Hydrogen fuel vehicle driving data acquisition method and system
By dynamically adjusting the wavelet packet threshold in hydrogen fuel cell vehicles and combining hydrogen concentration, speed, and humidity characteristics, the problem of false alarms and false negatives in hydrogen leak identification under complex environments by traditional algorithms is solved, and accurate acquisition of hydrogen concentration data and accurate identification of leak risks are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- QINGDAO MEIJIN NEW ENERGY VEHICLE MANUFACTURING CO LTD
- Filing Date
- 2026-05-12
- Publication Date
- 2026-06-09
AI Technical Summary
Traditional wavelet packet thresholding denoising algorithms are prone to underreporting and false alarms of hydrogen leaks when acquiring driving data of hydrogen fuel cell vehicles, and cannot accurately identify potential leak risks. Especially in high-speed driving and high-humidity environments, fixed thresholds cannot adapt to non-stationary and non-Gaussian noise interference.
By acquiring data on hydrogen concentration, speed, and ambient humidity during vehicle operation, the median, variance, and maximum energy of the highest frequency subband coefficient are calculated. Combined with driving speed and humidity characteristic factors, the wavelet packet threshold is dynamically adjusted to achieve adaptive noise reduction.
It improves the accuracy and robustness of hydrogen concentration data, enabling timely and accurate identification of potential leakage risks and enhancing the reliability of vehicle safety monitoring and fault early warning.
Smart Images
Figure CN122165956A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of monitoring technology of operating parameters of electric vehicles, and relates to a method and system for acquiring driving data of hydrogen fuel cell vehicles, which is used to collect and process the operating information of hydrogen fuel cell vehicles in operation in real time to ensure vehicle operation safety. Background Technology
[0002] Hydrogen fuel cell vehicles, as a zero-emission, high-efficiency new energy vehicle, are gradually becoming an important direction for the development of the automotive industry. During the operation of hydrogen fuel cell vehicles, hydrogen safety is paramount in the overall vehicle control strategy. To ensure safe vehicle operation, hydrogen concentration sensors are typically installed in key components such as the chassis to monitor for hydrogen leaks in real time. When the detected hydrogen concentration exceeds a safety threshold, the vehicle controller immediately implements protective measures such as shutting off the hydrogen supply, forced ventilation, or powering off the entire vehicle. Therefore, the accuracy of hydrogen monitoring data directly affects vehicle driving safety and uptime. One existing method for denoising non-stationary signals is the wavelet packet thresholding algorithm, which possesses excellent time-frequency localization characteristics and can perform refined analysis and processing of signals containing rich frequency components.
[0003] Traditional wavelet packet thresholding denoising algorithms typically employ a general threshold calculation method based on signal statistical characteristics when processing hydrogen concentration monitoring data. This method assumes that the noise is uniformly distributed Gaussian white noise. However, during actual driving of hydrogen fuel cell vehicles, especially at high speeds and in rainy or snowy weather, the hydrogen concentration sensor installed in the chassis area is subject to significant environmental interference. The strong airflow turbulence generated by the vehicle's high-speed movement creates wind pressure noise on the sensor surface, while high humidity causes a water film effect on the surface of the sensor's electrochemical probe, generating nonlinear humidity interference noise. The superposition of these two types of noise introduces a large number of non-Gaussian, non-stationary pseudo-leakage features into the hydrogen concentration data. If the threshold is calculated solely based on the statistical characteristics of the hydrogen concentration data itself, the algorithm cannot distinguish between the severe noise fluctuations caused by high wind speeds and high humidity and the concentration mutations caused by actual hydrogen leaks. When the vehicle is in special conditions of high speed and high humidity, the general threshold calculated by traditional algorithms is often too small, leading to the incorrect retention of noise and false alarms. Alternatively, the threshold may be too large in certain frequency bands, easily filtering out weak early leak signals as noise, resulting in missed detections.
