High-speed railway vehicle dynamic response detection data mileage correction method and device

Abstract

This invention discloses a method and apparatus for correcting mileage in high-speed rail vehicle dynamic response detection data. The method includes: acquiring a track geometry detection data sequence and a vehicle dynamic response detection data sequence; extracting multiple track geometry detection data sub-sequences; extracting a candidate set corresponding to each track geometry detection data sub-sequence; determining the vehicle dynamic response detection data sub-sequence with the smallest cumulative distance to the track geometry detection data sub-sequence using an optimized dynamic time warping algorithm; the optimized dynamic time warping algorithm calculates the cumulative distance as follows: aligning sampling points in two sub-sequences according to a specified priority, and terminating the calculation early when the cumulative distance exceeds a preset threshold; correcting the original mileage in the vehicle dynamic response detection data sequence based on the mileage offset between the aforementioned sub-sequences. This invention can correct the mileage system deviation between high-speed rail vehicle dynamic response data and the spatial location of the track.

Description

Technical Field

[0001] This invention relates to the field of rail transit testing technology, and in particular to a method and apparatus for correcting mileage of dynamic response testing data for high-speed trains. Background Technology

[0002] The operational safety of high-speed railways heavily relies on the accurate detection and assessment of track conditions. Currently, vehicle dynamic response-based detection technology has become an important means of obtaining track condition information. However, the acquisition of vehicle dynamic response detection data is based on a time system, while the actual location of the track is based on a spatial mileage system. There is an inherent systemic bias between the two, making it difficult for the time series of the detection data to accurately correspond to the actual spatial mileage of the track.

[0003] In the field of rail transit inspection, the correction of mileage data mainly focuses on correcting the mileage deviation of the track geometry inspection data itself. Common methods include using fixed feature points of the line for reference, or using algorithms such as dynamic time warping to perform sequence matching and alignment of track geometry data.

[0004] While existing solutions can improve the mileage consistency of geometric data itself, their technical approach is limited to matching within the geometric data itself or between static and dynamic geometric data. They have not yet systematically integrated and correlated vehicle dynamic response data with track geometric data from multiple sources. Dynamic Time Warping (DTW) is widely used for track geometric data alignment because it can handle nonlinear scaling sequence matching problems. However, traditional DTW has high computational complexity, making it difficult to meet the processing requirements of long-distance, high-frequency sampled vehicle dynamic response data. Summary of the Invention

[0005] This invention provides a method for correcting mileage in dynamic response detection data of high-speed train vehicles, used to correct the mileage system deviation between the dynamic response data of high-speed train vehicles and the spatial location of the line. The method includes: Acquire track geometry detection data sequences and vehicle dynamic response detection data sequences; Extract multiple subsequences of orbit geometry detection data from the orbit geometry detection data sequence; Based on the mileage information of the track geometry detection data subsequence, a candidate set corresponding to each track geometry detection data subsequence is extracted from the vehicle dynamic response detection data sequence; the candidate set includes: multiple vehicle dynamic response detection data subsequences; An optimized dynamic time warping algorithm is used to determine the vehicle dynamic response detection data subsequence with the smallest cumulative distance to the corresponding track geometry detection data subsequence from each candidate set. The optimized dynamic time warping algorithm calculates the cumulative distance as follows: based on the data characteristics of the track geometry detection data subsequence, the priority of each sampling point in the track geometry detection data subsequence is determined; according to the priority, the sampling points in the track geometry detection data subsequence are sequentially aligned with the sampling points in the vehicle dynamic response detection data subsequence, and the cumulative distance of the aligned sampling points is calculated. The cumulative distance of the aligned sampling points is determined as the cumulative distance between the track geometry detection data subsequence and the vehicle dynamic response detection data subsequence; during the alignment and calculation process, if the cumulative distance is greater than a preset threshold, the calculation of the cumulative distance is terminated. Based on the mileage offset between each track geometry detection data subsequence and the corresponding vehicle dynamic response detection data subsequence with the smallest cumulative distance, the original mileage in the vehicle dynamic response detection data sequence is corrected to obtain the mileage-corrected vehicle dynamic response detection data sequence.

