Motorcycle muffler electrical performance test method and system

Abstract

The application discloses a motorcycle muffler electrical performance test method and system, and relates to the technical field of function test.In the method, firstly, the sensor plate spacing change amount and the wire harness connector deformation coefficient are obtained and filtered to obtain a cumulative offset sequence;when the change rate and the acceleration of the sequence meet specific abnormal conditions, it is determined that the connector has abnormal displacement, and the contact resistance and the capacitance drift sequence are extracted;then, the frequency domain features are extracted by calculating the derivatives, the phase difference and the frequency spectrum distribution of the two, and the first feature vector is generated based on the peak value and the energy interval;next, the electrical response component and the attitude angular velocity component are spliced and normalized into the second feature vector;finally, the inner product of the angular velocity direction cosine and the electrical weight coefficient is calculated, and if the correlation degree exceeds the threshold value, it is determined that the coupling is strong, and the internal relation matrix is generated accordingly.The technical scheme provided by the application can improve the accuracy of the muffler electrical performance test.

Description

Technical Field

[0001] This application belongs to the field of functional testing technology, specifically a method and system for testing the electrical performance of a motorcycle muffler. Background Technology

[0002] With increasingly stringent motorcycle emission regulations and continuous improvements in vehicle intelligence, the muffler, as a core component of the exhaust system, relies heavily on the stability of its built-in sensors' electrical performance, which directly impacts engine control accuracy and emission compliance. Muffler sensors must maintain measurement accuracy and electrical connection reliability over long periods in harsh operating environments characterized by high temperature, high pressure, and strong vibration. To meet the needs of engines with different displacements, operating conditions, and service life requirements, muffler electrical performance testing technologies are showing diversified development trends. In particular, intelligent diagnostic systems based on multi-parameter fusion demonstrate significant advantages in fault identification and performance evaluation. Their testing accuracy and stability in complex vibration environments are crucial for ensuring sensor reliability and extending muffler lifespan.

[0003] Currently, the electrical performance testing methods for motorcycle mufflers mainly employ static or dynamic monitoring techniques based on a single electrical parameter. These methods determine the sensor's operating status by measuring electrical parameters such as contact resistance and capacitance. This single-parameter-based testing method evaluates sensor performance by setting acceptable ranges for resistance or capacitance and monitoring the degree of parameter deviation. It directly reflects the sensor's electrical characteristics and is widely used in sensor testing under normal operating conditions. To improve detection sensitivity, this method also enhances the ability to identify abnormal states by increasing the sampling frequency and reducing the judgment threshold.

[0004] However, due to the unique working environment and complex electrical connection structure of mufflers, the practical application scenarios for muffler electrical performance testing are becoming increasingly complex and varied. In practical applications, testing methods based on a single electrical parameter are insufficient to suppress the interference of factors such as mechanical vibration, thermal expansion, and installation stress on the test results, and the accuracy of determining the root cause of muffler electrical performance degradation is difficult to guarantee. In particular, for internal sensors of mufflers, due to the narrow spacing between their plates and the vibration and impact they experience during operation, the wiring harness connectors are prone to slight displacement, forming localized poor contact areas. Existing testing methods, which use a single electrical parameter for performance evaluation, are prone to insufficient correlation between the test signal and changes in the sensor's physical state, or even failure to effectively identify electrical performance fluctuations caused by mechanical factors, thereby reducing the accuracy of muffler electrical performance testing. Summary of the Invention

[0005] To address the above issues, this application provides a method and system for testing the electrical performance of motorcycle mufflers, which solves the problems that existing testing methods have difficulty in effectively identifying the impact of changes in physical state on electrical performance and the lack of accuracy in determining the root causes of muffler electrical performance degradation, thereby improving the accuracy of muffler electrical performance testing.

[0006] To achieve the above objectives, the technical solution adopted in this application is as follows: The change in the electrode spacing of the internal sensor of the muffler and the deformation coefficient of the wire harness connector material are obtained, and the change in the electrode spacing and the deformation coefficient of the wire harness connector material are filtered to obtain the cumulative offset sequence. Extract the rate of change and acceleration component of the cumulative offset sequence. When the rate of change exceeds the preset threshold for the change in electrode spacing and the acceleration component shows non-linear growth, it is determined that the wire harness connector has an abnormal displacement, and the contact resistance sequence and capacitance drift sequence are obtained. Calculate the time derivative of the contact resistance sequence and the phase difference of the capacitance drift sequence, and extract the spectral distribution of the main frequency bandwidth of the time derivative of the resistance sequence and the phase difference of the capacitance drift sequence to obtain the frequency domain characteristics of the exhaust monitoring signal. Extract the peak position and energy concentration interval of the frequency domain features, calculate the ratio of the time domain interval of the peak position to the span of the energy concentration interval, and generate a first feature vector; The electrical response feature component of the first feature vector is concatenated with the angular velocity component of the attitude sequence corresponding to the first feature vector and normalized to obtain the second feature vector. Extract the direction cosine of the attitude sequence angular velocity component and the weight coefficient of the electrical response feature component from the second feature vector, and calculate the inner product of the direction cosine and the weight coefficient of the electrical response feature component to obtain the correlation value. If the correlation value is greater than a preset threshold, it is determined that there is a strong coupling relationship between the electrical response feature component and the attitude sequence angular velocity component, and an intrinsic relationship matrix is ​​generated using the correlation value as matrix elements.

[0007] By employing the above technical solution, the rate of change and acceleration components of the cumulative offset sequence are extracted. Abnormal displacement of the wiring harness connector is determined when the rate of change exceeds a preset threshold for the change in electrode spacing and the acceleration component exhibits non-linear growth. This enables early identification of electrical performance fluctuations caused by mechanical factors. Calculating the time derivative of the contact resistance sequence and the phase difference of the capacitance drift sequence reveals the dynamic correlation between the two electrical parameters, overcoming the limitations of traditional single-parameter testing methods. The frequency domain characteristics of the exhaust monitoring signal are obtained by extracting the spectral distribution of the dominant bandwidth of the time derivative of the contact resistance sequence and the phase difference of the capacitance drift sequence. The ratio of the time interval of the peak position to the span of the energy concentration interval is calculated to generate a first feature vector, transforming the complex multi-parameter electrical signal into a feature expression with clear physical meaning. The electrical response feature components of the first feature vector are concatenated with the angular velocity components of the attitude sequence and normalized to obtain a second feature vector. This achieves a unified representation of electrical parameters and mechanical motion states, establishing a sufficient correlation between the test signal and changes in the sensor's physical state. The correlation value is obtained by calculating the inner product of the direction cosine of the attitude sequence angular velocity component and the weight coefficient of the electrical response feature component in the second feature vector. When the correlation value is greater than the preset threshold, it is determined that there is a strong coupling relationship between the electrical response feature component and the attitude sequence angular velocity component. This can accurately identify the specific impact of physical factors such as mechanical vibration, thermal expansion, and installation stress on electrical performance, thereby improving the accuracy of determining the root cause of the muffler's electrical performance degradation.

[0008] In a second aspect, embodiments of this application provide a motorcycle muffler electrical performance testing system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, and the memory is used to store computer program code, which includes computer instructions, and the one or more processors call the computer instructions to cause the motorcycle muffler electrical performance testing system to perform the method described in the first aspect and any possible implementation thereof.

[0009] In summary, one or more technical solutions provided in this application have at least the following technical effects or advantages: By acquiring the changes in the electrode spacing of the sensors inside the muffler and the deformation coefficient of the wiring harness connector material, and performing filtering to obtain the cumulative offset sequence, and then extracting its rate of change and acceleration components for nonlinear growth judgment, it can accurately capture the minute displacement of the wiring harness connector caused by changes in physical state such as mechanical vibration. This overcomes the shortcomings of existing technologies that cannot effectively identify mechanical interference by relying on a single electrical parameter, and achieves effective judgment of abnormal displacement. After determining that abnormal displacement has occurred in the wiring harness connector, the contact resistance sequence and capacitance drift sequence are further acquired. By calculating their time derivative and phase difference, and extracting the main frequency bandwidth and spectral distribution, the frequency domain characteristics of the exhaust monitoring signal are obtained. Then, the peak position and energy concentration interval are extracted to generate the first feature vector. This process deeply mines the dynamic change characteristics of electrical parameters and establishes the relationship between electrical signal fluctuations and changes in physical state. This method effectively expresses the characteristics of the signal, overcoming the deficiency in existing technologies where the correlation between test signals and changes in the physical state of sensors is insufficient. Based on this, the electrical response feature component of the first feature vector is concatenated with the corresponding attitude sequence angular velocity component and normalized to obtain the second feature vector. The correlation value is obtained by calculating the inner product of the direction cosine of the attitude sequence angular velocity component and the weight coefficient of the electrical response feature component, accurately quantifying the intrinsic relationship between electrical performance fluctuations and mechanical vibration attitude. Finally, the correlation value is used to determine whether there is a strong coupling relationship between the electrical response feature component and the attitude sequence angular velocity component, generating an intrinsic relationship matrix. This allows for accurate tracing and location of the mechanical root cause of muffler electrical performance degradation, solving the problem that existing testing methods struggle to effectively identify the impact of physical state changes on electrical performance, thereby improving the accuracy of determining the root cause of muffler electrical performance degradation. Attached Figure Description

[0010] Figure 1 This is a flowchart illustrating a method for testing the electrical performance of a motorcycle muffler, as disclosed in an embodiment of this application. Figure 2 This is another schematic flowchart of a method for testing the electrical performance of a motorcycle muffler disclosed in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of a motorcycle muffler electrical performance testing system provided in an embodiment of this application.

[0011] In the diagram: 301, Central Processing Unit; 302, Read-Only Memory; 303, Random Access Memory; 304, Bus; 305, Input / Output Interface; 306, Input Section; 307, Output Section; 308, Storage Section; 309, Communication Section; 310, Driver; 311, Removable Media. Detailed Implementation

[0012] To enable those skilled in the art to better understand the technical solution, the present application will be described in detail below with reference to the embodiments. The description in this section is only exemplary and explanatory, and should not be used to limit the scope of protection of the present application in any way.