[0004] Among existing related technologies, Chinese patent document CN114987284B, entitled "A Multi-Sensor Battery Thermal Runaway Early Warning System Based on Edge-End Collaborative Computing," discloses an early warning system involving multi-sensor collaboration. In the data processing stage, after receiving raw data from various sensors such as hydrogen concentration, its edge computing module uses the traditional wavelet thresholding method to remove noise from the raw data. Subsequently, it performs data feature extraction and thermal runaway risk assessment. When denoising hydrogen concentration data, this patent application uses a standard wavelet thresholding algorithm. Its threshold is usually a fixed parameter set based on global noise statistical characteristics. Its shortcomings are: it does not take into account the non-stationary, non-Gaussian noise interference caused by wind pressure turbulence and water film effect on sensors in actual high-speed driving or high-humidity environments of hydrogen fuel cell vehicles. When faced with complex vehicle driving conditions, the fixed denoising threshold is prone to misjudging the real hydrogen concentration change as noise filtering or misretaining strong environmental wind pressure noise as a real signal, resulting in insufficient adaptive capability.
[0005] Among existing related technologies, Chinese patent document CN112895900B, entitled "A Hydrogen Redundancy Monitoring and Protection Device and Method for Hydrogen-Energy Tram," discloses an on-board hydrogen monitoring and protection technology. Its core lies in acquiring and processing the pulse-width modulation (PWM) signal output from the hydrogen concentration sensor through a hardware and software combined filtering module. The system uses a fixed filtering algorithm to filter out high-frequency electromagnetic interference in the PWM signal. Then, the duty cycle of the processed PWM signal is converted into a corresponding hydrogen concentration value. Finally, the acquired hydrogen concentration value is compared with multiple preset fixed concentration safety thresholds, thereby achieving graded fault warnings and corresponding system power-off protection. The patent application relies on fixed-parameter filtering for electromagnetic interference. However, when hydrogen fuel cell vehicles are driving on complex roads, the strong wind pressure turbulence and water film effect caused by high speed and high humidity can cause severe non-steady mechanical vibrations and environmental interference to the hydrogen sensor. The fixed filtering parameters in the patent application cannot be adaptively adjusted according to the actual driving speed and ambient humidity of the vehicle. This can easily lead to the distortion of the actual small leakage signal as noise under harsh conditions, or the misinterpretation of strong ambient wind pressure noise as a real leakage signal, resulting in serious distortion of the monitoring data and false alarms from the early warning system.
[0006] Among existing related technologies, Chinese patent document CN113581015B, entitled "Safety Early Warning Method and Device for Fuel Cell Systems," discloses a safety diagnosis method for fuel cell vehicles. This method first acquires various onboard sensor data during vehicle operation or charging. Then, it uses a traditional discrete wavelet decomposition algorithm to perform multi-scale analysis on the onboard sensor data. Fixed global thresholds are set in each frequency band of the wavelet decomposition to extract abnormal mutation feature components. Finally, by judging whether these abnormal mutation feature components meet preset risk judgment conditions, early online identification and graded warning of potential safety risks in the fuel cell system are achieved. This patent application document... The data denoising and anomaly feature extraction stages employ wavelet decomposition technology with a fixed threshold. Its drawback lies in the lack of adaptability to dynamic physical environment changes due to the fixed global threshold. When hydrogen fuel cell vehicles are in conditions of drastic speed changes or extremely high ambient humidity, the noise energy introduced by the external environment often fluctuates sharply and generates modal aliasing with the real sensor anomaly signals. Since the patent application fails to introduce environmental indicators to dynamically correct the wavelet denoising threshold, the fixed execution threshold often appears too low when facing high-frequency, high-energy environmental noise, thus missing the noise, or too high when the environment is stable, thus weakening the real weak feature signals, ultimately reducing the accuracy of safety warnings for the fuel cell system. Summary of the Invention
[0007] The purpose of this invention is to overcome the shortcomings of the prior art and solve the technical problem that the traditional wavelet packet threshold denoising algorithm is prone to underreporting and false reporting of hydrogen leakage when acquiring driving data of hydrogen fuel cell vehicles, and cannot accurately identify potential leakage risks. The invention seeks to design and provide a method and system for acquiring driving data of hydrogen fuel cell vehicles.