[0006] This invention also provides a mileage correction device for dynamic response detection data of high-speed train vehicles, used to correct the mileage system deviation between the dynamic response data of high-speed train vehicles and the spatial position of the line. The device includes: The data acquisition module is used to acquire track geometry detection data sequences and vehicle dynamic response detection data sequences. The track geometry detection data segmentation module is used to: extract multiple track geometry detection data subsequences from the track geometry detection data sequence; The vehicle dynamic response detection data segmentation module is used to: extract a candidate set corresponding to each track geometry detection data subsequence from the vehicle dynamic response detection data sequence based on the mileage information of the track geometry detection data subsequence; the candidate set includes: multiple vehicle dynamic response detection data subsequences; The data alignment module is used to: determine the vehicle dynamic response detection data subsequence with the smallest cumulative distance to the corresponding track geometry detection data subsequence from each candidate set using an optimized dynamic time warping algorithm; the optimized dynamic time warping algorithm calculates the cumulative distance as follows: based on the data characteristics of the track geometry detection data subsequence, determine the priority of each sampling point in the track geometry detection data subsequence; according to the priority, align the sampling points in the track geometry detection data subsequence with the sampling points in the vehicle dynamic response detection data subsequence in turn, calculate the cumulative distance of the aligned sampling points, and determine the cumulative distance of the aligned sampling points as the cumulative distance between the track geometry detection data subsequence and the vehicle dynamic response detection data subsequence; during the alignment and calculation process, if the cumulative distance is greater than a preset threshold, the calculation of the cumulative distance is terminated; The mileage correction module is used to correct the original mileage in the vehicle dynamic response detection data sequence based on the mileage offset between each track geometry detection data subsequence and the corresponding vehicle dynamic response detection data subsequence with the smallest cumulative distance, so as to obtain the mileage-corrected vehicle dynamic response detection data sequence.

[0007] This invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-mentioned method for correcting the mileage of dynamic response detection data of high-speed train vehicles.

[0008] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for correcting mileage of dynamic response detection data for high-speed train vehicles.

[0009] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described method for correcting mileage of dynamic response detection data for high-speed train vehicles.

[0010] Compared with existing technologies that are limited to track geometry data mileage correction, this invention establishes a correlation between vehicle dynamic response detection data and track geometry detection data to achieve mapping and matching between vehicle dynamic response and track spatial features. On this basis, a sampling point reordering and dynamic early termination mechanism is introduced into the traditional DTW to improve computational efficiency while ensuring accuracy, and to achieve mileage correction of dynamic detection data sampled based on the time system, providing technical support for intelligent track operation and maintenance. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a flowchart of the method for correcting mileage of dynamic response detection data of high-speed train vehicles in an embodiment of the present invention; Figure 2 This is a schematic diagram of the downsampling process of the time sampling system in an embodiment of the present invention; Figure 3 This is an example diagram showing the lateral acceleration and trend of the vehicle body in an embodiment of the present invention; Figure 4This is a comparison diagram of the time-sampled vehicle acceleration trend term and the space-sampled track geometry trend term in an embodiment of the present invention; Figure 5 This is an example diagram illustrating the alignment of data points between the vehicle acceleration trend term and the track geometry trend term in an embodiment of the present invention; Figure 6 This is an example diagram of the minimum cost path in an embodiment of the present invention; Figure 7 This is another flowchart for vehicle dynamic response data mileage correction in an embodiment of the present invention; Figure 8 This is a flowchart of the REA-DTW algorithm in an embodiment of the present invention; Figure 9 This is an example diagram comparing the curve segment before and after correction in an embodiment of the present invention; Figure 10 This is an example diagram illustrating the mileage correction effect of the multi-cycle correction detection task in the curve segment according to an embodiment of the present invention; Figure 11 This is a schematic diagram of the mileage correction device for dynamic response detection data of high-speed train vehicles in an embodiment of the present invention; Figure 12 This is another schematic diagram of the high-speed rail vehicle dynamic response detection data mileage correction device in an embodiment of the present invention; Figure 13 This is another schematic diagram of the high-speed rail vehicle dynamic response detection data mileage correction device in an embodiment of the present invention. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0013] Mileage correction of vehicle dynamic response data is a key step in improving the spatial positioning accuracy of dynamic detection data. The sources of mileage deviation mainly include the mapping deviation between the time sampling system and the spatial sampling system, that is, the systematic deviation of the correspondence between mileage and time, as well as the mileage deviation between the track geometry data itself and the track ledger.

[0014] Given that track geometry mileage correction technology is relatively mature, this embodiment of the invention focuses on targeted correction of the mileage offset accumulated during the vehicle dynamic response acquisition process, aiming to solve the alignment problem between the vehicle dynamic time series and the track geometry spatial series.

[0015] Figure 1 This is a flowchart of the mileage correction method for dynamic response detection data of high-speed train vehicles in an embodiment of the present invention. Figure 1As shown, this method can be implemented in the following steps: Step 101: Obtain the track geometry detection data sequence and the vehicle dynamic response detection data sequence; Step 102: Extract multiple subsequences of orbital geometry detection data from the orbital geometry detection data sequence; Step 103: Based on the mileage information of the track geometry detection data subsequence, extract the candidate set corresponding to each track geometry detection data subsequence from the vehicle dynamic response detection data sequence; the candidate set includes: multiple vehicle dynamic response detection data subsequences; Step 104: Using an optimized dynamic time warping algorithm, determine the vehicle dynamic response detection data subsequence with the smallest cumulative distance to the corresponding track geometry detection data subsequence from each candidate set; the optimized dynamic time warping algorithm calculates the cumulative distance as follows: based on the data characteristics of the track geometry detection data subsequence, determine the priority of each sampling point in the track geometry detection data subsequence; according to the priority, align the sampling points in the track geometry detection data subsequence with the sampling points in the vehicle dynamic response detection data subsequence in turn, calculate the cumulative distance of the aligned sampling points, and determine the cumulative distance of the aligned sampling points as the cumulative distance between the track geometry detection data subsequence and the vehicle dynamic response detection data subsequence; during the alignment and calculation process, if the cumulative distance is greater than a preset threshold, the calculation of the cumulative distance is terminated; Step 105: Based on the mileage offset between each track geometry detection data subsequence and the corresponding vehicle dynamic response detection data subsequence with the smallest cumulative distance, correct the original mileage in the vehicle dynamic response detection data sequence to obtain the mileage-corrected vehicle dynamic response detection data sequence.