[0013] It should be noted that, in this document, the terms "comprising," "including," and any other variations 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 a process, method, article, or apparatus. Specific examples have been used in this document to illustrate the principles and implementation methods of the technical solutions of this application. The above examples are only for the purpose of helping to understand the methods and core ideas of this application. The above are merely preferred embodiments of this application. It should be pointed out that, due to the limitations of written expression and the objective existence of infinite specific structures, those skilled in the art can make several improvements, modifications, or changes without departing from the principles of this application, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes, or combinations, or the direct application of the concept and technical solutions of this application to other situations without modification, should all be considered within the scope of protection of this application.

[0014] This application provides a method for testing the electrical performance of a motorcycle muffler, referring to... Figure 1 , Figure 1 This is a flowchart illustrating a method for testing the electrical performance of a motorcycle muffler according to an embodiment of this application. The method is applied to a system, which refers to a hardware and software integrated platform capable of executing a motorcycle muffler electrical performance testing program. The system can execute a motorcycle muffler electrical performance testing program, and the method includes steps 101 to 107, as follows: Step 101: Obtain the change in electrode spacing of the internal sensor of the muffler and the deformation coefficient of the wire harness connector material, and filter the change in electrode spacing and the deformation coefficient of the wire harness connector material to obtain the cumulative offset sequence.

[0015] In this embodiment, the change in electrode spacing represents the change in distance between the capacitor plates of the sensor inside the muffler relative to the initial state. For example, when the distance changes from an initial 0.5 mm to 0.52 mm, the change in electrode spacing is 0.02 mm. The deformation coefficient of the wire harness connector material is a quantitative indicator of the degree of deformation of the wire harness connector material under stress. For example, the ratio of the deformation of the connector material under 100 Newtons of pressure to its original size is 0.003. Filtering refers to the process of noise suppression and smoothing of the acquired signal data, such as using moving average filtering to remove high-frequency noise components. The cumulative offset sequence refers to the sequence of displacement values ​​accumulated over time obtained by integrating the change in electrode spacing and the deformation coefficient of the wire harness connector material. For example, the data sequence formed by the cumulative offset of 0.01 mm in the first second, 0.025 mm in the second second, and 0.042 mm in the third second.

[0016] Specifically, firstly, real-time data on the electrode spacing of the sensors inside the muffler is collected using a displacement sensor. The difference between the electrode spacing at each sampling time and the initial calibration value is calculated to obtain the change in electrode spacing. Simultaneously, strain data of the wire harness connector material is collected using a strain sensor. The strain value is converted into the material elastic parameters to obtain the wire harness connector material deformation coefficient. Next, the length parameter of the filtering window is determined based on the current data acquisition frequency. A Gaussian function kernel weight distribution is constructed, and a weighted average operation is performed on the electrode spacing change sequence and the wire harness connector material deformation coefficient sequence to achieve smoothing filtering. Then, point-by-point numerical integration is performed on the filtered electrode spacing change sequence and the wire harness connector material deformation coefficient sequence. The change at each time point is multiplied by the corresponding time step and summed to obtain the cumulative offset value at each time point. The cumulative offset values ​​at all times are arranged in chronological order to form a cumulative offset sequence.

[0017] Step 102: Extract the rate of change and acceleration components of the cumulative offset sequence. When the rate of change exceeds the preset threshold for the change in electrode spacing and the acceleration component shows non-linear growth, it is determined that the wire harness connector has an abnormal displacement, and the contact resistance sequence and capacitance drift sequence are obtained.

[0018] In this embodiment, the rate of change represents the speed at which the value in the cumulative offset sequence changes over time. For example, if the cumulative offset increases from 0.01 mm to 0.025 mm from the 1st second to the 2nd second, the rate of change is 0.015 mm per second. The acceleration component refers to the derivative of the rate of change with respect to time, used to characterize the speed of change of the offset itself. For example, if the rate of change increases from 0.015 mm per second to 0.030 mm per second, the acceleration component is 0.015 mm per square second. The preset threshold for the change in electrode spacing refers to a pre-set critical value standard used to determine whether the change in electrode spacing is abnormal, for example, set to 0.02 mm per second. Nonlinear growth refers to the fact that the value change of the acceleration component does not conform to the linear function law but shows a curvilinear growth trend, for example, the acceleration component increases in the sequence of 0.01, 0.03, 0.08, and 0.20. A contact resistance sequence refers to a data sequence formed by arranging the resistance values ​​measured at the wire harness connector in chronological order, such as a sequence where the resistance is 0.5 ohms in the first second and 0.52 ohms in the second second. A capacitance drift sequence refers to a data sequence formed by arranging the deviations of the sensor electrode capacitance values ​​from a standard value over time.

[0019] Specifically, the values ​​at adjacent time points in the cumulative offset sequence are differentially calculated. The cumulative offset at the next time point is subtracted from the cumulative offset at the previous time point, and then divided by the time interval to obtain the rate of change value at each time point. The rate of change sequence is differentially calculated again, and the rate of change at the next time point is subtracted from the rate of change at the previous time point, and then divided by the time interval to obtain the acceleration component sequence. The acceleration component sequence is linearly fitted using the least squares method, and the sum of squared deviations between the actual acceleration component and the corresponding value of the fitted line at each time point is calculated as the linear fitting residual. It is determined whether the maximum value in the rate of change sequence exceeds the preset threshold for the change in electrode spacing, and simultaneously whether the linear fitting residual exceeds the preset nonlinearity judgment standard value. When both judgment conditions are met, it is confirmed that the acceleration component exhibits nonlinear growth and the rate of change exceeds the standard, indicating abnormal displacement of the wiring harness connector. The electrical parameter monitoring process is initiated. The resistance value at the wiring harness connector is continuously collected using an ohmmeter to form a contact resistance sequence, and the capacitance value of the sensor electrode is continuously collected using a capacitance meter and subtracted from the standard capacitance value to obtain the capacitance drift sequence.

[0020] In one possible implementation, the rate of change and acceleration component of the cumulative offset sequence are extracted. When the rate of change exceeds a preset threshold for the change in electrode spacing and the acceleration component exhibits non-linear growth, it is determined that the wire harness connector has experienced abnormal displacement, and the contact resistance sequence and capacitance drift sequence are obtained. Specifically, this includes steps 1021-1022, as follows: Step 1021: Calculate the first and second time derivatives of the cumulative offset sequence to obtain the rate of change and acceleration components, respectively.

[0021] In this embodiment, the cumulative offset sequence represents a sequence of displacement values ​​obtained by integrating and accumulating the sensor electrode spacing and wire harness connector material deformation, arranged over time. For example, a data sequence with a cumulative offset of 0.01 mm in the first second, 0.025 mm in the second second, and 0.045 mm in the third second. The first-order time derivative refers to the rate of change of the sequence values ​​with respect to time, obtained by calculating the difference between values ​​at adjacent moments and dividing by the time interval. For example, if the cumulative offset changes from 0.01 mm to 0.025 mm in 1 second, the first-order time derivative is 0.015 mm / s. The second-order time derivative represents the result of further differentiating the first-order time derivative with respect to time, reflecting the rate of change itself. For example, if the rate of change increases from 0.015 mm / s to 0.030 mm / s in 1 second, the second-order time derivative is 0.015 mm / s². The rate of change represents the speed at which the cumulative offset sequence changes over time, and has the same meaning as the first-order time derivative. For example, when the offset increases by 0.02 mm per second, the rate of change is 0.02 mm per second. The acceleration component refers to the derivative of the rate of change with respect to time, and has the same meaning as the second-order time derivative. It describes how fast the rate of change is. For example, when the rate of change increases by 0.01 mm per second, the acceleration component is 0.01 mm per square second.

[0022] Specifically, first, all values ​​in the cumulative offset sequence arranged in chronological order and their corresponding timestamps are obtained to determine the sequence length and the sampling time interval between adjacent time points. Then, each position in the sequence, starting from the second time point, is traversed. For the nth time point, the cumulative offset value of that time point is extracted and recorded as the current value. The cumulative offset value of the previous time point (n-1) is extracted and recorded as the previous value. The difference between the current value and the previous value is calculated and divided by the sampling time interval. The result is recorded as the first-order time derivative value of the nth time point. The same difference division operation is performed on all time points in the sequence except the first time point. The resulting first-order time derivative values ​​are arranged in chronological order to form a first-order time derivative sequence. This first-order time derivative sequence is identified as a rate of change sequence and stored in the data buffer. Next, a second derivative operation is performed on the first-order time derivative sequence. Starting from the second time point of the first-order time derivative sequence, the sequence is traversed. For the m-th time point, the first-order time derivative value at that time point is extracted and recorded as the current derivative value. The first-order time derivative value at the previous time point (m-1) is extracted and recorded as the preceding value. The difference between the current derivative value and the preceding value is calculated and divided by the sampling time interval to obtain the second-order time derivative value at the m-th time point. The same operation is performed on all time points in the first-order time derivative sequence except for the first time point. The obtained second-order time derivative values ​​are arranged in chronological order to form a second-order time derivative sequence. This sequence is identified as the acceleration component sequence, and the data acquisition is completed.

[0023] Step 1022: Perform linear fitting on the acceleration component and calculate the sum of squares of the linear fitting residuals; when the extreme value of the rate of change is greater than the preset threshold for the change in the electrode spacing and the sum of squares of the linear fitting residuals is greater than the preset nonlinearity judgment threshold, it is determined that the acceleration component exhibits nonlinear growth, and the operation of obtaining the contact resistance sequence and capacitance drift sequence is triggered.

