Method for correcting geometric irregularity mileage deviation of track

Abstract

The invention relates to the technical field of track geometric irregularity detection, and discloses a track geometric irregularity mileage deviation correction method which comprises the following steps: acquiring track geometric irregularity detection data; a principal component analysis method is adopted to extract principal component components representing the geometric irregularity data of the track; calculating a multivariate cross correlation coefficient of the geometric irregularity data sequence of the track detected at different times, and performing overall correction on the section mileage by using an offset when the multivariate cross correlation coefficient reaches a maximum value; and carrying out secondary mileage correction on the geometric irregularity of the track in the section based on a data sequence similarity measurement method of dynamic time warping. Unified mileage deviation correction of multiple parameters of track irregularity is achieved, it is ensured that mileage correction of different parameters is consistent, track detection data quality is improved, and accurate data support is provided for track maintenance and state analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of track geometric smoothness detection, and particularly relates to a method for correcting the mileage deviation of track geometric irregularities. Background Art

[0002] The track geometric smoothness detection data is expressed as a function of the mileage position. The mileage positioning of the high-speed comprehensive inspection train on the operating line mainly uses RFID radio frequency technology and GNSS space positioning technology for absolute mileage positioning, and at the same time integrates the mileage accumulation technology of the speed encoder to jointly determine the line mileage position. Due to factors such as relative slip between the wheel and rail, wheel rolling radius, and absolute mileage positioning accuracy, there are problems such as uneven distribution, error accumulation, duplicate identification, and missing of the mileage in the track detection data. The mileage deviation directly affects the accurate positioning and analysis of track diseases, increases the difficulty of on-site operation and maintenance work, and even affects the safety and stability of railway operation. At the same time, the mileage deviation also causes uncertainty in deeply exploring the spatio-temporal evolution characteristics and laws of the track geometric smoothness state. Therefore, accurately correcting the mileage of the moving inspection vehicle data has become an urgent need to improve the quality of the moving inspection vehicle data and ensure the scientific and efficient development of railway maintenance work, and is also of great significance for accurately evaluating the track smoothness and the reliability of in-depth analysis of the long-term evolution law of the track.

[0003] The absolute mileage deviation of the track detection data is the difference relative to the true mileage information of the line. In terms of hardware, mainly a multi-source fusion spatial positioning method such as high-precision differential GNSS, inertial navigation, RFID, and speed encoder is used to achieve higher-precision positioning of the mileage position of the high-speed inspection train; in terms of data processing, mainly based on the true mileage of the line plane curve characteristics and equipment information such as switches, bridges, and ATDs as a reference object to process the mileage deviation of the track detection data. The relative mileage deviation between different-time detection data mainly uses mathematical algorithms such as the least squares method, correlation coefficient, cross-correlation function, and dynamic time warping to determine the mileage difference in different sections and correct it. In addition to the relative mileage correction between different-period dynamic detection data, the literature proposes using static detection data with high accuracy and strong reliability as a reference benchmark, and using the waveform matching of dynamic and static measured data to evaluate and correct the mileage error of the dynamic detection data.

[0004] For the absolute mileage deviation, mainly based on obvious parameters such as superelevation and curvature in the line geometric parameters for matching to obtain the mileage deviation value. For the relative mileage deviation, mainly based on the data of a single track geometric irregularity to calculate its deviation, and there is a problem that the mileage deviation values are not equal when performing similarity matching on different track geometric parameters. Summary of the Invention

[0005] In view of the above deficiencies in the prior art, the present invention provides a method for correcting the mileage deviation of track geometric irregularities.

[0006] To achieve the above object of the invention, the technical solution adopted by the present invention is as follows: A method for correcting the mileage deviation of track geometric irregularities, comprising the following steps: Obtain track geometric irregularity detection data; Use the principal component analysis method to extract the principal component components characterizing the track geometric irregularity data; Calculate the multivariate cross-correlation coefficient for the track geometric irregularity data sequences detected at different times, and perform overall mileage correction for the section mileage with the offset when the multivariate cross-correlation coefficient reaches the maximum value; Perform secondary mileage correction for the track geometric irregularities within the section based on the dynamic time warping-based data sequence similarity measurement method.

[0007] Further, calculating the multivariate cross-correlation coefficient for the track geometric irregularity data sequences detected at different times includes: Use the minimum covariance determinant method to calculate the robust estimates of the mean and covariance of the track geometric irregularity data detected at different times; Calculate the Mahalanobis distance of the track geometric irregularity data based on the robust estimates of the mean and covariance of the track geometric irregularity data; Divide the track geometric irregularity data into multiple non-overlapping sections, and calculate the cross-covariance and variance of two multivariate data sequences in each section; Calculate the multivariate data sequence cross-correlation coefficient based on the cross-covariance and variance of the two multivariate data sequences.