[0016] Compared with existing technologies that are limited to track geometry data mileage correction, this invention establishes a correlation between vehicle dynamic response detection data and track geometry detection data to achieve mapping and matching between vehicle dynamic response and track spatial features. On this basis, a sampling point reordering and dynamic early termination mechanism is introduced into the traditional DTW to improve computational efficiency while ensuring accuracy, and to achieve mileage correction of dynamic detection data sampled based on the time system, providing technical support for intelligent track operation and maintenance.

[0017] In this embodiment of the invention, track geometry detection data sequence and vehicle dynamic response detection data sequence are obtained.

[0018] Because vehicle dynamic response detection data and track geometry data have fundamentally different sampling mechanisms, directly performing data fusion and comparative analysis can lead to problems of temporal misalignment and scale distortion.

[0019] To address this issue, in one embodiment, a trend term is extracted from the vehicle dynamic response detection data sequence, and the trend term is downsampled to obtain a vehicle dynamic trend sequence; a trend term is extracted from the track geometry detection data sequence to obtain a track geometry reference sequence; multiple track geometry detection data sub-sequences are extracted from the track geometry reference sequence; and a candidate set corresponding to each track geometry detection data sub-sequence is extracted from the vehicle dynamic trend sequence based on the mileage information of the track geometry detection data sub-sequences.

[0020] By placing the vehicle dynamic trend sequence and the track geometric reference sequence on the same analytical scale, a precise mapping and logical association between the vehicle dynamic response and the track spatial geometric state can be achieved, based on eliminating high-frequency vibration interference and differences in sampling mechanisms.

[0021] Figure 2 This is a schematic diagram of the downsampling process of the time sampling system in an embodiment of the present invention.

[0022] Taking the lateral acceleration of the vehicle body as an example, the vehicle dynamic response detection data is as follows: Figure 2 As shown, the vehicle's lateral acceleration signal is sampled using time sampling at a high frequency of 5000 points per second, while the track geometry data is sampled at equal intervals with a density of 0.25m. Downsampling processing achieved through interpolation makes the vehicle's lateral acceleration data scale closer to the track geometry detection data sequence, laying the foundation for subsequent sequence comparison.

[0023] In one embodiment, the vehicle dynamic response detection data sequence is subjected to low-frequency filtering to extract the trend term of the vehicle dynamic response detection data sequence, and the trend term is downsampled to obtain the vehicle dynamic trend sequence; the track geometry detection data sequence is subjected to low-frequency filtering to extract the trend term of the track geometry detection data sequence to obtain the track geometry reference sequence.

[0024] Figure 3 This is an example diagram of the lateral acceleration and trend of the vehicle body in an embodiment of the present invention. Figure 3 The blue signal represents the lateral acceleration detection data of the vehicle body, and the orange curve represents the trend term obtained through low-frequency filtering. The trend term of the lateral acceleration effectively characterizes the intensity of the systematic lateral sway generated when the vehicle passes through the curved section. Based on this, establishing a correlation mapping between the vehicle's dynamic response and the track geometry provides a common feature comparison benchmark for subsequent mileage alignment.

[0025] Figure 4 This is a comparison diagram of the time-sampled vehicle acceleration trend term and the spatial-sampled track geometric trend term in an embodiment of the present invention. Figure 4 The blue curve represents the trend term of the track geometry detection data sequence, and the orange curve represents the trend term of the vehicle dynamic response detection data sequence. Figure 4 It can be seen that there is a significant relative time delay between the two trend term sequences, which causes a systematic deviation in the dynamic detection data for actual mileage positioning.

[0026] After extracting and scaling the trend sequence from the two data sources, preliminary alignment at the macro-trend level was achieved. However, each track geometry subsequence corresponds to a fixed spatial span in the actual line, while the vehicle dynamic response is sampled in the time domain, and its data density is directly related to the vehicle's operating speed.

[0027] Figure 5 This is an example diagram illustrating the alignment of data points between the vehicle acceleration trend term and the track geometry trend term in an embodiment of the present invention. Figure 5 As shown, the red and green data points represent two sequences of different lengths.

[0028] Therefore, after downsampling, the time-sampled vehicle dynamic response detection data sequence still differs in density from the space-sampled track geometry detection data sequence, and when the two are aligned, they may form a one-to-many or many-to-one non-matching relationship.

[0029] For this type of asymmetric alignment problem, the DTW algorithm can be used.