[0024] In this embodiment, the acceleration component represents the second-order time derivative of the cumulative offset sequence, reflecting the rate of change of the offset. For example, at a certain moment, the acceleration component is 0.015 mm / s². Linear fitting refers to a mathematical method of approximating a set of data points using a linear function. For example, for data points with times ranging from 1 to 2 to 3 seconds and accelerations ranging from 0.01 to 0.02 to 0.03, a straight line with a slope of 0.01 is fitted. The sum of squares of the linear fitting residuals represents the sum of the squares of the differences between the actual values ​​of each data point and the predicted values ​​of the fitted straight line. For example, if the residuals of three data points are 0.001 to 0.002 to 0.001, then the sum of squares of the residuals is 0.000006. The rate of change is used to represent the first-order time derivative of the cumulative offset sequence. For example, when the offset increases by 0.02 mm per second, the rate of change is 0.02 mm / s². Extreme values ​​refer to the maximum or minimum values ​​in a rate of change sequence. For example, if the rate of change sequence is 0.01 to 0.03 to 0.02, the extreme value is 0.03. The preset threshold for electrode spacing variation is a pre-set critical value used to determine whether the rate of change is abnormal, for example, set to 0.025 mm / s. The preset nonlinearity judgment threshold represents the critical value for determining whether the acceleration component exhibits nonlinear changes, such as set to 0.00001. Nonlinear growth refers to the acceleration component's growth trend not conforming to a linear function law but exhibiting a curvilinear growth. The contact resistance sequence represents the data sequence of resistance values ​​at the wiring harness connector arranged over time. The capacitance drift sequence refers to the data sequence of sensor electrode capacitance values ​​deviating from the standard value arranged over time. A trigger operation indicates the action of initiating the execution of a specific functional process.

[0025] Specifically, firstly, all values ​​and corresponding timestamps of the acceleration component sequence are obtained, constructing a data point set with time as the independent variable and acceleration components as the dependent variable. The least squares method is used to fit a straight line to the data point set, setting the equation of the fitted line as the dependent variable equal to the slope multiplied by the independent variable plus the intercept. The slope and intercept parameters that minimize the sum of squared distances from all data points to the fitted line are calculated. The predicted acceleration component value for each time point is calculated based on the fitted line equation, and the predicted value is subtracted from the actual acceleration component value at that time point to obtain the residual. The residuals for all time points are squared, and the squared residuals are summed to obtain the sum of squared residuals for the linear fitting. The sequence of rates of change is traversed to find the element with the largest and smallest values, and the value with the larger absolute value is taken as the extreme value of the rate of change. The extreme value of the rate of change is compared with a preset threshold for the change in electrode spacing, and the sum of squared residuals for the linear fitting is also compared with a preset nonlinearity judgment threshold. When the extreme value of the rate of change exceeds the preset threshold for the change in electrode spacing and the sum of squares of the linear fitting residuals exceeds the preset threshold for nonlinearity determination, the marked acceleration component exhibits a nonlinear growth state. After completing the nonlinear growth determination, the resistance measurement program is started to continuously collect the resistance values ​​of the wire harness connectors using an ohmmeter and sort them by timestamp to form a contact resistance sequence. At the same time, the capacitance measurement program is started to collect the sensor electrode capacitance values ​​and calculate the difference between them and the standard capacitance values ​​using a capacitance meter and sort them by timestamp to form a capacitance drift sequence, thus completing the data acquisition operation trigger.

[0026] Step 103: Calculate the time derivative of the contact resistance sequence and the phase difference of the capacitance drift sequence, and extract the spectral distribution of the main frequency bandwidth of the time derivative of the resistance sequence and the phase difference of the capacitance drift sequence to obtain the frequency domain characteristics of the exhaust monitoring signal.

[0027] In this embodiment, the time derivative represents the rate of change of the resistance value with time in the contact resistance sequence. For example, if the resistance is 0.5 ohms in the first second and 0.53 ohms in the second second, the time derivative is 0.03 ohms per second. The phase difference refers to the angular difference between the instantaneous phase and the reference phase of the capacitance drift sequence. For example, if the instantaneous phase of the capacitance drift sequence is 45 degrees and the reference phase is 0 degrees, the phase difference is 45 degrees. The dominant bandwidth represents the frequency range where the energy of the time derivative signal is concentrated in the frequency domain. For example, if the energy is mainly distributed between 8 Hz and 12 Hz, the dominant bandwidth is 4 Hz. The spectral distribution refers to the amplitude distribution of the phase difference signal at different frequency points. For example, the amplitude is 0.8 at 10 Hz and 0.3 at 15 Hz. The frequency domain characteristics of the exhaust monitoring signal are used to represent the comprehensive characteristics of the contact resistance time derivative and the capacitance drift phase difference in the frequency domain space, including the combined information of the dominant bandwidth and spectral distribution.

[0028] Specifically, the resistance values ​​at adjacent time points in the contact resistance sequence are differentially calculated. The resistance value at the next time moment is subtracted from the resistance value at the previous time moment, and then divided by the time interval to obtain the time derivative sequence. A Hilbert transform is applied to the capacitance drift sequence to construct an analytic signal. The ratio of the imaginary to real parts of the analytic signal is extracted, and the arctangent function is calculated to obtain the instantaneous phase angle value. The instantaneous phase is subtracted from the initial phase to obtain the phase difference sequence. A fast Fourier transform is applied to the time derivative sequence to convert the time-domain signal into a frequency-domain signal. The frequency range with an amplitude exceeding twice the average amplitude is searched in the frequency-domain amplitude spectrum, and the difference between the upper and lower limits of this range is calculated to obtain the dominant frequency bandwidth. A fast Fourier transform is also applied to the phase difference sequence to obtain the frequency-domain amplitude spectrum. The amplitude data corresponding to each frequency point is recorded to form a spectral distribution array. The dominant frequency bandwidth value and the spectral distribution array are combined and encapsulated into a frequency-domain feature data structure for the exhaust monitoring signal.

[0029] In one possible implementation, the time derivative of the contact resistance sequence and the phase difference of the capacitance drift sequence are calculated, and the spectral distribution of the main frequency bandwidth of the time derivative of the resistance sequence and the phase difference of the capacitance drift sequence are extracted to obtain the frequency domain characteristics of the exhaust monitoring signal. Specifically, this includes steps 1031-1032, as follows: Step 1031: Perform differential operation on the contact resistance sequence to obtain the time derivative; perform Hilbert transform on the capacitor drift sequence to extract the analytic signal, and calculate the instantaneous phase of the analytic signal to obtain the phase difference.

[0030] In this embodiment, the contact resistance sequence refers to a data sequence of resistance values ​​at the wire harness connector arranged in chronological order, such as a sequence where the resistance is 0.5 ohms in the first second, 0.52 ohms in the second second, and 0.55 ohms in the third second. The difference operation refers to the operation of calculating the difference between values ​​at adjacent moments in the sequence and dividing by the time interval. For example, if the resistance changes from 0.5 ohms to 0.52 ohms in 1 second, the difference result is 0.02 ohms per second. The time derivative represents the rate of change of the contact resistance sequence with respect to time, that is, how quickly the sequence value changes with time. For example, if the resistance increases by 0.02 ohms per second, the time derivative is 0.02 ohms per second. The capacitance drift sequence refers to a sequence of data where the sensor electrode capacitance value deviates from the standard value, arranged chronologically, such as a sequence where the drift is 0.001 picofarads in the first second and 0.0015 picofarads in the second second. The Hilbert transform represents a mathematical operation that performs a special integral transformation on a real signal to obtain its orthogonal components, used to convert a real signal into a complex signal. An analytic signal is a complex signal composed of a real signal and its Hilbert transform, containing two components: a real part and an imaginary part. The instantaneous phase represents the phase angle value of the analytic signal at a specific moment, obtained by calculating the arctangent of the imaginary and real parts. For example, if the real part is 1 and the imaginary part is 1 at a certain moment, the instantaneous phase is 45 degrees. The phase difference is used to represent the phase angle change sequence obtained after the capacitance drift sequence undergoes Hilbert transform and instantaneous phase extraction.

[0031] Specifically, first, all resistance values ​​and corresponding timestamps in the contact resistance sequence, arranged chronologically, are obtained to determine the sequence length and sampling time interval. Starting from the second time point, the sequence is traversed. For each position in the sequence, the resistance value at the nth time point is extracted and recorded as the current value. The resistance value at the n-1th time point is extracted and recorded as the previous value. The difference between the current value and the previous value is calculated and divided by the sampling time interval to obtain the difference value at the nth time point. Differential division is performed sequentially on all time points except the first time point, and the resulting difference values ​​are arranged chronologically to form a time derivative sequence. Next, all values ​​in the capacitance drift sequence are obtained. A Hilbert transform is performed on the capacitance drift value at each time point, and the weighted integral of that point with all other time point values ​​is calculated to obtain the Hilbert transform output value. The original capacitance drift sequence values ​​are combined as the real part and the Hilbert transform output value as the imaginary part to form the analytic signal complex value for each time point. For each time point, the imaginary and real parts of the analytical signal are extracted from the complex values. The quotient of the imaginary part divided by the real part is calculated, and the arctangent function is applied to this quotient to obtain the angle value in radians, which is taken as the instantaneous phase at that time point. The instantaneous phase values ​​of all time points are arranged in chronological order to form a phase difference sequence, and the data extraction is completed.

[0032] Step 1032: Apply Fourier transform to the time derivative and phase difference respectively to convert the time domain signal into a frequency domain signal; extract the main frequency bandwidth corresponding to the time derivative and the spectral distribution corresponding to the phase difference from the frequency domain signal, and combine the main frequency bandwidth and spectral distribution to obtain the frequency domain characteristics of the exhaust monitoring signal.

[0033] In this embodiment, the time derivative represents the rate of change of the contact resistance sequence with respect to time; for example, the time derivative at a certain moment is 0.02 ohms per second. The phase difference represents the instantaneous phase angle change sequence obtained after processing the capacitance drift sequence; for example, the phase difference at a certain moment is 30 degrees. The Fourier transform is a mathematical transformation method that decomposes a time-domain signal into a superposition of sine waves of different frequencies, used to extract the frequency components of the signal. The time-domain signal represents the representation of the signal amplitude changing with time; for example, both the time derivative sequence and the phase difference sequence are time-domain signals. The frequency-domain signal refers to the representation of the signal in the frequency dimension, showing the amplitude and phase distribution of different frequency components. The dominant frequency bandwidth represents the frequency range where the energy is most concentrated in the frequency-domain signal corresponding to the time derivative; for example, if the dominant frequency is 50 Hz and the bandwidth is 10 Hz, then the dominant frequency bandwidth is 45 Hz to 55 Hz. The spectral distribution refers to the energy distribution at each frequency point in the frequency-domain signal corresponding to the phase difference, describing the intensity of different frequency components. The exhaust monitoring signal represents comprehensive signal data used to monitor the status of the exhaust system. Frequency domain features are used to represent the combination of characteristic parameters of a signal in the frequency domain, including the dominant frequency bandwidth and spectral distribution information.