[0008] To achieve the aforementioned objectives, this invention provides a method and system for acquiring driving data of hydrogen fuel cell vehicles. The main process includes the following steps: acquiring hydrogen concentration data, driving speed data, and ambient humidity data during vehicle operation; preprocessing these data to obtain hydrogen concentration data, driving speed data, and ambient humidity data within a time window; determining the fluctuation level of the hydrogen concentration data within the time window based on the median of the absolute values of all coefficients in the highest frequency subband obtained after wavelet packet decomposition, the variance of the coefficients, and the maximum energy value among the coefficients; and determining the time window based on the driving speed data values within the time window. The driving speed characteristic factor within the time window is determined based on the ambient humidity data within the time window. The actual interference level of the hydrogen concentration data within the time window is determined by combining the driving speed characteristic factor and the ambient humidity characteristic factor with the fluctuation level of the hydrogen concentration data within the time window. A general threshold is corrected based on the actual interference level of the hydrogen concentration data within the time window to determine the final execution threshold used for the hydrogen concentration data within the time window. Wavelet packet thresholding denoising is then performed on the hydrogen concentration data within the time window using the final execution threshold to obtain the denoised hydrogen concentration data within the time window.
[0009] This invention highlights the fluctuation characteristics of hydrogen concentration data affected by noise by calculating the fluctuation degree of hydrogen concentration data within a time window and utilizing the median, variance, and maximum energy of the highest frequency subband coefficients. It quantifies the intensity of external interference by calculating driving speed and ambient humidity characteristic factors, thus achieving noise intensity analysis. By combining driving speed, ambient humidity, and fluctuation degree within the time window to calculate the true interference level of hydrogen concentration data, it adaptively corrects the general threshold, enhancing the anti-interference capability of the wavelet packet threshold denoising algorithm in complex backgrounds. Finally, by obtaining the denoised hydrogen concentration data within the time window based on the final execution threshold, it achieves the cleaning of hydrogen concentration data from hydrogen fuel cell vehicles, improving the accuracy and robustness of hydrogen concentration data acquisition.
[0010] The preprocessing described in this invention to obtain hydrogen concentration data, driving speed data, and ambient humidity data within a time window includes: aligning the collected hydrogen concentration data, driving speed data, and ambient humidity data on timestamps, and then dividing the time window according to a preset length with the last moment on the time axis as the end point to obtain hydrogen concentration data, driving speed data, and ambient humidity data within the time window.
[0011] The degree of fluctuation in hydrogen concentration data within the time window described in this invention satisfies the expression: In the formula, For the first The degree of fluctuation in hydrogen concentration data within a time window For median operations, For the first The hydrogen concentration data within the time window, obtained after wavelet packet decomposition, is in the highest frequency subband. Each coefficient value It is the absolute value symbol. For the first The number of coefficients in the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within a time window. For the first The mean of all coefficients in the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within a time window. For the first The maximum energy value among all coefficients in the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within a time window. It is a natural exponential function.
[0012] This invention constructs a positive correlation function relationship between the product term of the variance of the highest frequency subband coefficient and the maximum energy value, and the ratio term of the median of the highest frequency subband coefficient. This enables the assessment of the fluctuation degree of hydrogen concentration data within a time window. The product term of the variance of the highest frequency subband coefficient and the maximum energy value reflects the high-frequency fluctuation intensity of the hydrogen concentration data within the time window, while the ratio term of the median of the highest frequency subband coefficient increases the contribution of the noise region. This results in a larger numerical fluctuation degree being calculated for the hydrogen concentration data affected by noise, thus providing a reliable basis for calculating the true interference degree of hydrogen concentration data within a time window.
[0013] The method for obtaining the driving speed feature factor within the time window described in this invention is as follows: the product of the average driving speed data within the time window and the average rate of change of the driving speed data within the time window is used as the first value, and the ratio between the first value and the maximum designed speed of the vehicle is used as the driving speed feature factor within the time window.
[0014] The method for obtaining the environmental humidity characteristic factor within the time window described in this invention is as follows: the ratio of the mean environmental humidity data within the time window to the saturated humidity is used as the environmental humidity characteristic factor within the time window.
[0015] The true interference level of the hydrogen concentration data within the time window described in this invention satisfies the expression: In the formula, For the first The true extent of interference in hydrogen concentration data within a time window. For the first The degree of fluctuation in hydrogen concentration data within a time window For the first Vehicle speed characteristic factors within a time window For the first Environmental humidity characteristics within a time window For the minimum normalization function, It is a natural exponential function.