[0030] For example, suppose the orbital geometric sequence is... The vehicle dynamic response sequence is .in, M This represents the length of the orbital geometric sequence, corresponding to the number of spatial sampling points; N This represents the length of the vehicle dynamic response sequence, corresponding to the number of time sampling points after downsampling.

[0031] Dynamically regularized paths are denoted as ;in, , For time index number, The spatial sampling sequence number is used, and the regularized path needs to satisfy boundary, continuity, and monotonicity constraints.

[0032] The goal of the DTW algorithm is to find the regularized path that minimizes the cumulative cost (i.e., the cumulative distance). ;in, Synchronization point The local distance between corresponding points is represented by Euclidean distance in this example. The minimum cumulative distance is calculated recursively using dynamic programming. ;in, Time sampling point i Spatial Mileage Sampling Points j The cumulative distance.

[0033] The final DTW algorithm output is The regularized path obtained is the optimal flexible correspondence between vehicle response and track geometry in time and space, which can provide a reliable alignment basis for subsequent mileage deviation correction.

[0034] Figure 6 This is an example diagram of the minimum cost path in an embodiment of the present invention. Figure 6 As shown, the horizontal and vertical axes represent the indices of the two sequences, and the red broken line in the figure represents the output of the REA-DTW algorithm.

[0035] Although the DTW algorithm provides a theoretical feasibility for achieving flexible alignment between vehicle response and track geometry in time and space, the inherent computational complexity of the DTW algorithm is proportional to the square of the sequence length. Faced with the massive vehicle dynamic response data generated by long-distance, high-frequency sampling of high-speed railways, traditional DTW is difficult to meet the real-time requirements of processing efficiency for actual detection tasks.

[0036] To address this, this invention proposes an optimized Dynamic Time Warping (REA-DTW) algorithm, transforming the traditional global full computation into a progressive computation based on an optimized path with the ability to exit early. By optimizing the order, strong discriminative information is obtained early to quickly distinguish whether two sequences match. Furthermore, by using a threshold determination, obviously mismatched candidates are eliminated early in the computation process, significantly reducing unnecessary computation and achieving a substantial improvement in computational efficiency while maintaining DTW alignment accuracy.

[0037] First, in this embodiment of the invention, multiple orbital geometry detection data subsequences are extracted from the orbital geometry detection data sequence.

[0038] In this embodiment of the invention, based on the mileage information of the track geometry detection data subsequence, a candidate set corresponding to each track geometry detection data subsequence is extracted from the vehicle dynamic response detection data sequence; the candidate set includes: multiple vehicle dynamic response detection data subsequences.

[0039] In one embodiment, within a preset range centered on the mileage information of the track geometry detection data subsequence in the vehicle dynamic response detection data sequence, multiple vehicle dynamic response detection data subsequences are extracted by a sliding window to obtain a candidate set corresponding to each track geometry detection data subsequence.

[0040] Based on the dynamic time warping framework, by introducing flexible alignment and local similarity measurement mechanisms, it is possible to achieve accurate matching of mileage offset between vehicle dynamic response sequences and track geometric reference sequences.

[0041] The waveform characteristics of straight sections of a track (such as the trend of lateral acceleration of the vehicle body) are often gentle, similar, and lack variation. In contrast, each curved section (with a specific radius, superelevation, and transition curve length) exhibits stable characteristics in both vehicle dynamic response (especially lateral acceleration) and track geometry data. Furthermore, the curve shape itself is a fixed design attribute of the track, remaining stable over the long term. Therefore, data sequence matching can be performed based on curve characteristics, which is not easily affected by daily changes in track conditions or detection noise, resulting in extremely high robustness in matching.

[0042] Figure 7 This is another flowchart for vehicle dynamic response data mileage correction in an embodiment of the present invention. Figure 7 Taking the lateral acceleration of the vehicle body in the dynamic response as an example, the vehicle body dynamic response data is first downsampled, and curve recognition is performed based on the track geometry data. Using the REA-DTW algorithm proposed in this embodiment of the invention, the cumulative distance between each curve segment and the track geometry detection data subsequence is found through iterative loops. Meets the criteria (less than the preset value) The vehicle dynamic response detection data subsequence effectively captures the sequence morphology and phase differences caused by mileage deviation.

[0043] In this embodiment of the invention, an optimized dynamic time warping algorithm is used to determine the vehicle dynamic response detection data subsequence with the smallest cumulative distance to the corresponding track geometry detection data subsequence from each candidate set. The optimized dynamic time warping algorithm calculates the cumulative distance as follows: based on the data characteristics of the track geometry detection data subsequence, the priority of each sampling point in the track geometry detection data subsequence is determined; according to the priority, the sampling points in the track geometry detection data subsequence are sequentially aligned with the sampling points in the vehicle dynamic response detection data subsequence, and the cumulative distance of the aligned sampling points is calculated. The cumulative distance of the aligned sampling points is determined as the cumulative distance between the track geometry detection data subsequence and the vehicle dynamic response detection data subsequence; during the alignment and calculation process, if the cumulative distance is greater than a preset threshold, the calculation of the cumulative distance is terminated.