[0008] Further, when performing overall mileage correction for the section mileage with the offset when the multivariate cross-correlation coefficient reaches the maximum value, determine the mileage deviation of each section with the offset when the multivariate data sequence cross-correlation coefficient reaches the maximum value, and perform overall mileage correction according to the mileage deviation of each section.

[0009] Further, determining the mileage deviation of each section with the offset when the multivariate data sequence cross-correlation coefficient reaches the maximum value specifically includes:

[0010] wherein, is the mileage deviation of the l -th k section of track geometric irregularities with a section length of is the input value that makes the function reach the maximum value, is the cross-correlation coefficient with a multivariate time series offset of m , is the l -thk The maximum correlation coefficient of two data sequences of segments is the maximum value function m and is the offset of two multivariate data sequences.

[0011] Furthermore, mileage correction is performed according to the mileage deviation of each segment, specifically:

[0012] Among them, is the mileage of the track geometric irregularity of the l th k segment after correction with the corrected segment length is the original mileage of the track geometric irregularity of the l th k segment with the segment length is the mileage deviation of the segment ( l , k ).

[0013] Furthermore, a method for measuring the similarity of track geometric irregularity mileage within a segment based on dynamic time warping includes: Constructing a distance matrix of two multivariate data sequences according to the distance between the two multivariate data sequences; Constructing a cumulative distance matrix of two multivariate data sequences according to the distance matrix of the two multivariate data sequences; Starting from the lower right corner of the cumulative distance matrix and tracing back to the upper left corner, at each step, select the point with the smallest cumulative distance among the adjacent upper left, left, or upper directions until returning to the starting point to obtain the optimal alignment path and the minimum cumulative distance; Stretching or compressing the track geometric irregularity mileage data of each segment according to the optimal alignment path and the minimum cumulative distance.

[0014] Furthermore, the calculation method of the distance between two multivariate data sequences is:

[0015] Among them, is the distance between the y 1 th element in the multivariate data sequence i and the y 2 th element in the multivariate data sequence j , is the y 1 th element in the multivariate data sequence i containing the p th multivariate variable is the y2 The j th element contains the p th multivariate variable, P is the number of multivariate variables, w is the label for distance calculation.

[0016] Furthermore, the distance matrix of two multivariate data sequences is specifically:

[0017] where is the distance matrix between a multivariate data sequence with data length M and a multivariate data sequence with data length N.

[0018] Furthermore, the elements in the cumulative distance matrix of two multivariate data sequences are specifically:

[0019] where is the cumulative distance between the y 1 th element in i and the y 2 th element in j , and min is the minimum value function.

[0020] The present invention has the following beneficial effects: The present invention first segments the track irregularity data detected multiple times, and selects the principal component components according to the variance of the principal components by using the principal component analysis method. Using the principal component components as input parameters, the overall mileage deviation value of the track detection data in each section is calculated by using the multivariate correlation function, so that the multi-parameters of track irregularity have consistent mileage deviation. Since the mileage deviations of each section are not exactly equal, the multivariate dynamic time warping method is used to perform secondary correction on the mileage within each section to achieve the purpose of accurately aligning the mileage of different track inspection data. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 is a schematic flow chart of a method for correcting the mileage deviation of track geometric irregularities. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] The following describes the specific embodiments of the present invention to facilitate those skilled in the art of the present technology to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those of ordinary skill in the art of the present technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions created using the concept of the present invention are within the scope of protection.

[0023] In order to solve the problem of inconsistent mileage deviation when correcting mileage for different track geometric parameters, overcome the defect that the existing mileage deviation correction method cannot handle the overall mileage deviation of the section and the mileage stretching and compression within the section at the same time, eliminate the inaccurate positioning of track diseases caused by mileage deviation, and the interference with track condition analysis and maintenance work, the present invention proposes a method for correcting mileage deviation of track geometric detection data. By extracting independent principal component components to characterize the track geometric information features, combined with the multivariate cross-correlation coefficient and multivariate dynamic time warping technology based on Mahalanobis distance, the unified mileage deviation correction of multiple parameters of track unevenness is realized, ensuring consistent mileage correction for different parameters, improving the quality of track detection data, and providing accurate data support for track maintenance and condition analysis.