[0016] This invention constructs a composite function that includes a positive correlation term for the fluctuation of hydrogen concentration data within a time window, as well as a product term between the driving speed characteristic factor and the ambient humidity characteristic factor. This function enables the assessment of the true interference level of hydrogen concentration data within a time window. The fluctuation term of hydrogen concentration data within the time window enhances the interference weight of the noise region, while the driving speed characteristic factor and ambient humidity characteristic factor amplify the contribution of external noise interference. This results in a larger true interference level being calculated for the hydrogen concentration data affected by noise, thereby effectively distinguishing the real leakage signal from the background noise.
[0017] The universal threshold described in this invention satisfies the expression: In the formula, For the first The common threshold used for hydrogen concentration data within each time window For median operations, For the first The hydrogen concentration data within the time window, obtained after wavelet packet decomposition, is in the highest frequency subband. Each coefficient value It is the absolute value symbol. For the first The number of coefficients in the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within a time window, where ln is the natural logarithm function with base e.
[0018] The final execution threshold used for hydrogen concentration data within the time window described in this invention satisfies the expression: In the formula, For the first The final execution threshold used for hydrogen concentration data within each time window For the first The common threshold used for hydrogen concentration data within each time window This is the gain adjustment coefficient. For the first The true extent of interference in hydrogen concentration data within a given time window.
[0019] This invention dynamically adjusts the general threshold based on the actual interference level of hydrogen concentration data within a time window, achieving adaptive final execution threshold. In areas with strong noise, the threshold is automatically increased to suppress false filtering, while in areas with obvious leakage characteristics, a lower threshold is maintained to preserve the true response, ensuring the accuracy of the denoised hydrogen concentration data.
[0020] The present invention describes obtaining denoised hydrogen concentration data within a time window, comprising: for any coefficient in any frequency band sub-band, setting the coefficient value to zero in response to the absolute value of the coefficient being less than the final execution threshold, and shrinking the coefficient towards zero in response to the absolute value of the coefficient being not less than the final execution threshold; and reconstructing the hydrogen concentration data within the time window by performing wavelet packet inverse transform using all frequency band sub-bands and all processed coefficients of all frequency band sub-bands.
[0021] This invention denoises hydrogen concentration data by using the final execution threshold of hydrogen concentration data within a time window, enabling the noise-affected area to map the low-frequency trend characteristics of the real hydrogen concentration data, thus achieving accurate acquisition of hydrogen concentration data.
[0022] The present invention also provides a hydrogen fuel cell vehicle driving data acquisition system, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned hydrogen fuel cell vehicle driving data acquisition method is implemented. By generating a computer program from the above-mentioned method for acquiring driving data of hydrogen fuel cell vehicles and storing it in a memory, it can be loaded and executed by a processor, thereby creating a terminal device based on the memory and processor for convenient use.
[0023] Compared with the prior art, the present invention has at least the following beneficial effects: First, by introducing an adaptive final execution threshold correction mechanism based on multi-feature fusion, the technical problem of traditional wavelet packet threshold denoising algorithms easily confusing noise with leakage signals in hydrogen fuel cell vehicle environments, leading to false alarms or missed alarms for hydrogen leakage, is solved.
[0024] Secondly, by analyzing the median, variance, and maximum energy of the highest frequency subband coefficient, the intrinsic mapping relationship between the fluctuation of hydrogen concentration data and the hydrogen concentration data affected by noise was established. On this basis, the driving speed characteristic factor and the environmental humidity characteristic factor were further integrated, and the indicators reflecting the fluctuation intensity and the indicators reflecting external interference were closely coupled, thus realizing the accurate identification of noise background.
[0025] Third, it can calculate the true interference level of hydrogen concentration data for each time window. Through adaptive correction of the general threshold, it achieves accurate matching between the threshold and the local features of the hydrogen concentration data. For noise-affected areas, the threshold is automatically increased to prevent false filtering, while for actual leakage areas, the threshold is maintained to preserve integrity. Ultimately, this invention improves the accuracy and anti-interference capability of hydrogen concentration data acquisition for hydrogen fuel cell vehicles, enhances the robustness of noise reduction processing, and enables timely and accurate identification of potential leakage risks, providing a solid technical guarantee for vehicle safety monitoring and fault early warning. Attached Figure Description
[0026] Figure 1 This is a flowchart illustrating the method for acquiring driving data of hydrogen fuel cell vehicles, which relates to the present invention.