[0044] Figure 8 This is a flowchart of the REA-DTW algorithm in an embodiment of the present invention.

[0045] For example, such as Figure 8As shown, the orbital geometry detection data subsequence is used as the query set, and each sampling point of the orbital geometry detection data subsequence is used as multiple features of the query set. The initial optimal distance for each candidate set is calculated and set as a preset threshold. Based on the orbital geometry features reflected in the data, an optimization order is set for the features of the query set, and the query set is traversed according to the optimization order. The cumulative distance (DTW distance) between the query set and each candidate object in the candidate set is calculated. If the cumulative distance is greater than the current optimal distance during the calculation process (in the first round of processing, the initial optimal distance is the current optimal distance; if the current cumulative distance is less than the initial optimal distance, the preset threshold is updated to the current cumulative distance), the candidate object being analyzed is abandoned early. Finally, the REA-DTW algorithm outputs the candidate object with the smallest cumulative distance.

[0046] exist Figure 8 Before the core computation loop of the REA-DTW algorithm starts, the algorithm first performs a crucial preprocessing step: dynamically reordering the feature computation order. Traditional DTW processes compute features in their natural order, such as "1, 2, 3, 4…", while the REA-DTW algorithm's optimization strategy intelligently reorders features based on their discriminative capabilities, generating new computation orders such as "…5, 3, 1, 2, 4…", thus constructing a more efficient optimized computation path. The purpose of this reordering mechanism is to prioritize features most likely to accumulate a distance exceeding a preset threshold in the early stages of computation, thereby achieving rapid identification and elimination of non-matching candidates.

[0047] The reordering and early termination mechanisms significantly improve computational efficiency. When the calculated maximum normalized distance (cumulative distance) exceeds a preset threshold, a significant mileage deviation is identified in the corresponding segment. Subsequently, within the identified curve segment, specific deviation points can be located, and the original mileage associated with the vehicle dynamic response data can be locally adjusted based on these deviation points.

[0048] In one embodiment, the absolute values ​​of the orbital geometric parameters at each sampling point in the orbital geometric detection data subsequence are obtained; the priority of each sampling point in the orbital geometric detection data subsequence is determined according to the descending order of the absolute values; wherein, the larger the absolute value, the higher the priority.

[0049] In one embodiment, the orbital geometry parameters include curvature and superelevation.

[0050] By introducing the REA-DTW algorithm, it is possible to efficiently and accurately correct systematic mileage deviations in vehicle dynamic response data, thereby significantly improving the reliability and application value of this data in accurate infrastructure condition assessment and fault location.

[0051] In this embodiment of the invention, the original mileage in the vehicle dynamic response detection data sequence is corrected based on the mileage offset between each track geometry detection data subsequence and the corresponding vehicle dynamic response detection data subsequence with the smallest cumulative distance, so as to obtain the mileage-corrected vehicle dynamic response detection data sequence.

[0052] In one embodiment, the vehicle dynamic response detection data sequence after mileage correction is upsampled using an interpolation method.

[0053] For example, the process of accurately matching the mileage deviation between the vehicle dynamic response detection data subsequence and the track geometry detection data subsequence is as follows: 1. Record the input vehicle dynamic response lateral acceleration as... ,in For the first i Each time sampling point (i.e., the sampling point in the vehicle dynamic response detection data sequence, hereinafter referred to as time sampling point). The number of time sampling points; the lateral acceleration of the track geometry vehicle body is denoted as... ,in For the first j Spatial mileage sampling points (i.e., sampling points in the orbital geometry detection data sequence, hereinafter referred to as spatial mileage sampling points). This represents the number of spatial mileage sampling points.

[0054] 2. Regarding Perform low-frequency filtering to extract the trend term. ;in, Indicates the cutoff frequency as The low-frequency filter. Similarly, we obtain... Trend items .

[0055] 3. For trend items Perform downsampling to make its length equal to Approximate, denoted as: ;in, The vehicle dynamic trend sequence obtained by downsampling processing, For trend items The first in k One sampling point, K For trend items The number of sampling points in the sample.

[0056] 4. Based on the orbital geometric trend term Identify the curve and extract the corresponding subsequence of track geometry detection data. Denote the subsequence of track geometry detection data as the curve. ;in, For the spatial mileage sampling point at the starting end of the m-th curve, Let M be the spatial mileage sampling point at the m-th curve termination point, where M is the number of curves; Based on the mileage information corresponding to the spatial mileage sampling points, it is possible to The vehicle dynamic response detection data subsequence is extracted from the corresponding curve segment and denoted as . ;in, The time sampling point at the starting end of the m-th curve. Let M be the time sampling point at the termination point of the m-th curve, and M be the number of curves; and This reflects the initial mileage deviation.

[0057] 5. Taking the m-th curve segment as an example, the track geometry detection data subsequence of the m-th curve segment is used as the query set. The vehicle dynamic response detection data subsequence of the m-th curve segment is used as the candidate set. .