[0034] Specifically, first, obtain all values ​​and corresponding timestamps of the time derivative sequence, and determine the sequence length as N. Apply a Discrete Fourier Transform to the time derivative sequence, traversing each frequency point from 0 to N-1. For the k-th frequency point, calculate the sum of the products of all time point values ​​in the time derivative sequence and their corresponding phase factors to obtain the complex frequency domain value of that frequency point. Arrange the complex frequency domain values ​​of all frequency points in frequency order to form the frequency domain signal of the time derivative. Next, perform the same Discrete Fourier Transform operation on the phase difference sequence, traversing all frequency numbers to calculate the complex frequency domain value of each frequency point, and arrange them in frequency order to form the frequency domain signal of the phase difference. Calculate the square of the magnitude of the complex value of each frequency point from the frequency domain signal of the time derivative as the energy value. Find the frequency point with the largest energy value and mark it as the main frequency point. Traverse from the main frequency point to both sides until the energy value drops to half of the main frequency point's energy value. Record the left and right frequency points at the half-maximum of energy, and calculate the difference between the right and left frequency points to obtain the main bandwidth. Energy values ​​at all frequency points are extracted from the frequency domain signal of the phase difference. A frequency-energy distribution curve is constructed using the frequency value as the x-axis and the energy value as the y-axis, representing the spectral distribution. The main frequency bandwidth value and the spectral distribution curve data are merged and stored in the same data structure, labeled as the frequency domain features of the exhaust monitoring signal, and feature extraction is completed.

[0035] Step 104: Extract the peak position and energy concentration interval of the frequency domain features, calculate the ratio of the time interval of the peak position to the span of the energy concentration interval, and generate the first feature vector.

[0036] In this embodiment, the peak position refers to the coordinates of the frequency point in the frequency domain feature where the amplitude reaches a local maximum. For example, if the amplitude at a frequency of 10 Hz is 0.9 and greater than the amplitude of the adjacent frequency point, then this point is the peak position. The energy concentration interval refers to a continuous frequency range in the frequency domain feature where the signal energy accounts for a preset proportion of the total energy. For example, if the energy in the frequency range from 5 Hz to 15 Hz accounts for 80% of the total energy, then this range is the energy concentration interval. The time domain interval represents the difference in time periods corresponding to adjacent peak positions. For example, if the time domain periods corresponding to the peak positions at frequencies of 10 Hz and 20 Hz are 0.1 seconds and 0.05 seconds respectively, then the time domain interval is 0.05 seconds. The span ratio refers to the ratio of the time domain interval to the frequency span of the energy concentration interval. For example, if the time domain interval is 0.05 seconds and the frequency span of the energy concentration interval is 10 Hz, then the span ratio is 0.005. The first feature vector is used to represent a vector data structure composed of multiple span ratios arranged according to frequency distribution, for example, a vector containing span ratios corresponding to 10 frequency points arranged in sequence.

[0037] Specifically, local maxima of the amplitude spectrum are searched within the frequency range corresponding to the main frequency bandwidth. Frequency positions where the amplitude is greater than that of the adjacent frequency points are marked as peak positions and their corresponding frequency values ​​are recorded. The total energy value is obtained by summing the squares of the amplitudes of each frequency point in the spectral distribution array. The energy values ​​of each frequency point are accumulated from low frequency to high frequency to form a cumulative energy sequence. The frequency point where the cumulative energy first reaches 80% of the total energy is found in the cumulative energy sequence as the upper limit of the energy concentration interval, and the frequency point where the cumulative energy first exceeds 10% of the total energy is found as the lower limit of the energy concentration interval. The reciprocal of the frequency value corresponding to each peak position is taken to obtain the time-domain period value. The difference in time-domain periods between adjacent peak positions is calculated to obtain the time-domain interval value. The difference between the upper and lower limit frequencies of the energy concentration interval is calculated to obtain the frequency span value. The time-domain interval is divided by the frequency span to obtain the span ratio. The span ratio is calculated for multiple peak positions within the main frequency bandwidth, and the obtained span ratios are arranged in ascending order of their corresponding frequencies to form the first feature vector.

[0038] In one possible implementation, the peak position and energy concentration interval of the frequency domain features are extracted, the ratio of the time domain interval of the peak position to the span of the energy concentration interval is calculated, and a first feature vector is generated. Specifically, this includes steps 1041-1043, as follows: Step 1041: Search for the frequency point corresponding to the maximum amplitude within the main frequency bandwidth as the peak position, and calculate the time domain interval between adjacent peak positions.

[0039] In this embodiment, the dominant bandwidth represents the frequency range where energy is concentrated in the time derivative frequency domain signal, for example, 45 Hz to 55 Hz. The maximum amplitude refers to the maximum value of the complex magnitude of all frequency points within the frequency range; for example, the maximum magnitude within the range is 7. A frequency point represents a single frequency sampling position in the frequency domain signal; for example, the 10th frequency point corresponds to 10 Hz. The peak position is used to represent the frequency point corresponding to the maximum amplitude within the dominant bandwidth; for example, if the maximum amplitude occurs at 48 Hz, then the peak position is 48 Hz. Adjacent peak positions refer to two adjacent peak positions when sorted by frequency; for example, the first peak is at 48 Hz and the second is at 52 Hz. The time domain interval represents the reciprocal of the frequency difference between adjacent peak positions; for example, if the frequency difference is 4 Hz, then the time domain interval is 0.25 seconds.

[0040] Specifically, first, the start and end frequencies of the main frequency bandwidth are obtained, and all frequency points within this range are extracted from the time derivative frequency domain signal. For each frequency point, the square root of the sum of the square of the real and imaginary parts of the complex value is calculated to obtain the amplitude. All frequency points are iterated through, and the maximum amplitude and its corresponding frequency point are recorded as the first peak position. Other local peaks are then searched, and for each frequency point, its amplitude is checked to see if it is greater than its left and right adjacent points; if so, it is marked as a peak position. All peak positions are sorted from low to high frequency, and the frequency values ​​of the first and second peak positions are extracted. The frequency difference between the second and first peak positions is calculated, and the reciprocal of the frequency difference is taken to obtain the first set of time-domain intervals. The frequency differences between the second and third peak positions, and between the third and fourth peak positions, are calculated sequentially, and their reciprocals are taken to obtain multiple time-domain interval values, which are stored sequentially to complete the calculation.

[0041] Step 1042: Calculate the cumulative energy distribution sequence of the spectrum distribution, extract the frequency interval in the cumulative energy distribution sequence that has an energy proportion that reaches a preset ratio as the energy concentration interval, and calculate the difference between the upper and lower limits of the energy concentration interval to obtain the span value.

[0042] In this embodiment, spectral distribution refers to the energy distribution at each frequency point of the phase difference frequency domain signal. The cumulative energy distribution sequence represents a sequence formed by accumulating the energy of each frequency point from the lowest frequency; for example, if the energy of the first three points is 1, 2, and 3, the cumulative sequence is 1, 3, and 6. Energy percentage refers to the ratio of the cumulative energy at a certain frequency point to the total energy; for example, if the total energy is 100 and the cumulative energy at a certain point is 75, the percentage is 0.75. A preset ratio represents a preset energy percentage threshold; for example, 0.95 means 95%. A frequency range refers to a continuous frequency range, such as 20 Hz to 80 Hz. An energy concentration range is used to represent the frequency range where the cumulative energy percentage reaches the preset ratio. The upper and lower limit difference refers to the difference between the highest and lowest frequencies in the range; for example, if the upper limit is 80 Hz and the lower limit is 20 Hz, the difference is 60 Hz. The span value represents the difference between the upper and lower limits of the energy concentration range.

[0043] Specifically, the process involves obtaining all frequency points and energy values ​​of the spectral distribution, sorted by frequency. The sum of all energy values ​​is calculated and recorded as the total energy. Energy is accumulated starting from the lowest frequency; the accumulated energy of the first frequency point equals its energy value, the second accumulated energy equals the sum of the previous two, and so on, forming a cumulative energy distribution sequence. For each frequency point, the cumulative energy is calculated and divided by the total energy to obtain the energy percentage. The process iterates from the lowest frequency to find the frequency point where the energy percentage first exceeds a preset proportion; the preceding frequency point is recorded as the lower limit of the energy concentration interval, and the highest frequency point as the upper limit. The span value is obtained by subtracting the lower limit frequency from the upper limit frequency, completing the calculation.

[0044] Step 1043: Divide the time interval by the span value to obtain the span ratio, arrange the span ratios according to the frequency distribution, and generate the first feature vector.

[0045] In this embodiment, the time-domain interval represents the reciprocal of the frequency difference between adjacent peak positions; for example, a frequency difference of 4 Hz corresponds to a time-domain interval of 0.25 seconds. The span value represents the difference between the upper and lower limits of the energy concentration range; for example, an upper limit of 80 Hz and a lower limit of 20 Hz corresponds to a span value of 60 Hz. The span ratio is the quotient obtained by dividing the time-domain interval by the span value; for example, a time-domain interval of 0.25 seconds and a span value of 60 Hz corresponds to a span ratio of 0.00417. The frequency distribution represents the way the span ratios are arranged according to their corresponding frequency positions, reflecting the distribution pattern of span ratios corresponding to different frequencies. Arrangement refers to the operation of organizing multiple span ratios into a sequence in a specific order; for example, arranging the span ratios corresponding to frequencies of 10 Hz, 20 Hz, and 30 Hz in sequence. The first feature vector is used to represent the vector composed of frequency domain feature values ​​of the exhaust monitoring signal, containing an array formed by arranging multiple span ratios according to the frequency distribution.