[0027] Figure 2 A schematic diagram comparing the denoising effects of using a general threshold in existing technologies and the adaptive final execution threshold used in this invention on hydrogen concentration data. Detailed Implementation
[0028] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments and accompanying drawings.
[0029] Example 1: This embodiment relates to a method for acquiring driving data of a hydrogen fuel cell vehicle, referring to... Figure 1 The main process includes steps S1 to S5, specifically: S1: Obtain hydrogen concentration data, driving speed data, and ambient humidity data.
[0030] Specifically, hydrogen concentration data is collected using a hydrogen concentration sensor installed in an area prone to hydrogen accumulation on the chassis, driving speed data is collected using the vehicle's ABS system, and ambient humidity data is collected using an onboard weather sensor. The collected hydrogen concentration data, driving speed data, and ambient humidity data are aligned on a timestamp, and then a time window is divided according to a preset length, with the last moment on the time axis as the end point, to obtain the hydrogen concentration data, driving speed data, and ambient humidity data within the time window. In this embodiment, the collection frequency of hydrogen concentration data, driving speed data, and ambient humidity data is 50Hz, and the length of the time window is set to 200. In other embodiments, the implementer can set the collection frequency of hydrogen concentration data, driving speed data, and ambient humidity data, and the length of the time window according to the actual implementation situation.
[0031] S2: Determine the degree of fluctuation in hydrogen concentration data.
[0032] It should be noted that existing wavelet packet thresholding denoising algorithms typically use general threshold rules to estimate noise levels when processing hydrogen concentration data, assuming the noise is uniformly distributed Gaussian white noise. This makes it difficult to adapt to the complex, non-stationary, and non-Gaussian environmental noise during vehicle operation, leading to a decrease in the accuracy of analyzing the true noise level of the current hydrogen concentration data. This, in turn, affects the adaptive adjustment effect of the subsequent denoising threshold. According to wavelet multi-resolution analysis theory, the actual hydrogen leakage process usually manifests as a gradual signal with energy concentrated in the low-frequency band, while circuit noise and environmental wind noise manifest as random fluctuation signals with energy concentrated in the high-frequency band. Furthermore, the statistical distribution of the high-frequency subband coefficients can reflect the noise intensity to some extent. Therefore, this invention determines an intuitive fluctuation intensity index for hydrogen concentration data to intuitively characterize the degree of disorder in the hydrogen concentration data within a time window from a frequency domain perspective.
[0033] Specifically, the wavelet packet thresholding denoising algorithm is used to decompose the hydrogen concentration data within the time window using wavelet packets to obtain sub-bands of all frequency bands and all coefficients corresponding to all frequency band sub-bands. In this embodiment, the wavelet basis function of the wavelet packet thresholding denoising algorithm is set to db4, and the number of wavelet packet decomposition layers is set to 3. In other embodiments, implementers can set the wavelet basis function and the number of wavelet packet decomposition layers of the wavelet packet thresholding denoising algorithm according to the actual implementation situation.
[0034] Specifically, the degree of fluctuation in hydrogen concentration data within the time window satisfies the expression: ; In the formula, For the first The degree of fluctuation in hydrogen concentration data within a time window For median operations, For the first The hydrogen concentration data within the time window, obtained after wavelet packet decomposition, is in the highest frequency subband. Each coefficient value It is the absolute value symbol. For the first The number of coefficients in the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within a time window. For the first The mean of all coefficients in the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within a time window. For the first The maximum energy value among all coefficients in the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within a time window is used to normalize the variance of all coefficients in the highest frequency subband. It is a natural exponential function.
[0035] In the formula, This is a classic estimate of the standard deviation of wavelet noise; the larger the value, the better. The higher the overall amplitude level of all coefficients in the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within a time window, the higher the overall amplitude level of all coefficients in the highest frequency subband. The greater the chance that hydrogen concentration data within a given time window will be affected by noise, the more likely the data will be to be affected by noise. The greater the fluctuation in hydrogen concentration data within a time window, the greater the volatility. The larger the value, the more likely it is to be the first. The more discrete the coefficients of the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within a time window, the more significant the difference between the first and second time windows. The greater the likelihood that fluctuations in hydrogen concentration data within a given time window are not white noise but rather abrupt changes, the more likely the first time window is to be affected. The greater the likelihood that the hydrogen concentration data within a given time window is affected by noise, the higher its reliability. The greater the fluctuation in hydrogen concentration data within a time window, the greater the volatility.