[0058] 6. In this example, the orbital geometry detection data subsequence is expanded before and after. L Within a 500m range, for A sliding window selection process is used to obtain the candidate set. The REA-DTW algorithm is then used to calculate the relationship between each subsequence in the candidate set and the query set. cumulative distance .

[0059] 7. Determine the query set The subsequence with the smallest cumulative distance from the candidate set Sampling point offset between ;in, This represents the number of offset sampling points corresponding to the initial alignment position. The offset is determined based on the number of sampling points. The mileage of the vehicle dynamic response detection data sequence in the m-th curve segment is evaluated and corrected. ;in, This is the corrected vehicle dynamic response mileage sequence. This is the original vehicle dynamic response mileage sequence.

[0060] 8. The corrected vehicle dynamic response mileage sequence is upsampled to the original time resolution using an interpolation method to obtain the complete vehicle dynamic response data after mileage correction. .

[0061] According to the curve segment, the above steps 6 to 8 are executed iteratively segment by segment to achieve the mileage system correction of the entire line.

[0062] Figure 9This is an example diagram comparing the curve segment before and after correction in an embodiment of the present invention. Based on the optimal matching point pair relationship, the vehicle dynamic response data sequence is shifted along the time axis accordingly to obtain the corrected alignment result, as shown below. Figure 9 The lower curve is shown. It is misaligned with the original. Figure 9 Compared to the upper curve, the DTW distance between the two sequences decreased from 2.356 to 0.431 after this correction, indicating that the vehicle response data and track geometry data achieved a better match in waveform structure, verifying the effectiveness of this alignment method in reducing the mileage difference between sequences.

[0063] Figure 10 This is an example diagram illustrating the mileage correction effect of a multi-cycle correction and detection task in a curve section according to an embodiment of the present invention. Mileage correction is performed on the vehicle acceleration data collected in multiple repeated detection tasks, and the results after correction for each cycle are compared, as shown below. Figure 10 As shown, the blue curve represents the data collected in April, the orange curve represents the data collected in May, and the green curve represents the data collected in June. Figure 10 The curves from each cycle showed good consistency at key feature points. This indicates that the mileage correction method not only has high single-process accuracy but also maintains stable correction effects at different detection times, demonstrating good generalization performance.

[0064] Furthermore, the execution efficiency of the high-speed rail vehicle dynamic response detection data mileage correction method in this embodiment of the invention was quantitatively analyzed, focusing on its correction speed performance when processing data segments of different lengths. The experiment selected dynamic detection data from typical line segments of varying mileage lengths as the test set. Comparative experiments with the traditional dynamic time warping method on the same hardware device verified the computational efficiency advantage of this model.

[0065] The experimental results are shown in Table 1. Compared with the traditional dynamic time warping algorithm, under the same hardware environment, the average processing time of this model for unit mileage data is reduced by more than 90%, and the efficiency improvement is better as the data scale increases. It can meet the processing needs of large-scale line data in actual engineering.

[0066] Table 1 Comparison of REA-DTW mileage correction efficiency for different mileage lengths

[0067] This invention also provides a mileage correction device for dynamic response detection data of high-speed train vehicles, as described in the following embodiments. Since the principle by which this device solves the problem is similar to the method for correcting mileage in dynamic response detection data of high-speed train vehicles, the implementation of this device can refer to the implementation of the method for correcting mileage in dynamic response detection data of high-speed train vehicles; repeated details will not be elaborated further.

[0068] Figure 11 This is a schematic diagram of the mileage correction device for dynamic response detection data of high-speed train vehicles in an embodiment of the present invention. Figure 11 As shown, the device includes: Data acquisition module 1101 is used to: acquire track geometry detection data sequence and vehicle dynamic response detection data sequence; The track geometry detection data segmentation module 1102 is used to: extract multiple track geometry detection data subsequences from the track geometry detection data sequence; The vehicle dynamic response detection data segmentation module 1103 is used to: extract a candidate set corresponding to each track geometry detection data subsequence from the vehicle dynamic response detection data sequence based on the mileage information of the track geometry detection data subsequence; the candidate set includes: multiple vehicle dynamic response detection data subsequences; The data alignment module 1104 is used to: determine the vehicle dynamic response detection data subsequence with the smallest cumulative distance to the corresponding track geometry detection data subsequence from each candidate set using an optimized dynamic time warping algorithm; the optimized dynamic time warping algorithm calculates the cumulative distance as follows: based on the data characteristics of the track geometry detection data subsequence, the priority of each sampling point in the track geometry detection data subsequence is determined; according to the priority, the sampling points in the track geometry detection data subsequence are sequentially aligned with the sampling points in the vehicle dynamic response detection data subsequence, the cumulative distance of the aligned sampling points is calculated, and the cumulative distance of the aligned sampling points is determined as the cumulative distance between the track geometry detection data subsequence and the vehicle dynamic response detection data subsequence; during the alignment and calculation process, if the cumulative distance is greater than a preset threshold, the calculation of the cumulative distance is terminated; The mileage correction module 1105 is used to correct the original mileage in the vehicle dynamic response detection data sequence based on the mileage offset between each track geometry detection data subsequence and the corresponding vehicle dynamic response detection data subsequence with the smallest cumulative distance, so as to obtain the mileage-corrected vehicle dynamic response detection data sequence.