[0024] The present invention first segments the track unevenness data detected multiple times, and selects the principal component components according to the variance of the principal components by using the principal component analysis method. Taking the principal component components as input parameters, the overall mileage deviation values of the track detection data in each section are calculated by using the multivariate correlation function, so that the multiple parameters of track unevenness have consistent mileage deviation. Since the mileage deviations of each section are not exactly equal, the multivariate dynamic time warping method is used to perform secondary correction on the mileage of each section to achieve the purpose of accurately aligning the mileage of different track inspection data.

[0025] As Figure 1 shown, a method for correcting mileage deviation of track geometric unevenness provided by an embodiment of the present invention includes the following steps S1 to S4: S1. Obtain track geometric unevenness detection data; S2. Use the principal component analysis method to extract the principal component components characterizing the track geometric unevenness data; In an optional embodiment of the present invention, the track geometric smoothness state is characterized by seven unevenness parameters, and the parameters are not completely independent. Independent principal component components are extracted by principal component analysis to more accurately analyze the track geometric state characteristics. For the track unevenness data set X, which measures N data and has p dimensional original feature vectors, which can be expressed as:

[0026] The principal component analysis method linearly transforms the above original vector X into the same number of principal component components Y, which is expressed as follows:

[0027] In the formula, y p is the N th p dimensional principal component vector with

[0028] The calculation steps of the covariance matrix eigenvalues in this embodiment are as follows: (1) Data preprocessing Since principal component analysis is sensitive to the scale of data, each original eigenvector x j (1 ≤ j ≤ p ) is centered and standardized:

[0029] In the formula, and are the mean and standard deviation of the -th eigenvector respectively. After standardization, the dimensional differences of each eigenvector can be removed, and the evaluation indexes can be unified.

[0030] (2) Eigenvalue decomposition Calculate the covariance matrix between each standardized eigenvector according to the formula . From the eigenvalue equation , non-negative eigenvalues can be obtained, and the corresponding eigenvectors are the coefficients W of the original vector X transformed into the principal component component Y.

[0031] (3) Selection of principal component components Sort the eigenvalues of the principal components from large to small, and select the principal component components according to their cumulative contribution rate of variance. The cumulative contribution rate of the first r principal components is:

[0032] If the cumulative contribution rate exceeds a certain threshold and the characteristic information represented by these r principal components meets the requirements, the characteristic information of the original vector can be represented by fewer new principal component components.

[0033] After the track irregularity data is analyzed by principal component analysis, uncorrelated principal component components are formed, where is the number of observed data, and is the number of principal component components.

[0034] This embodiment extracts the main characteristic components of the track geometric irregularity data through principal component analysis, reduces the data dimension, and improves the calculation efficiency.

[0035] S3. Calculate the multiple cross-correlation coefficient for the track geometric irregularity data sequences detected at different times, and perform overall correction of the section mileage with the offset when the multiple cross-correlation coefficient reaches the maximum value; In an alternative embodiment of the present invention, step S3 calculates the multivariate cross-correlation coefficient for the track geometric irregularity data sequences detected at different times, including: Calculating robust estimates of the mean and covariance of the track geometric irregularity data detected at different times using the minimum covariance determinant method; Calculating the Mahalanobis distance of the track geometric irregularity data based on the robust estimates of the mean and covariance of the track geometric irregularity data; Dividing the track geometric irregularity data into multiple non-overlapping segments, and calculating the cross-covariance and variance of two multivariate time series in each segment; Calculating the multivariate cross-correlation coefficient based on the cross-covariance and variance of the two multivariate data sequences.

[0036] For the track irregularity data sequences detected at different times in this embodiment and , the calculation process of the multivariate cross-correlation coefficient is as follows: (1) Mahalanobis distance of the principal component components Calculating the robust estimates of the mean and covariance of the principal component components based on the MCD (Minimum covariance determinant) method , . Then the Mahalanobis distance of each data in the principal component components can be expressed as:

[0037] (2) Data segmentation Dividing the principal component component Y with a length of N into non-overlapping segments, then the data sequence of each segment is: ,

[0038] where is the segment number, l is the segment length.

[0039] (3) Multivariate cross-correlation coefficient of the data sequence The cross-covariance and variance of the two multivariate data sequences in each segment can be expressed as:

[0040]

[0041] where m is the relative translation amount of the two multivariate data sequences.

[0042] The cross - correlation coefficient of two multi - variable data sequences is defined as follows:

[0043] When step S3 performs overall correction of the section mileage with the offset at which the multi - variable cross - correlation coefficient reaches the maximum value, the mileage deviation of each section is determined by the offset at which the multi - variable data sequence cross - correlation coefficient reaches the maximum value, and the overall mileage of each section is corrected according to the mileage deviation of each section.