[0036] S3: Determine the true degree of interference in the hydrogen concentration data.
[0037] It should be noted that, due to the special operating conditions of hydrogen fuel cell vehicles traveling at high speeds and in high-humidity environments, the intense airflow turbulence in the chassis area and the water film effect on the sensor surface will superimpose to generate significant non-stationary background noise. This causes environmental noise to mix with the actual leakage signal, making it difficult to accurately retain weak leakage signals under strong noise interference. According to the physical interference coupling mechanism of the sensor, the vehicle's speed and acceleration directly determine the intensity of wind pressure turbulence acting on the hydrogen concentration sensor, while environmental humidity will nonlinearly amplify the hydrogen concentration sensor's response sensitivity to wind pressure noise by changing the electrochemical characteristics of the probe surface, i.e., the water film effect. Therefore, this invention determines the actual interference level of hydrogen concentration data within a time window to analyze the contribution of environmental interference to the apparent fluctuation intensity of hydrogen concentration data within the time window from a physical perspective, thereby guiding the adaptive correction of the denoising threshold.
[0038] Specifically, the product of the average driving speed data within the time window and the average rate of change of driving speed data within the time window is calculated to obtain the first value. The ratio between the first value and the maximum designed speed of the vehicle is calculated to obtain the driving speed characteristic factor within the time window. The ratio between the average ambient humidity data within the time window and the saturated humidity is calculated to obtain the ambient humidity characteristic factor within the time window.
[0039] Specifically, the true degree of interference in hydrogen concentration data within the time window satisfies the expression: ; In the formula, For the first The true extent of interference in hydrogen concentration data within a time window. For the first The degree of fluctuation in hydrogen concentration data within a time window For the first Vehicle speed characteristic factors within a time window For the first Environmental humidity characteristics within a time window For the minimum normalization function, It is a natural exponential function.
[0040] In the formula, The larger the value, the more likely it is to be the first. The greater the chance that hydrogen concentration data within a given time window will be affected by noise, the more likely the data will be to be affected by noise. The greater the degree of real interference in the hydrogen concentration data within a time window, the greater the interference. In the formula... The larger the value, the more likely it is to be the first. The larger the driving speed data and the rate of change of driving speed data within each time window, the more significant the change. The greater the wind pressure turbulence intensity acting on the hydrogen concentration sensor within the first time window, the more it indicates that the first... The greater the chance that hydrogen concentration data within a given time window will be affected by noise, the more likely the data will be to be affected by noise. The greater the degree of real interference in hydrogen concentration data within a time window. The larger the value, the more likely it is to be the first. The higher the ambient humidity data within a time window, the more it indicates that the... The greater the sensitivity of the hydrogen concentration sensor to wind pressure noise within each time window, the better. The greater the degree of real interference in hydrogen concentration data within a time window.
[0041] S4: Determine the final execution threshold for using hydrogen concentration data.
[0042] It should be noted that after obtaining the actual interference level of the hydrogen concentration data within the time window, this invention will correct the general threshold based on the actual interference level of the hydrogen concentration data within the time window. Traditional wavelet packet thresholding denoising algorithms typically use a general threshold to estimate the noise level when processing data. This invention, however, uses the actual interference level of the hydrogen concentration data within the time window to calculate the final execution threshold used for the hydrogen concentration data within the time window, so that the wavelet packet thresholding denoising algorithm can filter out noise more accurately.
[0043] Specifically, the general threshold satisfies the expression: ; In the formula, For the first The common threshold used for hydrogen concentration data within each time window For median operations, For the first The hydrogen concentration data within the time window, obtained after wavelet packet decomposition, is in the highest frequency subband. Each coefficient value It is the absolute value symbol. For the first The number of coefficients in the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within a time window, where ln is the natural logarithm function with base e.