[0069] Figure 12 This is another schematic diagram of the high-speed rail vehicle dynamic response detection data mileage correction device in an embodiment of the present invention.

[0070] In one embodiment, such as Figure 12 As shown, Figure 11 The device also includes a data preprocessing module 1201, used for: The trend term of the vehicle dynamic response detection data sequence is extracted, and the trend term is downsampled to obtain the vehicle dynamic trend sequence. The trend term of the track geometry detection data sequence is extracted to obtain the track geometry reference sequence; The track geometry detection data segmentation module 1102 is specifically used to: extract multiple track geometry detection data subsequences from the track geometry reference sequence; The vehicle dynamic response detection data segmentation module 1103 is specifically used to: extract the candidate set corresponding to each track geometry detection data subsequence from the vehicle dynamic trend sequence based on the mileage information of the track geometry detection data subsequence.

[0071] In one embodiment, the data preprocessing module 1201 is specifically used for: The vehicle dynamic response detection data sequence is subjected to low-frequency filtering to extract the trend term of the vehicle dynamic response detection data sequence, and the trend term is downsampled to obtain the vehicle dynamic trend sequence. Low-frequency filtering is applied to the track geometry detection data sequence to extract the trend term, thus obtaining the track geometry reference sequence.

[0072] In one embodiment, the vehicle dynamic response detection data segmentation module 1103 is specifically used for: Within a preset range centered on the mileage information of the track geometry detection data subsequence in the vehicle dynamic response detection data sequence, multiple vehicle dynamic response detection data subsequences are extracted by a sliding window to obtain a candidate set corresponding to each track geometry detection data subsequence.

[0073] In one embodiment, the data alignment module 1104 is specifically used for: Obtain the absolute values ​​of the orbital geometric parameters at each sampling point in the orbital geometric detection data subsequence; The priority of each sampling point in the orbital geometry detection data subsequence is determined according to the descending order of the absolute value; wherein, the larger the absolute value, the higher the priority.

[0074] In one embodiment, the orbital geometry parameters include curvature and superelevation.

[0075] Figure 13 This is another schematic diagram of the high-speed rail vehicle dynamic response detection data mileage correction device in an embodiment of the present invention.

[0076] In one embodiment, such as Figure 13 As shown, Figure 11 The device also includes a data recovery module 1301, used for: The vehicle dynamic response detection data sequence after mileage correction is upsampled using an interpolation method.

[0077] This invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-mentioned method for correcting the mileage of dynamic response detection data of high-speed train vehicles.

[0078] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for correcting mileage of dynamic response detection data for high-speed train vehicles.

[0079] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described method for correcting mileage of dynamic response detection data for high-speed train vehicles.

[0080] Compared with existing technologies that are limited to track geometry data mileage correction, this invention establishes a correlation between vehicle dynamic response detection data and track geometry detection data to achieve mapping and matching between vehicle dynamic response and track spatial features. On this basis, a sampling point reordering and dynamic early termination mechanism is introduced into the traditional DTW to improve computational efficiency while ensuring accuracy, and to achieve mileage correction of dynamic detection data sampled based on the time system, providing technical support for intelligent track operation and maintenance.

[0081] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0082] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0083] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0084] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0085] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for correcting mileage in dynamic response detection data of high-speed train vehicles, characterized in that, include: Acquire track geometry detection data sequences and vehicle dynamic response detection data sequences; Extract multiple subsequences of orbital geometry detection data from the orbital geometry detection data sequence; Based on the mileage information of the track geometry detection data subsequence, a candidate set corresponding to each track geometry detection data subsequence is extracted from the vehicle dynamic response detection data sequence; the candidate set includes: multiple vehicle dynamic response detection data subsequences; An optimized dynamic time warping algorithm is used to determine the vehicle dynamic response detection data subsequence with the smallest cumulative distance to the corresponding track geometry detection data subsequence from each candidate set. The optimized dynamic time warping algorithm calculates the cumulative distance as follows: based on the data characteristics of the track geometry detection data subsequence, the priority of each sampling point in the track geometry detection data subsequence is determined; according to the priority, the sampling points in the track geometry detection data subsequence are sequentially aligned with the sampling points in the vehicle dynamic response detection data subsequence, and the cumulative distance of the aligned sampling points is calculated. The cumulative distance of the aligned sampling points is determined as the cumulative distance between the track geometry detection data subsequence and the vehicle dynamic response detection data subsequence; during the alignment and calculation process, if the cumulative distance is greater than a preset threshold, the calculation of the cumulative distance is terminated. Based on the mileage offset between each track geometry detection data subsequence and the corresponding vehicle dynamic response detection data subsequence with the smallest cumulative distance, the original mileage in the vehicle dynamic response detection data sequence is corrected to obtain the mileage-corrected vehicle dynamic response detection data sequence.