[0046] Specifically, all time-domain interval and span values ​​are obtained. The first span ratio is obtained by dividing the first time-domain interval value by the span value; the second span ratio is obtained by dividing the second time-domain interval value by the span value; and so on, calculating the quotient of each time-domain interval and span value to obtain multiple span ratios. The peak position frequency value corresponding to each time-domain interval is obtained. The first span ratio is associated with the first peak position frequency, and the second span ratio is associated with the second peak position frequency, establishing a correspondence between span ratios and frequencies. The span ratios are sorted from low to high frequency, and the sorted span ratios are stored sequentially in an array. The first element of the array is the span ratio corresponding to the lowest frequency, the second element is the span ratio corresponding to the second lowest frequency, and so on up to the highest frequency. The array storing the span ratios is marked as the first feature vector to complete the generation process.

[0047] Step 105: Concatenate the electrical response feature component of the first feature vector with the angular velocity component of the attitude sequence corresponding to the first feature vector and normalize them to obtain the second feature vector.

[0048] In this embodiment, the electrical response feature component represents the numerical elements reflecting the electrical performance response characteristics extracted from the first feature vector, such as a sub-vector composed of span ratios related to contact resistance changes and capacitance drift. The attitude sequence angular velocity component refers to the component sequence formed by arranging the angular velocity values ​​of the muffler rotating around three coordinate axes in three-dimensional space in time, for example, a three-dimensional vector sequence with an angular velocity of 2 radians per second around the X-axis, 1.5 radians per second around the Y-axis, and 0.8 radians per second around the Z-axis. Concatenation refers to the operation of joining two vectors end-to-end in element-wise order to form a new vector; for example, vector A is 1 to 2 to 3, and vector B is 4 to 5, concatenating them yields 1 to 2 to 3 to 4 to 5. Normalization refers to the transformation process of mapping each element in a vector to a uniform numerical range; for example, normalizing a vector from 0.5 to 5 to 50 to the range of 0 to 0.09 to 1. The second feature vector is used to represent the comprehensive feature vector containing electrical response characteristics and attitude angular velocity information after concatenation and normalization.

[0049] Specifically, the element position indices related to electrical parameters are extracted from the first feature vector. Based on the index positions, the corresponding span ratio data is read to form the electrical response feature component sub-vector. Angular velocity sampling data from the muffler's three-axis gyroscope is acquired during the test period. The X-axis, Y-axis, and Z-axis angular velocity sequences are aligned by time points and combined to form the attitude sequence angular velocity components. The values ​​of each element of the electrical response feature component are sequentially written to the beginning of the new vector. Then, the values ​​of each element of the attitude sequence angular velocity components are sequentially written to the end of the new vector to complete the splicing operation and obtain the spliced ​​feature vector. The maximum and minimum values ​​of all elements in the spliced ​​feature vector are calculated. For each element in the spliced ​​feature vector, the minimum value is subtracted, and then the result is divided by the difference between the maximum and minimum values. The result is multiplied by the preset upper limit of the target value range to obtain the normalized element value. The normalized element value replaces the corresponding element in the original spliced ​​feature vector to obtain the second feature vector.

[0050] It should be noted that the electrical response feature component is formed by extracting the position index of the element related to contact resistance and capacitance drift in the first feature vector and reading the corresponding span ratio data. It is used to characterize the electrical response characteristics of the wire harness connector. The attitude sequence angular velocity component is formed by using attitude sensors such as a three-axis gyroscope mounted on the muffler to collect the angular velocity data of the muffler around the X-axis, Y-axis and Z-axis in real time within the test time window, and then aligning and combining the three-axis angular velocity data synchronously according to the time points. It is used to characterize the spatial motion state of the muffler in a complex vibration environment.

[0051] In one possible implementation, the electrical response feature component of the first feature vector is concatenated with the angular velocity component of the attitude sequence corresponding to the first feature vector and normalized to obtain the second feature vector. Specifically, this includes steps 1051-1053, as follows: Step 1051: Extract the three-axis angular velocity data of the muffler within the test time window as the angular velocity components of the attitude sequence.

[0052] In this embodiment, the test time window represents the time period used to collect exhaust monitoring data, such as from second 0 to second 10. A muffler refers to a device in a vehicle's exhaust system used to reduce exhaust noise. The triaxial angular velocity data represents the muffler's rotational speed in three vertical directions, for example, an X-axis angular velocity of 2 degrees per second, a Y-axis angular velocity of 3 degrees per second, and a Z-axis angular velocity of 1 degree per second. The attitude sequence angular velocity components are used to represent the data sequence formed by the change of the muffler's triaxial angular velocity over time within the test time window.

[0053] Specifically, the start and end times of the test time window are first determined. X-axis, Y-axis, and Z-axis angular velocity data sequences are acquired from the three-axis gyroscope sensor mounted on the silencer. For the X-axis angular velocity data sequence, angular velocity values ​​are extracted sequentially at sampling time intervals starting from the start time until the end time, and the extracted values ​​are then arranged in chronological order to form the X-axis angular velocity component. For the Y-axis angular velocity data sequence, angular velocity values ​​are extracted sequentially at sampling time intervals starting from the start time until the end time, and the extracted values ​​are then arranged in chronological order to form the Y-axis angular velocity component. For the Z-axis angular velocity data sequence, angular velocity values ​​are extracted sequentially at sampling time intervals starting from the start time until the end time, and the extracted values ​​are then arranged in chronological order to form the Z-axis angular velocity component. The X-axis, Y-axis, and Z-axis angular velocity components are then combined to form the attitude sequence angular velocity component, completing the extraction process.

[0054] Step 1052: Obtain the maximum and minimum values ​​of the angular velocity components of the attitude sequence within the test time window. Subtract the minimum value from the angular velocity components of the attitude sequence and divide by the difference between the maximum and minimum values ​​to obtain the normalized result.

[0055] In this embodiment, the attitude sequence angular velocity component represents a data sequence formed by the change of the muffler's three-axis angular velocity over time within a test time window. The test time window represents the time period used to collect exhaust monitoring data, for example, from second 0 to second 10. The maximum value refers to the largest value among all values ​​in the attitude sequence angular velocity component, for example, a maximum of 8 degrees per second. The minimum value refers to the smallest value among all values ​​in the attitude sequence angular velocity component, for example, a minimum of -2 degrees per second. The difference represents the value obtained by subtracting the minimum value from the maximum value; for example, the difference between the maximum value of 8 degrees per second and the minimum value of -2 degrees per second is 10 degrees per second. The normalization result is used to represent the numerical sequence mapping the attitude sequence angular velocity component to a range of 0 to 1.

[0056] Specifically, the process involves obtaining all values ​​of the X-axis, Y-axis, and Z-axis angular velocity components from the attitude sequence. For the X-axis angular velocity component, the maximum and minimum values ​​are found, and the difference between the maximum and minimum values ​​is calculated to obtain the X-axis difference. For each X-axis angular velocity component, the minimum X-axis value is subtracted, and the result is divided by the X-axis difference to obtain the normalized X-axis component. Similarly, for the Z-axis angular velocity component, the minimum Z-axis value is subtracted, and the result is divided by the Z-axis difference to obtain the normalized Z-axis component. Finally, the normalized X-axis, Y-axis, and Z-axis components are combined to form the normalized result, completing the calculation.

[0057] Step 1053: Arrange the electrical response feature components and the normalization result to obtain the spliced ​​feature vector; map the spliced ​​feature vector to a preset numerical range to obtain the second feature vector.

[0058] In this embodiment, the electrical response feature components represent a sequence of characteristic values ​​extracted from the electrical characteristics of the exhaust monitoring signal. The normalization result represents the attitude sequence angular velocity component mapped to a numerical sequence within the range of 0 to 1. Arrangement refers to the operation of organizing multiple numerical sequences into a single sequence in a specific order. The concatenated feature vector represents the vector formed by arranging the electrical response feature components and the normalization result in order. Mapping refers to the operation of converting a value from one interval to another, for example, mapping 0.5 to the range of -1 to 1 to obtain 0. The preset numerical range represents a pre-defined target numerical range, such as -1 to 1. The second feature vector represents the vector after mapping the concatenated feature vector to the preset numerical range, used to characterize the comprehensive features of the exhaust monitoring signal.

[0059] Specifically, obtain all values ​​of the electrical response feature components and all values ​​of the normalized results. Create an empty vector as the concatenated feature vector. Place the first value of the electrical response feature component into the first position of the concatenated feature vector, the second value into the second position, and so on. Continue by placing the first value of the normalized result into the concatenated feature vector immediately following the electrical response feature component, the second value into the next position, and so on, to complete the concatenation. Obtain the lower and upper limits of the preset numerical range. Multiply each value in the concatenated feature vector by the upper limit minus the lower limit, and then add the lower limit to obtain the mapped value. Store all mapped values ​​in the vector in their original order, marking it as the second feature vector to complete the generation.

[0060] Step 106: Extract the direction cosine of the attitude sequence angular velocity component and the weight coefficient of the electrical response feature component from the second feature vector, and calculate the inner product of the direction cosine and the weight coefficient of the electrical response feature component to obtain the correlation value.

[0061] In this embodiment, the direction cosine represents the ratio of the projection of the angular velocity component onto each coordinate axis of the three-dimensional spatial coordinate system to the magnitude of the angular velocity vector. For example, when the angular velocity vector is projected onto the X-axis as 2, the Y-axis as 1, and the Z-axis as 2, and the vector magnitude is 3, the direction cosine of the X-axis is 0.67, the direction cosine of the Y-axis is 0.33, and the direction cosine of the Z-axis is 0.67. The weighting coefficient refers to the quantification coefficient of the contribution of each element in the electrical response characteristic component to the overall characteristic. For example, the weight of the first element is 0.4, the weight of the second element is 0.35, and the weight of the third element is 0.25. The inner product represents the mathematical result of multiplying the corresponding elements of two vectors and then summing them. For example, the inner product of vectors 1 to 2 to 3 and vectors 4 to 5 to 6 is 1 x 4 + 2 x 5 + 3 x 6, which equals 32. The correlation value is used to represent the quantification index of the degree of correlation between the direction cosine and the weighting coefficient vectors; the larger the value obtained through the inner product operation, the stronger the correlation between the two.