[0044] Specifically, the final execution threshold used for hydrogen concentration data within the time window satisfies the expression: ; In the formula, For the first The final execution threshold used for hydrogen concentration data within each time window For the first The common threshold used for hydrogen concentration data within each time window For the first The true extent of interference in hydrogen concentration data within a time window. The gain adjustment coefficient is set to 2 in this embodiment. In other embodiments, the implementer can set it according to the actual implementation situation. For example, when the vehicle operating environment is harsh, wind noise and humidity interference are strong, and the requirements for the system's anti-false alarm capability are strict, the gain adjustment coefficient can be appropriately increased to enhance the threshold increase, thereby filtering out environmental noise more thoroughly. When it is necessary to monitor extremely weak early leakage signals and the system's detection sensitivity is required to be high, the gain adjustment coefficient can be appropriately decreased to reduce the threshold rise, thereby preserving the characteristics of weak real leakage signals as much as possible and improving the accuracy of defect detection.
[0045] In the formula, The larger the value, the more likely it is to be the first. The greater the likelihood that hydrogen concentration data within a given time window will be affected by noise, the higher the probability of the hydrogen concentration data being affected by noise. The larger the final execution threshold used for hydrogen concentration data within a time window, the more effectively noise can be filtered out.
[0046] S5: Obtain the noise-reduced hydrogen concentration data.
[0047] Specifically, the denoised hydrogen concentration data within the time window is obtained, including: For any coefficient in any frequency band sub-band corresponding to the hydrogen concentration data within the time window, the coefficient value is set to zero in response to the absolute value of the coefficient being less than the final execution threshold, and the coefficient is shrunk towards zero in response to the absolute value of the coefficient being not less than the final execution threshold. For any coefficient whose absolute value is not less than the final execution threshold, the coefficient value is subtracted from the threshold in response to the coefficient value being greater than the threshold, thus completing the coefficient shrinkage towards zero. In response to the negative number of the coefficient value being less than the threshold, the coefficient value is added to the threshold, thus completing the coefficient shrinkage towards zero.
[0048] By using the subbands of all frequency bands and all processed coefficients of all frequency band subbands, wavelet packet inverse transform is performed to reconstruct the hydrogen concentration data within the time window after denoising.
[0049] like Figure 2 As shown in the figure, this diagram compares the noise reduction performance of hydrogen concentration data under complex conditions of high speed and high humidity, using an adaptive final execution threshold as described in this invention versus a general threshold as described in existing technologies. The horizontal axis represents time, and the vertical axis represents hydrogen concentration. In the latter half of the time axis, in the region with strong noise interference, the curve processed by the existing technology exhibits violent oscillations and deviates significantly from the true trend, indicating that it cannot effectively filter out environmental noise. In contrast, the curve of the method described in this invention closely follows the reference line of the true hydrogen leakage trend, with a smooth trajectory and no obvious noise. This directly demonstrates that the present invention, through its adaptive threshold correction mechanism, can accurately eliminate non-stationary interference caused by speed and humidity in harsh environments, significantly improving the accuracy of hydrogen concentration data while preserving the true characteristics of hydrogen leakage.
[0050] This embodiment also discloses a hydrogen fuel cell vehicle driving data acquisition system, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a hydrogen fuel cell vehicle driving data acquisition method according to the present invention is implemented. The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the setup and functions of which are known in the art.
Claims
1. A method for acquiring driving data of a hydrogen fuel cell vehicle, characterized in that the steps include... include: The hydrogen concentration data, driving speed data, and ambient humidity data during vehicle operation are acquired, preprocessed, and then used to obtain hydrogen concentration data, driving speed data, and ambient humidity data within a time window. The degree of fluctuation of hydrogen concentration data within the time window is determined based on the median of the absolute values of all coefficients in the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within the time window, the variance of the coefficients, and the maximum energy value among the coefficients. The driving speed characteristic factor within the time window is determined based on the driving speed data value within the time window. The environmental humidity characteristic factor within the time window is determined based on the environmental humidity data value within the time window. Based on the driving speed characteristic factor and the environmental humidity characteristic factor within the time window, combined with the fluctuation degree of hydrogen concentration data within the time window, the true interference degree of hydrogen concentration data within the time window is determined. The general threshold is corrected based on the actual interference level of the hydrogen concentration data within the time window, and the final execution threshold used for the hydrogen concentration data within the time window is determined. The hydrogen concentration data within the time window is denoised using wavelet packet thresholding based on the final execution threshold to obtain the denoised hydrogen concentration data within the time window.