[0044] In this embodiment, when the cross - correlation coefficient of the multi - variable data of the two track irregularity data reaches the maximum value, its offset m is the mileage deviation Δ l , k of this section ( S , that is:

[0045] According to the mileage deviation of each section, the original mileage of each section is corrected according to the following formula, that is

[0046] In the formula, is the mileage of the l th segment of track irregularity with a segment length of k , and its corresponding track irregularity data is , is the mileage distribution after mileage deviation correction.

[0047] In this embodiment, based on the principal component components, the multi - variable cross - correlation coefficient is calculated to determine the overall mileage deviation of each section.

[0048] S4. Perform secondary mileage correction of the track geometric irregularity within the section based on the data sequence similarity measurement method of dynamic time warping.

[0049] In an alternative embodiment of the present invention, step S4 performs secondary mileage correction of the track geometric irregularity within the section based on the data sequence similarity measurement method of dynamic time warping, including: Construct a distance matrix of two multi - variable data sequences according to the distance between the two multi - variable data sequences; Construct a cumulative distance matrix of two multi - variable data sequences according to the distance matrix of the two multi - variable data sequences; Starting from the lower - right corner of the cumulative distance matrix and backtracking to the upper - left corner, select the point with the minimum cumulative distance in the adjacent upper - left, left, or upper direction at each step until returning to the starting point to obtain the optimal alignment path and the minimum cumulative distance; Stretch or compress the track geometric irregularity mileage data of each section according to the optimal alignment path and the minimum cumulative distance.

[0050] Since the mileage deviations of each section along the line direction are not exactly equal, mileage compression or stretching occurs in the section overlapping area, resulting in uneven mileage distribution. To address this issue, the dynamic time warping method is adopted in this embodiment to perform secondary correction on the mileage of each section.

[0051] Dynamic time warping is a time series similarity measurement method based on the idea of dynamic programming. It supports amplitude translation, time axis stretching and bending, can handle the similarity measurement between non-equal length time series, and has good accuracy and robustness. The calculation process of dynamic time warping is as follows: (1) Construct a distance matrix For two P dimensional data sequences and , construct the distance matrix D of the two data sequences M,N as:

[0052] In the formula, and are the data lengths, is the number of multivariate variables.

[0053] Each element in the matrix represents the distance between the 1 th element in Y i and the 2 th element in Y j , that is:

[0054] Among them, is the distance between the y 1 th element in the multivariate data sequence i and the y 2 th element in the multivariate data sequence j , is the y 1 th element in the multivariate data sequence i contains the p th multivariate variable, is the y 2 th element in the multivariate data sequence j contains the p th multivariate variable, P is the number of multivariate variables, w is the distance calculation label, w When w = 2, it is the Euclidean distance.

[0055] (2)Calculate the cumulative distance matrix using dynamic programming Cumulative distance matrix R The elements of the

[0056] (3)Calculate the minimum cumulative distance and backtrack the path Starting from the lower right corner of the cumulative distance matrix r M,N Backtrack to the upper left corner r 1,1 to find the path with the minimum cumulative distance, which represents the optimal alignment between two data sequences. The final minimum cumulative distance is the value in the lower right corner of the cumulative distance matrix r M,N , which represents the minimum cumulative distance between two data sequences and is used to measure their similarity. By performing dynamic time warping calculations on the different sub-track irregularity data of each section, more reasonable stretching and compression of the waveform of the track detection data are achieved to achieve a more accurate alignment of the mileage.

[0057] In this embodiment, the dynamic time warping method is used to perform secondary correction on the mileage within the section to solve the problem of mileage stretching or compression and achieve high-precision mileage alignment.

[0058] In summary, the present invention has the following advantages compared with the prior art: Consistency: By performing principal component analysis and calculating the multiple correlation coefficient, the problem of inconsistent mileage deviation correction results for different track geometric parameters is solved.

[0059] High precision: Combining the dynamic time warping method, the problem of mileage stretching or compression within the section is solved, significantly improving the matching accuracy of the track detection data.

[0060] Applicability: This method is not only applicable to the mileage deviation correction between dynamic detection data, but also can be applied to the correction between dynamic and static detection data.

[0061] Efficiency: Through dimensionality reduction processing by principal component analysis, the computational complexity is reduced and the correction efficiency is improved.

[0062] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and combinations of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices produce means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in one block or multiple blocks.

[0063] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in one block or multiple blocks.

[0064] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in one block or multiple blocks.