[0062] Specifically, the attitude sequence angular velocity component and the electrical response feature component are separated from the second feature vector based on the data length boundary. For the attitude sequence angular velocity component, angular velocity values ​​in the three coordinate axes are extracted. The square root of the sum of the squares of the three angular velocity values ​​is calculated to obtain the magnitude of the angular velocity vector. The angular velocity value in each coordinate axis direction is divided by the magnitude to obtain the direction cosine. The direction cosines of the three coordinate axes are arranged in the order X to Y to Z to form the direction cosines. For the electrical response feature component, the proportion of each element's value to the sum of all element values ​​is calculated. This proportion is used as the weight coefficient for the corresponding element. All weight coefficients are arranged in their original order to form a weight coefficient vector. The direction cosines and the weight coefficient vector are dimensionally aligned. The shorter vector is taken as the effective calculation length. The first element of the direction cosine is multiplied by the first element of the weight coefficient vector. The same multiplication operation is performed on subsequent elements. All product results are summed to obtain the correlation value.

[0063] In one possible implementation, the direction cosine of the attitude sequence angular velocity component and the weighting coefficients of the electrical response feature component are extracted from the second feature vector, and the inner product of the direction cosine and the weighting coefficients of the electrical response feature component is calculated to obtain the correlation value. Specifically, this includes steps 1061-1062, as follows: Step 1061: Obtain the data length of the electrical response feature component and the data length of the attitude sequence angular velocity component; based on the data length of the electrical response feature component and the data length of the attitude sequence angular velocity component, segment the second feature vector to separate the attitude feature sequence characterizing the spatial motion state of the muffler and the electrical feature sequence characterizing the electrical response of the wire harness connector.

[0064] In this embodiment, the electrical response feature component represents a sequence of feature values ​​extracted from the electrical characteristics of the exhaust monitoring signal. Data length refers to the number of values ​​contained in the sequence; for example, if a sequence contains 50 values, the data length is 50. The attitude sequence angular velocity component represents a data sequence formed by the change of the muffler's three-axis angular velocities over time within the test time window. The second feature vector represents the vector obtained by concatenating and mapping the electrical response feature components and the normalized angular velocity components. Segmentation refers to the operation of dividing the vector into sub-vectors based on the position index. The attitude feature sequence represents the separated sequence describing the muffler's attitude. The electrical feature sequence represents the separated sequence describing the electrical response.

[0065] Specifically, the number of values ​​in the electrical response feature components is denoted as the first length, and the number of values ​​in the attitude sequence angular velocity components is denoted as the second length. Starting from index 1 of the second feature vector, all values ​​up to the index of the first length are read sequentially and stored in a new array to form the electrical feature sequence. Starting from index 1 plus 1 of the second feature vector, all values ​​up to the index of the first length plus the second length are read sequentially and stored in another new array to form the attitude feature sequence. The length of the electrical feature sequence is verified to be equal to the first length, and the length of the attitude feature sequence is equal to the second length. The electrical feature sequence is labeled as representing the electrical response of the wiring harness connector, and the attitude feature sequence is labeled as representing the spatial motion state of the muffler, completing the segmentation.

[0066] Step 1062: Calculate the direction cosine of the angular velocity component of the attitude sequence relative to the preset reference coordinate axis in the attitude feature sequence, and extract the weight coefficients corresponding to the electrical response feature components in the electrical feature sequence; multiply the corresponding elements of the direction cosine and the weight coefficients and sum them to obtain the inner product and use it as the correlation value.

[0067] In this embodiment, the attitude feature sequence represents a numerical sequence describing the attitude changes of the muffler, separated from the second feature vector. The angular velocity component represents the rotational speed of the muffler along a specific axis. The preset reference coordinate axis refers to the coordinate axes in a pre-defined reference coordinate system, such as the X-axis, Y-axis, and Z-axis. The direction cosine represents the cosine of the angle between the vector and the coordinate axis; for example, if the vector makes a 60-degree angle with the X-axis, the direction cosine is 0.5. The electrical feature sequence represents a numerical sequence describing the electrical response characteristics, separated from the second feature vector. The weighting coefficient represents the numerical value indicating the importance of the electrical response feature components. Element-wise multiplication refers to the operation of multiplying the elements at the same position in two sequences. The inner product represents the sum of the elements-wise multiplications of two vectors. The correlation value represents the degree of association between the attitude features and the electrical features.

[0068] Specifically, the X-axis, Y-axis, and Z-axis angular velocity components are extracted from the attitude feature sequence. The square root of the sum of the squares of the X-axis, Y-axis, and Z-axis angular velocity components is taken to obtain the vector magnitude. The X-axis angular velocity component is divided by the vector magnitude to obtain the X-axis direction cosine, the Y-axis angular velocity component is divided by the vector magnitude to obtain the Y-axis direction cosine, and the Z-axis angular velocity component is divided by the vector magnitude to obtain the Z-axis direction cosine. Weight coefficients corresponding to the X, Y, and Z axes are obtained from the electrical feature sequence. The X-axis direction cosine is multiplied by the X-axis weight coefficient to obtain the first product, the Y-axis direction cosine is multiplied by the Y-axis weight coefficient to obtain the second product, and the Z-axis direction cosine is multiplied by the Z-axis weight coefficient to obtain the third product. The sum of the first, second, and third products is calculated to obtain the inner product, which is then marked as the correlation value to complete the calculation.

[0069] Step 107: If the correlation value is greater than the preset threshold, it is determined that there is a strong coupling relationship between the electrical response feature component and the attitude sequence angular velocity component, and an intrinsic relationship matrix is ​​generated using the correlation value as matrix element.

[0070] In this embodiment, the preset threshold represents a critical value used to determine whether the correlation value reaches the strong coupling standard. For example, setting it to 0.75 means that a correlation value exceeding 0.75 is considered a strong correlation. Strong coupling refers to a state where there is a significant mutual influence and correlation between the electrical response feature components and the attitude sequence angular velocity components. For example, if the electrical response feature components increase significantly and synchronously with the angular velocity, then there is a strong coupling relationship between the two. The intrinsic relationship matrix is ​​used to represent a two-dimensional matrix data structure formed by arranging the correlation values ​​at different spatial locations and time points in rows and columns. For example, rows represent the spatial coordinates of the electrical response feature components, columns represent timestamps, and matrix elements are the correlation values ​​at corresponding locations and times.

[0071] Specifically, the calculated correlation value is compared with a pre-stored preset threshold to determine if the correlation value is greater than the preset threshold. When the correlation value is greater than the preset threshold, a judgment operation is performed to mark a strong coupling relationship between the electrical response feature component and the attitude sequence angular velocity component and record the judgment result. The spatial coordinate information corresponding to the electrical response feature component is obtained as the row index identifier of the matrix, and the timestamp information corresponding to the attitude sequence angular velocity component is obtained as the column index identifier of the matrix. The number of rows and columns of the matrix is ​​determined according to the number of row indices and column indices, and a two-dimensional array of the corresponding size is initialized as the storage structure of the intrinsic relationship matrix. All combinations of electrical response feature components and timestamps are traversed, and the corresponding correlation value is calculated for each combination. The matrix row position is determined according to the spatial coordinates, and the matrix column position is determined according to the timestamp. The correlation value is filled into the element at the intersection of the corresponding row and column of the matrix. After filling all positions, the complete intrinsic relationship matrix is ​​obtained.

[0072] In one possible implementation, an intrinsic relationship matrix is ​​generated using relevance values ​​as matrix elements, specifically including step 1071, as follows: Step 1071: Using the spatial coordinates of the electrical response feature components as row indices and the timestamps of the attitude sequence angular velocity components as column indices, fill the correlation values ​​into the corresponding matrix positions to generate the intrinsic relationship matrix.

[0073] In this embodiment, spatial coordinates are the positional identifiers of the electrical response feature components in three-dimensional space, such as X-axis coordinate 1, Y-axis coordinate 2, and Z-axis coordinate 3. Row index indicates the row number position in the matrix. Attitude sequence angular velocity components represent the sequence of angular velocity values ​​of the muffler at various moments within the test period. Timestamps are time markers used to record the data acquisition time, such as the 1st millisecond, 2nd millisecond, etc. Column indexes indicate the column number position in the matrix. Correlation values ​​represent the calculated result of the degree of association between attitude features and electrical features. Matrix position refers to the position of a matrix element determined by both row and column indices. The intrinsic relationship matrix represents a two-dimensional array characterizing the intrinsic relationship between the electrical response feature components and the attitude sequence angular velocity components.

[0074] Specifically, the X-axis, Y-axis, and Z-axis coordinates corresponding to the electrical response feature components are obtained, and these three coordinate values ​​are used as 3D spatial position identifiers. An empty matrix is ​​created, mapping the spatial coordinates of the electrical response feature components to the row numbers of the matrix, for example, spatial coordinate 1 corresponds to row 1, spatial coordinate 2 corresponds to row 2. The timestamps of the attitude sequence angular velocity components at each acquisition time are obtained, and the timestamps are mapped to the column numbers of the matrix, for example, timestamp 100 milliseconds corresponds to column 1, timestamp 200 milliseconds corresponds to column 2. The spatial coordinates of all electrical response feature components are traversed, and the corresponding row number is found for each spatial coordinate. The timestamps of all angular velocity components are traversed, and the corresponding column number is found for each timestamp. Based on the row and column numbers determined by the current spatial coordinates and timestamps, the position is located in the matrix. The calculated correlation values ​​are filled into the located matrix positions. The filling process is repeated until all row and column positions are filled. The filled matrix is ​​marked as the intrinsic relationship matrix.