2. The method for acquiring driving data of a hydrogen fuel cell vehicle according to claim 1, characterized in that, The preprocessing process to obtain hydrogen concentration data, driving speed data, and ambient humidity data within a time window includes: aligning the collected hydrogen concentration data, driving speed data, and ambient humidity data on timestamps, and then dividing the time window according to a preset length with the last moment on the time axis as the end point to obtain hydrogen concentration data, driving speed data, and ambient humidity data within the time window.
3. The method for acquiring driving data of a hydrogen fuel cell vehicle according to claim 1, characterized in that, The degree of fluctuation in hydrogen concentration data within the time window satisfies the expression: ； In the formula, For the first The degree of fluctuation in hydrogen concentration data within a time window For median operations, For the first The hydrogen concentration data within the time window, obtained after wavelet packet decomposition, is in the highest frequency subband. Each coefficient value It is the absolute value symbol. For the first The number of coefficients in the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within a time window. For the first The mean of all coefficients in the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within a time window. For the first The maximum energy value among all coefficients in the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within a time window. It is a natural exponential function.
4. The method for acquiring driving data of a hydrogen fuel cell vehicle according to claim 1, characterized in that, The method for obtaining the driving speed characteristic factor within the time window is as follows: the product of the average driving speed data within the time window and the average rate of change of the driving speed data within the time window is used as the first value, and the ratio between the first value and the maximum designed speed of the vehicle is used as the driving speed characteristic factor within the time window.
5. The method for acquiring driving data of a hydrogen fuel cell vehicle according to claim 1, characterized in that, The method for obtaining the environmental humidity characteristic factor within the time window is as follows: the ratio of the mean environmental humidity data within the time window to the saturated humidity is used as the environmental humidity characteristic factor within the time window.
6. A method for acquiring driving data of a hydrogen fuel cell vehicle according to claim 1, 4, or 5, characterized in that, The true degree of interference in the hydrogen concentration data within the time window satisfies the expression: ； In the formula, For the first The true extent of interference in hydrogen concentration data within a time window. For the first The degree of fluctuation in hydrogen concentration data within a time window For the first Vehicle speed characteristic factors within a time window For the first Environmental humidity characteristics within a time window For the minimum normalization function, It is a natural exponential function.
7. The method for acquiring driving data of a hydrogen fuel cell vehicle according to claim 1, characterized in that, The general threshold satisfies the expression: ； In the formula, For the first The common threshold used for hydrogen concentration data within each time window For median operations, For the first The hydrogen concentration data within the time window, obtained after wavelet packet decomposition, is in the highest frequency subband. Each coefficient value It is the absolute value symbol. For the first The number of coefficients in the highest frequency subband obtained after wavelet packet decomposition of hydrogen concentration data within a time window, where ln is the natural logarithm function with base e.
8. A method for acquiring driving data of a hydrogen fuel cell vehicle according to claim 1 or 7, characterized in that, The final execution threshold used for the hydrogen concentration data within the time window satisfies the expression: ； In the formula, For the first The final execution threshold used for hydrogen concentration data within each time window For the first The common threshold used for hydrogen concentration data within each time window This is the gain adjustment coefficient. For the first The true extent of interference in hydrogen concentration data within a given time window.
9. The method for acquiring driving data of a hydrogen fuel cell vehicle according to claim 1, characterized in that, The process of obtaining the denoised hydrogen concentration data within the time window includes: for any coefficient in any frequency band sub-band, setting the coefficient value to zero in response to the absolute value of the coefficient being less than the final execution threshold, and shrinking the coefficient towards zero in response to the absolute value of the coefficient being not less than the final execution threshold; and reconstructing the hydrogen concentration data within the time window by performing wavelet packet inverse transform using all frequency band sub-bands and all processed coefficients of all frequency band sub-bands.
10. A system for acquiring driving data of a hydrogen fuel cell vehicle, characterized in that, Its main structure includes a processor and a memory, the memory storing computer program instructions, which, when executed by the processor, implement a method for acquiring driving data of a hydrogen fuel cell vehicle according to any one of claims 1-9.