[0065] Specific embodiments are applied in the present invention to elaborate on the principles and implementation manners of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

[0066] Those of ordinary skill in the art will realize that the embodiments described herein are for helping readers understand the principles of the present invention, and it should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various other specific deformations and combinations that do not depart from the essence of the present invention based on the technical revelations disclosed in the present invention, and these deformations and combinations are still within the protection scope of the present invention.

Claims

1. A method for correcting mileage deviation of track geometric irregularity, characterized in that: The following steps are involved: Obtain track geometry irregularity detection data; The principal component analysis method is used to extract the principal component components that characterize the track geometric irregularity data; The multivariate correlation coefficients of the track geometric irregularity data sequences detected at different times are calculated, and the overall section mileage correction is performed based on the offset when the multivariate correlation coefficient reaches the maximum value. A data sequence similarity measurement method based on dynamic time warping is used to perform secondary mileage correction for track geometric irregularities within a section.

2. A method for correcting mileage deviation of track geometric irregularity according to claim 1, characterized in that: The multivariate correlation coefficients of the track geometric irregularity data sequences detected at different times are calculated, including: The minimum covariance determinant method is used to calculate the robust estimation of the mean and covariance of track geometry irregularity data detected at different times; The Mahalanobis distance of the track geometry irregularity data is calculated based on the robust estimation of the mean and covariance of the track geometry irregularity data; The track geometric irregularity data are divided into multiple non-overlapping segments, and the cross-covariance and variance of the two multivariate data sequences in each segment are calculated; Computes the cross-correlation coefficient of two multivariate data series based on their cross-covariance and variance.

3. A method for correcting mileage deviation of track geometric irregularity according to claim 1, characterized in that: When the overall mileage of the section is corrected by the offset when the multivariate cross-correlation coefficient reaches the maximum value, the mileage deviation of each section is determined by the offset when the multivariate time series cross-correlation coefficient reaches the maximum value, and the overall mileage of each section is corrected according to the mileage deviation of each section.

4. A method for correcting mileage deviation of track geometric irregularity according to claim 3, characterized in that: The mileage deviation of each section is determined by the offset when the correlation coefficient of the multivariate data sequence reaches the maximum value, specifically: in, The segment length is l No. k Mileage deviation of track geometry irregularity, In order to maximize the input value of the function, The offset for the multivariate time series is m The mutual correlation coefficient of The segment length is l No. k The maximum correlation coefficient of two data sequences in the segment, is the maximum value function, m is the offset between two multivariate data sequences.

5. A method for correcting mileage deviation of track geometric irregularity according to claim 4, characterized in that: Mileage correction is performed based on the mileage deviation of each section, specifically: in, The corrected segment length is l No. k Mileage with uneven track geometry, The segment length is l No. k The original mileage of the track with uneven geometry, For the segment ( l , k ) mileage deviation.

6. A method for correcting mileage deviation of track geometric irregularity according to claim 1, characterized in that: The data sequence similarity measurement method based on dynamic time warping is used to perform secondary mileage correction of track geometric irregularities within the section, including: Construct a distance matrix of two multivariate data sequences based on the distances of the two multivariate data sequences; Construct a cumulative distance matrix of two multivariate data sequences based on their distance matrices; Start from the lower right corner of the cumulative distance matrix and trace back to the upper left corner. At each step, select the point with the smallest cumulative distance in the adjacent upper left, left or upper direction until you return to the starting point to obtain the optimal alignment path and the minimum cumulative distance. The track geometric irregularity mileage data of each section is stretched or compressed according to the optimal alignment path and the minimum cumulative distance.

7. A method for correcting mileage deviation of track geometric irregularity according to claim 6, characterized in that: The distance between two multivariate data sequences is calculated as: in, For multivariate time series y 1 Middle i Elements and Multivariate Time Series y 2 Middle j The distance between elements, For multivariate time series y 1 Middle i The element contains the p multivariate variables, For multivariate time series y 2 Middle j The element contains the p multivariate variables, P is the number of multivariate variables, w Label the distance calculation.

8. A method for correcting mileage deviation of track geometric irregularity according to claim 7, characterized in that: The distance matrix of two multivariate data sequences is specifically: in, It is the distance matrix between the multivariate data sequence with data length M and the multivariate data sequence with data length N.

9. A method for correcting mileage deviation of track geometric irregularity according to claim 8, characterized in that: The elements in the cumulative distance matrix of two multivariate data sequences are specifically: in, For multivariate data series y 1 Middle i Elements and multivariate data sequences y 2 Middle j The cumulative distance between elements, min is the minimum function.

Images