[0075] In the above embodiments, by constructing an intrinsic relationship matrix and extracting attitude-sensitive features, a quantitative description of the correlation between the change in capacitor plate spacing and the muffler's attitude angular velocity and a preliminary identification of vibration response characteristics were achieved. To further improve the accuracy of deformation cumulative effect calculation and the precision of time-series data synchronization processing, this application also provides a method for testing the electrical performance of a motorcycle muffler. This method obtains the current sampling frequency and adjusts the smoothing filter window length according to the sampling frequency; determines a Gaussian filter kernel based on the window length to smooth the change in plate spacing and the deformation coefficient of the wiring harness connector material to obtain a filtered sequence; interpolates the change in plate spacing using the sampling time of the deformation coefficient of the wiring harness connector material as a common time reference and establishes a timestamp alignment relationship; and performs numerical integration on the filtered deformation coefficient sequence and the filtered plate spacing change sequence to obtain a cumulative offset sequence. This accurately eliminates the interference of sampling noise on deformation data and solves the data misalignment problem caused by inconsistent sampling times of different sensors, providing high-precision time-synchronized deformation data and reliable cumulative offset calculation results for subsequent time-frequency analysis of cumulative offset and accurate assessment of vibration energy distribution. The following is a combination of... Figure 2 Another method for testing the electrical performance of a motorcycle muffler, as described in the embodiments of this application, is as follows: Please see Figure 2 This is another flowchart illustrating a method for testing the electrical performance of a motorcycle muffler in an embodiment of this application.

[0076] Step 201: Obtain the current sampling frequency and adjust the smoothing filter window length according to the current sampling frequency.

[0077] In this embodiment, the current sampling frequency represents the data acquisition rate. The sampling frequency refers to the number of times the signal is sampled per unit time; for example, a sampling frequency of 1000 Hz means 1000 samples per second. The smoothing filter window length represents the number of data points used during the filtering operation. The window length refers to the number of consecutive data points processed at once during the filtering process; for example, a window length of 5 means using 5 consecutive data points for filtering calculations each time.

[0078] Specifically, the current sampling frequency of the sensor is read. The magnitude of the current sampling frequency is determined. When the sampling frequency is high, the smoothing filter window length is set to a larger value, for example, 20 for a sampling frequency of 2000 Hz. When the sampling frequency is low, the smoothing filter window length is set to a smaller value, for example, 5 for a sampling frequency of 500 Hz. The ratio of the sampling frequency to the reference frequency is calculated, with the reference frequency set to 1000 Hz. This ratio is multiplied by the reference window length to obtain the adjusted window length, which is set to 10. The calculated window length value is rounded down. The adjusted window length is checked to see if it is odd; if it is even, it is incremented by 1 to convert it to an odd number to ensure window symmetry. The adjusted window length is verified to be within a preset valid range, with a lower limit of 3 and an upper limit of 51. The adjusted window length value is stored in the buffer area as a parameter for subsequent filtering operations.

[0079] Step 202: Determine the Gaussian filter kernel based on the smoothing filter window length, and use the Gaussian filter kernel to smooth the change in electrode spacing and the deformation coefficient of the wire harness connector material, so as to obtain the filtered deformation coefficient sequence and the filtered electrode spacing change sequence.

[0080] In this embodiment, the smoothing filter window length represents the number of data points used in the filtering operation. The Gaussian filter kernel represents a filter weight array constructed based on a Gaussian function. A filter kernel is a one-dimensional array containing weight coefficients for each position; for example, a filter kernel of length 5 contains 5 weight values. The electrode spacing change represents the numerical sequence of changes in the distance between the two electrodes of the capacitive sensor. The wire harness connector material deformation coefficient represents the numerical sequence of coefficients indicating the deformation of the wire harness connector material under stress. Smoothing filtering represents a method of eliminating data fluctuations through weighted averaging. The filtered deformation coefficient sequence represents the smoothed deformation coefficient data sequence. The filtered electrode spacing change sequence represents the smoothed electrode spacing change data sequence.

[0081] Specifically, a Gaussian filter kernel array of corresponding length is generated based on the smoothing filter window length. The center position index of the filter kernel is calculated, and the distance from each position within the window to the center position is calculated. The distances at each position are substituted into the Gaussian function to calculate the corresponding weight value. All weight values ​​are normalized so that the sum of the weights equals 1. The Gaussian filter kernel is aligned to the starting position of the electrode spacing change sequence. Each data point within the window is multiplied by its corresponding weight value and summed to obtain the filtering result at the current position. The filter kernel is moved one data point forward, and the weighted summation process is repeated. After traversing all data points in the electrode spacing change sequence, the filtered electrode spacing change sequence is obtained. The Gaussian filter kernel is applied to the deformation coefficient sequence of the wire harness connector material in the same way. The deformation coefficient sequence is weighted and summed to obtain the filtered deformation coefficient sequence. The two filtered sequences are stored separately.

[0082] In one possible implementation, before smoothing the change in electrode spacing and the deformation coefficient of the wire harness connector material using a Gaussian filter kernel, steps 2021-2022 are further included, as follows: Step 2021: Using the sampling time of the deformation coefficient of the wire harness connector material as a common time reference, interpolate the change in electrode spacing.

[0083] In this embodiment, the deformation coefficient of the wire harness connector material represents a numerical sequence of coefficients indicating the deformation of the wire harness connector material under stress. The sampling time represents the point in time when data is acquired. The sampling time refers to a specific time marker for recording data, such as the 100th millisecond, the 200th millisecond, etc. The common time reference represents a reference time standard used to unify the time axes of different data sequences. The time reference refers to a time sequence used as an alignment reference, such as using a specific time point of a data sequence as a reference. The change in electrode spacing represents a numerical sequence of changes in the distance between the two electrodes of the capacitive sensor. Interpolation processing represents the method of calculating intermediate values ​​between known data points. Interpolation is the calculation process of extrapolating values ​​at other times based on existing data points; for example, knowing the values ​​at the 100th millisecond and the 200th millisecond, calculating the value at the 150th millisecond.

[0084] Specifically, the process involves acquiring all sampling times of the deformation coefficient sequence of the wire harness connector material and using these times as the target time point sequence. The original sampling times of the electrode spacing change sequence are also acquired and used as the known time point sequence. Each sampling time of the deformation coefficient is iterated through, checking if a corresponding data point exists in the electrode spacing change sequence. If the target time exists in the known time point sequence, the electrode spacing change value at that time is directly read. If the target time does not exist in the known time point sequence, the two closest known time points before and after the target time are searched. The electrode spacing change values ​​corresponding to the two known time points are read. The time interval between the target time and the previous time point is calculated, and the total time interval between the next time point and the previous time point is calculated. The two time intervals are divided to obtain the interpolation weighting coefficient. The value of the next time point is subtracted from the value of the previous time point to obtain the numerical change. The numerical change is multiplied by the interpolation weighting coefficient to obtain the increment value. The value of the previous time point is added to the increment value to obtain the interpolation result for the target time. Interpolation calculations are performed sequentially for all target times to generate a sequence of changes in electrode spacing corresponding to the sampling times of the deformation coefficient.

[0085] Step 2022: Correspond point-by-point to the change in electrode spacing after interpolation with the deformation coefficient of the wire harness connector material, and establish a timestamp alignment relationship.

[0086] In this embodiment, the interpolated electrode spacing change represents the numerical sequence of electrode spacing changes calculated through interpolation. The wire harness connector material deformation coefficient represents the numerical sequence of coefficients that cause deformation of the wire harness connector material under stress. Point-to-point correspondence refers to the processing method of pairing two data sequences at the same time. Point-to-point correspondence means establishing a one-to-one mapping relationship between data at the same time point, for example, pairing the electrode spacing change at the 100th millisecond with the deformation coefficient at the 100th millisecond. Timestamp alignment relationship represents the correspondence between different data sequences in the time dimension. Alignment relationship refers to the association relationship that ensures the time points of multiple data sequences remain synchronized.

[0087] Specifically, the process involves: obtaining the first data point and its timestamp of the deformation coefficient sequence of the wire harness connector material; obtaining the first data point and its timestamp of the interpolated electrode spacing change sequence; verifying that the two timestamps are identical; pairing the deformation coefficient values ​​with the electrode spacing change values ​​at the same timestamp to form the first data pair; recording the timestamp, deformation coefficient value, and electrode spacing change value of this data pair; moving to the second time point and reading the values ​​of the deformation coefficient sequence and electrode spacing change sequence at that moment; verifying the consistency of the second timestamp and pairing the two values ​​to form the second data pair; repeating the data pairing process for all time points; creating a data structure containing the timestamp, deformation coefficient, and electrode spacing change at each time point; arranging all data pairs in chronological order to construct a complete correspondence sequence; checking the continuity and consistency of the timestamps for each data pair in the correspondence sequence; and storing the established timestamp alignment relationship in a relation table to complete the point-to-point correspondence between the two data sequences.

[0088] Step 203: Perform numerical integration on the filtered deformation coefficient sequence and the filtered electrode spacing change sequence to obtain the cumulative offset sequence.

[0089] In this embodiment, the filtered deformation coefficient sequence represents a smoothed deformation coefficient data sequence. The filtered electrode spacing change sequence represents a smoothed electrode spacing change data sequence. Numerical integration represents the calculation process of summing the discrete data sequences. Integration refers to accumulating the values ​​of each data point in the sequence one by one; for example, integrating the sequences 1, 2, and 3 yields 1, 3, and 6. The cumulative offset sequence represents the cumulative offset value sequence obtained through integration.

[0090] Specifically, the first value of the initial cumulative offset sequence is set to 0. The first data point value of the filtered deformation coefficient sequence is read. This value is multiplied by the time interval to obtain the offset for the first time interval. The first offset is added to the initial cumulative value to obtain the first cumulative offset value. The second data point value of the filtered deformation coefficient sequence is read and multiplied by the time interval to obtain the second offset. The second offset is added to the first cumulative offset to obtain the second cumulative offset value. This process is repeated for all data points in the filtered deformation coefficient sequence to obtain the cumulative offset sequence of deformation coefficients. The same method is used to perform numerical integration on the filtered electrode spacing change sequence. Each electrode spacing change value is read, multiplied by the time interval, and added to the cumulative value from the previous time step. After traversing all data points, the cumulative offset sequence of electrode spacing is obtained. The two cumulative offset sequences are then merged and stored.