2. The method as described in claim 1, characterized in that, After acquiring the track geometry detection data sequence and the vehicle dynamic response detection data sequence, the following is also included: The trend term of the vehicle dynamic response detection data sequence is extracted, and the trend term is downsampled to obtain the vehicle dynamic trend sequence. The trend term of the track geometry detection data sequence is extracted to obtain the track geometry reference sequence; Extracting multiple subsequences of orbital geometry detection data from the orbital geometry detection data sequence, including: extracting multiple subsequences of orbital geometry detection data from the orbital geometry reference sequence; Based on the mileage information of the track geometry detection data subsequences, a candidate set corresponding to each track geometry detection data subsequence is extracted from the vehicle dynamic response detection data sequence, including: Based on the mileage information of the track geometry detection data subsequence, a candidate set corresponding to each track geometry detection data subsequence is extracted from the vehicle dynamic trend sequence.

3. The method as described in claim 2, characterized in that, Extract the trend term from the vehicle dynamic response detection data sequence, and downsample the trend term to obtain the vehicle dynamic trend sequence, including: The vehicle dynamic response detection data sequence is subjected to low-frequency filtering to extract the trend term of the vehicle dynamic response detection data sequence, and the trend term is downsampled to obtain the vehicle dynamic trend sequence. Extracting the trend term from the track geometry detection data sequence to obtain the track geometry reference sequence includes: performing low-frequency filtering on the track geometry detection data sequence, extracting the trend term from the track geometry detection data sequence, and obtaining the track geometry reference sequence.

4. The method as described in claim 1, characterized in that, Based on the mileage information of the track geometry detection data subsequences, a candidate set corresponding to each track geometry detection data subsequence is extracted from the vehicle dynamic response detection data sequence, including: Within a preset range centered on the mileage information of the track geometry detection data subsequence in the vehicle dynamic response detection data sequence, multiple vehicle dynamic response detection data subsequences are extracted by a sliding window to obtain a candidate set corresponding to each track geometry detection data subsequence.

5. The method as described in claim 1, characterized in that, Based on the data characteristics of the orbital geometry detection data subsequence, the priority of each sampling point in the orbital geometry detection data subsequence is determined, including: Obtain the absolute values ​​of the orbital geometric parameters at each sampling point in the orbital geometric detection data subsequence; The priority of each sampling point in the orbital geometry detection data subsequence is determined according to the descending order of the absolute value; wherein, the larger the absolute value, the higher the priority.

6. The method as described in claim 5, characterized in that, The track geometry parameters include curvature and superelevation.

7. The method as described in claim 1, characterized in that, After correcting the original mileage in the vehicle dynamic response detection data sequence based on the mileage offset between each track geometry detection data subsequence and the corresponding vehicle dynamic response detection data subsequence with the smallest cumulative distance, to obtain the mileage-corrected vehicle dynamic response detection data sequence, the following steps are also included: The vehicle dynamic response detection data sequence after mileage correction is upsampled using an interpolation method.

8. A mileage correction device for dynamic response detection data of high-speed train vehicles, characterized in that, include: The data acquisition module is used to acquire track geometry detection data sequences and vehicle dynamic response detection data sequences. The track geometry detection data segmentation module is used to: extract multiple track geometry detection data subsequences from the track geometry detection data sequence; The vehicle dynamic response detection data segmentation module is used to: extract a candidate set corresponding to each track geometry detection data subsequence from the vehicle dynamic response detection data sequence based on the mileage information of the track geometry detection data subsequence; the candidate set includes: multiple vehicle dynamic response detection data subsequences; The data alignment module is used to: determine the vehicle dynamic response detection data subsequence with the smallest cumulative distance to the corresponding track geometry detection data subsequence from each candidate set using an optimized dynamic time warping algorithm; the optimized dynamic time warping algorithm calculates the cumulative distance as follows: based on the data characteristics of the track geometry detection data subsequence, determine the priority of each sampling point in the track geometry detection data subsequence; according to the priority, align the sampling points in the track geometry detection data subsequence with the sampling points in the vehicle dynamic response detection data subsequence in turn, calculate the cumulative distance of the aligned sampling points, and determine the cumulative distance of the aligned sampling points as the cumulative distance between the track geometry detection data subsequence and the vehicle dynamic response detection data subsequence; during the alignment and calculation process, if the cumulative distance is greater than a preset threshold, the calculation of the cumulative distance is terminated; The mileage correction module is used to correct the original mileage in the vehicle dynamic response detection data sequence based on the mileage offset between each track geometry detection data subsequence and the corresponding vehicle dynamic response detection data subsequence with the smallest cumulative distance, so as to obtain the mileage-corrected vehicle dynamic response detection data sequence.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1 to 7.

11. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method of any one of claims 1 to 7.

Images