=Psmin*α*Var, α is an adjustment factor determined by calibration data, Psmin=5100, Pcomp and Var are positively correlated, and Var represents the variance of slip ratio, expressed as: S′ is the average of the relative slip ratio S of a set of samples, Sn={2.5%, 1.6%, 2.1%, 1.6%, 2.3%, 2.7%}. The expression for the relative slip ratio S is S=(Vfront-Vrear) / Vfront, where Vfront represents the average speed of the front axle and Vrear represents the average speed of the rear axle. The calculated Vfront=45kph and Vrear=45.8kph, when substituted into the aforementioned formula, S=1.7% and Var=1.77%.

[0048] To determine whether road surface compensation should be performed, a set value Vset is given, which is Vset = (Var1 + Var2) / 2 = 3.5%. Since Var < set value Vset, it is assumed that the vehicle is driving on a relatively smooth road surface. At this time, the road surface roughness compensation value is 0. An alarm can be triggered when the alarm confirmation value Pr(n) of each wheel is greater than Psum.

[0049] In summary, the wheel loosening monitoring method proposed in this invention, based on wheel speed signal spectrum analysis, collects wheel speed signals using the vehicle's existing Hall effect sensors. It then processes these signals using interpolation algorithms, special filtering, and road roughness compensation before estimating the power spectrum of relevant frequency bands. This yields the real-time power spectrum characteristics of each wheel. During the special filtering process, based on the relationship between frequency and energy distribution in the amplitude spectrum of the filtered signal, the 8-15Hz and 60-80Hz frequency bands are identified as the frequency ranges for judging wheel loosening. Feature extraction is then performed on these frequency ranges, ultimately achieving accurate monitoring of wheel loosening.

[0050] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for monitoring wheel loosening based on wheel speed signal spectrum analysis, characterized in that, Includes the following steps: S1. The vehicle speed signal during the motion process is collected by a Hall sensor installed inside the wheel hub, and then the discrete vehicle speed signal is converted into a digital signal based on a time uniform distribution by linear interpolation. S2. After filtering and removing interference from the digital signal obtained in step S1, a spectrum analysis is performed. The interference removal process first calculates the harmonic components F(n) of the wheel speed digital signal, and then calculates the current angular velocity ω based on the number of wheel teeth k and the timestamps X(n) and X(n-1) of adjacent teeth, thereby obtaining the fundamental frequency and the frequency values ​​of its Nth harmonic. Then, X(n) is interpolated and the amplitude spectrum e(n) is calculated using a fast Fourier transform. Finally, the amplitude spectrum e(n) is locally smoothed and filtered to obtain the frequency band that has energy characteristics with the wheel frequency. This frequency band is used as the frequency range for judging wheel looseness. The frequency band is 8-15Hz and 60-80Hz. S3. Calculate the average signal power of the 8-15Hz and 60-80Hz frequency bands respectively. Then, based on the average signal power of the two frequency bands, further calculate the wheel looseness judgment threshold Ps. When the alarm confirmation value Pr set by the system is greater than the alarm cumulative confirmation value Psum, the system issues a wheel looseness alarm and notifies the user through the vehicle instrument interaction interface. S4. Considering the impact of road surface roughness on the road excitation caused by the wheels during driving on the spectral energy of 8-15Hz and 60-80Hz, compensation calculations and disabling are performed. The slip ratio variance Var is obtained by using the average speed Vfront of the front axle wheels and the average speed Vrear of the two rear axle wheels. The road surface roughness compensation value Pcomp is further obtained by using the slip ratio variance Var. An alarm can be triggered when the alarm confirmation value Pr(n) of each wheel is greater than the alarm cumulative confirmation value Psum. To determine whether road surface compensation should be performed, a set value Vset is given. The set value Vset is the average of the variance Var1 when the vehicle is driving on a smooth road surface and the variance Var2 when it is driving on a rough road surface. When Var is less than the set value Vset, the vehicle is considered to be driving on a relatively smooth road surface, and the road surface roughness compensation value is 0. When Var is greater than the set value Vset, the road surface excitation of the four wheels in motion is relatively consistent within a certain period of time. The wheel loosening alarm estimates P(n) of the four wheels are sorted. If the difference between P(n) of the four wheels is large, the road surface excitation with the smallest value is used as the road surface roughness compensation value. If the difference between the wheel loosening alarm estimates P(n) of the four wheels is not large, the system is temporarily disabled and identification is required after the vehicle has traveled to a good road surface.

2. The wheel loosening monitoring method based on wheel speed signal spectrum analysis according to claim 1, characterized in that, In step S4, the expression for Var is: ,in, Let S be the average relative slip rate S of a set of samples, and Sn be the slip rate of that sample. The expression for the relative slip rate S is: Vfront represents the average speed of the front axle, and Vrear represents the average speed of the rear axle.

3. The wheel loosening monitoring method based on wheel speed signal spectrum analysis according to claim 1, characterized in that, In step S3, the wheel loosening judgment threshold Ps=P stable +(P) loss -P stable )*β, where P loss P represents the energy value when the wheel becomes loose. stable This represents the energy value when the wheel is not loose; β is typically around 0.

8. The P... loss and the P stable All passed The calculation yielded that, As adjustment factors, P1 is the signal power in the 8-15Hz frequency band, and P2 is the signal power in the 60-80Hz frequency band.

4. The wheel loosening monitoring method based on wheel speed signal spectrum analysis according to claim 1, characterized in that, In step S4, the expression for the alarm confirmation value Pr(n) is: Where Pcomp is the road surface roughness compensation value, its expression is: α is an adjustment factor determined by calibration data, and Psmin is the minimum wheel energy value.

5. The wheel loosening monitoring method based on wheel speed signal spectrum analysis according to claim 3, characterized in that, In step S3, the formula for calculating the signal power in the 8-15Hz frequency band is as follows: N=8, n=8-15; the formula for calculating the signal power in the 60-80Hz frequency band is: N=21, n=60-80.

6. The wheel loosening monitoring method based on wheel speed signal spectrum analysis according to claim 1, characterized in that, In step S2, the expression for the harmonic component F(n) is: Where N is a positive integer, f0 is the frequency of the fundamental signal, and its expression is f0=ω / 2π, where ω is the angular velocity; and the maximum value of F(n) is < 80.

7. The wheel loosening monitoring method based on wheel speed signal spectrum analysis according to claim 1, characterized in that, The formula for calculating the angular velocity ω is ω = 2π / ((X(n)-X(n-1))*k).

8. The wheel loosening monitoring method based on wheel speed signal spectrum analysis according to claim 1, characterized in that, In step S2, the expression for e(n) is e(n) = (e(n-1)+e(n+1)), where e(n) is the amplitude of the fundamental frequency and its Nth harmonic spectrum, e(n-1) is the amplitude of the frequency preceding the harmonic frequency, e(n+1) is the amplitude of the frequency following the harmonic frequency, and n<80Hz.

Images