[0091] It should be noted that the preset threshold for the change in electrode spacing should be set according to the vibration range of the muffler during normal operation. This can be achieved by collecting data on the change in electrode spacing during stable operation, calculating statistical characteristic values ​​such as the mean and standard deviation, and setting the threshold to the mean plus a certain multiple of the standard deviation to ensure effective differentiation between normal vibration and abnormal displacement.

[0092] A preset nonlinearity threshold is used to determine whether the acceleration component exhibits nonlinear growth. This can be achieved by performing linear fitting on known linear and nonlinear variation samples from historical data, statistically analyzing the residual sum of squares distributions for the two types of samples, and selecting a critical value that effectively distinguishes the two types of samples as the threshold.

[0093] The smoothing filter window length should be proportional to the sampling frequency. A reference sampling frequency and its corresponding reference window length can be set, and the window lengths for other sampling frequencies are calculated proportionally. It is also necessary to ensure that the window length is an odd number and within a reasonable range; too small a window will result in insufficient filtering, while too large a window will cause signal distortion.

[0094] The preset proportion of the energy concentration range is used to determine the degree of concentration of spectral energy. A selectable percentage of energy can be chosen to cover the main spectral characteristics; this is typically set to a proportion that covers the main frequency components of the signal, ensuring that the extracted frequency range contains sufficient signal information while avoiding excessive noise.

[0095] A preset threshold for the correlation value is used to determine whether there is a strong coupling relationship between the change in electrode spacing and the angular velocity components of the attitude sequence. This can be achieved by analyzing historical data of samples known to have strong and weak coupling, calculating their correlation value distribution, and selecting a critical value that effectively distinguishes between strong and weak coupling as the determination threshold.

[0096] The preset numerical range of the second feature vector is used to unify the value range of features with different dimensions. It can be set according to the requirements of subsequent processing for the numerical range. Commonly used ranges include zero to one or negative one to positive one, ensuring that feature components from different sources are processed and compared on the same scale.

[0097] The preset reference coordinate axis for the angular velocity direction cosine should be fixedly related to the muffler's installation direction and the testing environment. The muffler's main axis direction, the direction of gravity, or the vehicle's travel direction can be selected as the reference to ensure that the angular velocity components of different test batches and different test objects have a unified reference system, facilitating the comparison and analysis of results.

[0098] The following describes a motorcycle muffler electrical performance testing system from a hardware processing perspective, as described in this application embodiment. Please refer to [link / reference]. Figure 3 This is a schematic diagram of the structure of a motorcycle muffler electrical performance testing system according to an embodiment of this application.

[0099] It should be noted that, Figure 3 The structure of the motorcycle muffler electrical performance testing system shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0100] like Figure 3 As shown, a motorcycle muffler electrical performance testing system includes a Central Processing Unit (CPU) 301, which can perform various appropriate actions and processes according to a program stored in a Read-Only Memory (ROM) 302 or a program loaded from a storage section 308 into a Random Access Memory (RAM) 303, such as executing the methods described in the above embodiments. The RAM 303 also stores various programs and data required for system operation. The CPU 301, ROM 302, and RAM 303 are interconnected via a bus 304. An Input / Output (I / O) interface 305 is also connected to the bus 304.

[0101] The following components are connected to I / O interface 305: input section 306 including audio input devices, push-button switches, etc.; output section 307 including a liquid crystal display (LCD) and audio output devices, indicator lights, etc.; storage section 308 including a hard disk, etc.; and communication section 309 including a network interface card such as a LAN (Local Area Network) card, modem, etc. Communication section 309 performs communication processing via a network such as the Internet. Drive 310 is also connected to I / O interface 305 as needed. Removable media 311, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 310 as needed so that computer programs read from them can be installed into storage section 308 as needed.

[0102] The above description is merely a specific embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for testing the electrical performance of a motorcycle muffler, characterized in that, The method includes: The change in the electrode spacing of the internal sensor of the muffler and the deformation coefficient of the wire harness connector material are obtained, and the change in the electrode spacing and the deformation coefficient of the wire harness connector material are filtered to obtain the cumulative offset sequence. Extract the rate of change and acceleration components of the cumulative offset sequence. When the rate of change exceeds a preset threshold for the change in electrode spacing and the acceleration component exhibits non-linear growth, it is determined that the wire harness connector has experienced abnormal displacement, and the contact resistance sequence and capacitance drift sequence are obtained. Specifically, this includes: calculating the first and second time derivatives of the cumulative offset sequence to obtain the rate of change and the acceleration component, respectively; performing linear fitting on the acceleration component and calculating the sum of squared residuals of the linear fitting. When the extreme value of the rate of change is greater than a preset threshold for the change in the electrode spacing, and the sum of squares of the linear fitting residuals is greater than a preset nonlinearity determination threshold, it is determined that the acceleration component exhibits nonlinear growth, and the operation of acquiring the contact resistance sequence and the capacitance drift sequence is triggered. The process involves calculating the time derivative of the contact resistance sequence and the phase difference of the capacitance drift sequence, and extracting the spectral distribution of the dominant bandwidth of the time derivative of the resistance sequence and the phase difference of the capacitance drift sequence to obtain the frequency domain characteristics of the exhaust monitoring signal. Specifically, this includes: performing a differential operation on the contact resistance sequence to obtain the time derivative; performing a Hilbert transform on the capacitance drift sequence to extract an analytic signal, and calculating the instantaneous phase of the analytic signal to obtain the phase difference; applying Fourier transforms to the time derivative and the phase difference respectively to convert the time domain signal into a frequency domain signal; extracting the dominant bandwidth corresponding to the time derivative and the spectral distribution corresponding to the phase difference from the frequency domain signal, and combining the dominant bandwidth and the spectral distribution to obtain the frequency domain characteristics of the exhaust monitoring signal. Extract the peak position and energy concentration interval of the frequency domain features, calculate the ratio of the time domain interval of the peak position to the span of the energy concentration interval, and generate a first feature vector; The electrical response feature component of the first feature vector is concatenated with the angular velocity component of the attitude sequence corresponding to the first feature vector and normalized to obtain the second feature vector. Extracting the direction cosine of the attitude sequence angular velocity component and the weight coefficient of the electrical response feature component from the second feature vector, and calculating the inner product of the direction cosine and the weight coefficient of the electrical response feature component to obtain the correlation value, specifically includes: obtaining the data length of the electrical response feature component and the data length of the attitude sequence angular velocity component; segmenting the second feature vector according to the data length of the electrical response feature component and the data length of the attitude sequence angular velocity component to separate the attitude feature sequence representing the spatial motion state of the muffler and the electrical feature sequence representing the electrical response of the wiring harness connector; calculating the direction cosine of the attitude sequence angular velocity component relative to a preset reference coordinate axis in the attitude feature sequence, and extracting the weight coefficient corresponding to the electrical response feature component in the electrical feature sequence; multiplying the direction cosine and the weight coefficient element-wise and summing them to obtain the inner product, which is used as the correlation value; If the correlation value is greater than a preset threshold, it is determined that there is a strong coupling relationship between the electrical response feature component and the attitude sequence angular velocity component, and an intrinsic relationship matrix is ​​generated using the correlation value as matrix elements.

2. The method according to claim 1, characterized in that, The step of filtering the change in the electrode spacing and the deformation coefficient of the wire harness connector material to obtain a cumulative offset sequence includes: Obtain the current sampling frequency and adjust the length of the smoothing filter window based on the current sampling frequency; The Gaussian filter kernel is determined based on the length of the smoothing filter window, and the Gaussian filter kernel is used to smooth the change in the electrode spacing and the deformation coefficient of the wire harness connector material to obtain the filtered deformation coefficient sequence and the filtered electrode spacing change sequence. The cumulative offset sequence is obtained by numerically integrating the filtered deformation coefficient sequence and the filtered electrode spacing change sequence, respectively.

3. The method according to claim 2, characterized in that, Before using the Gaussian filter kernel to smooth the change in electrode spacing and the deformation coefficient of the wire harness connector material, the method further includes: Using the sampling time of the deformation coefficient of the wire harness connector material as a common time reference, the change in the electrode spacing is interpolated; The change in electrode spacing after interpolation is correlated point by point with the deformation coefficient of the wire harness connector material to establish a timestamp alignment relationship.

4. The method according to claim 1, characterized in that, The step of extracting the peak position and energy concentration interval of the frequency domain features, calculating the ratio of the time domain interval of the peak position to the span of the energy concentration interval, and generating a first feature vector includes: The frequency point corresponding to the maximum amplitude within the main frequency bandwidth is searched as the peak position, and the time domain interval between adjacent peak positions is calculated. Calculate the cumulative energy distribution sequence of the spectrum distribution, extract the frequency range in the cumulative energy distribution sequence that has an energy proportion that reaches a preset ratio as the energy concentration range, and calculate the difference between the upper and lower limits of the energy concentration range to obtain the span value. The span ratio is obtained by dividing the time interval by the span value, and the span ratio is arranged according to the frequency distribution to generate the first feature vector.

5. The method according to claim 1, characterized in that, The step of concatenating and normalizing the electrical response feature components of the first feature vector with the angular velocity components of the attitude sequence corresponding to the first feature vector to obtain the second feature vector includes: Extract the three-axis angular velocity data of the muffler within the test time window as the angular velocity components of the attitude sequence; The maximum and minimum values ​​of the angular velocity components of the attitude sequence within the test time window are obtained. The minimum value is subtracted from the angular velocity components of the attitude sequence, and then the result is divided by the difference between the maximum and minimum values ​​to obtain the normalized result. The electrical response feature components are arranged with the normalization result to obtain a concatenated feature vector; The concatenated feature vector is mapped to a preset numerical range to obtain the second feature vector.

6. The method according to claim 1, characterized in that, The step of generating an intrinsic relationship matrix using the correlation values ​​as matrix elements includes: Using the spatial coordinates of the electrical response feature components as row indices and the timestamps of the attitude sequence angular velocity components as column indices, the correlation values ​​are filled into the corresponding matrix positions to generate the intrinsic relationship matrix.

7. A motorcycle muffler electrical performance testing system, characterized in that, The motorcycle muffler electrical performance testing system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the motorcycle muffler electrical performance testing system to perform the method as described in any one of claims 1-6.

Images