CN112949696A - Track static inspection historical data matching method based on DTW and robust estimation - Google Patents
Track static inspection historical data matching method based on DTW and robust estimation Download PDFInfo
- Publication number
- CN112949696A CN112949696A CN202110167669.XA CN202110167669A CN112949696A CN 112949696 A CN112949696 A CN 112949696A CN 202110167669 A CN202110167669 A CN 202110167669A CN 112949696 A CN112949696 A CN 112949696A
- Authority
- CN
- China
- Prior art keywords
- matching
- mileage
- robust estimation
- dtw
- calculating
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2474—Sequence data queries, e.g. querying versioned data
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Fuzzy Systems (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a track static inspection historical data matching method based on DTW and robust estimation, which comprises the following steps: selecting [ t ]0,tend]The current static inspection data of the mileage range is used as a query sequence Q, and the same mileage t0,tend]The homonymous past static inspection data of the range is used as a matching sequence C; calculating a Q regular path P and a C regular path P by using a DTW algorithm, calculating a mileage offset delta by using robust estimation on the regular path P according to [ t ]0‑δ,tend‑δ]Reselecting a query sequence Q; calculating the Q and C regular paths P by using the DTW algorithm againbestNamely track static check historical data matching relation. The invention utilizes dynamic time warping elasticity to measure the similarity of static inspection historical data and carries out robust estimation on the mileage offset according to the linear relation similar to the matching path, thereby realizing the existence of noise, asynchronism, offset, drift and mileage extension and contraction in the static inspection dataThe matching precision is high, the matching speed is high, and the accuracy of track static inspection data analysis can be improved.
Description
Technical Field
The invention relates to the field of track detection, in particular to a track static inspection historical data matching method based on DTW and robust estimation.
Background
The smooth state evaluation and the maintenance are based on the track line inspection. China implements a line inspection principle of 'dynamic inspection as a main part and combination of dynamic inspection and static inspection'. However, the static inspection data of the same mileage lacks of definite matching relation in amplitude and mileage, so that the difficulty of track state evaluation, site identification of diseases and return inspection of operation quality is caused. Furthermore, for example, the status correction and the preventive correction of the line equipment also depend on the matching of the data.
Matching is to identify and align the data level of the content or structure with the same/similar attributes in different data sets from the same or similar scenes or objects. The matching method generally includes a region-based method, a feature-based method, a point set matching method, and the like. Regarding multiple dynamic examination records of the rail inspection vehicle on the same sampling point, the mileage matching can be realized by adopting a characteristic-based method such as DGPS (differential global positioning system), RFID (radio frequency identification) and the like of absolute position data, or the matching relation between the dynamic examination data is determined by adopting a region-based method such as a correlation function, gray correlation and the like of relative position data. The acquisition of absolute positions requires hardware overhead; classical point set matching methods such as euclidean distance are only suitable for "one-to-one" comparisons and are sensitive to shifts in time series, amplitude variations, and the like. Related technicians realize mileage correction of multiple Dynamic inspection records by using the correlation analysis and Dynamic Time Warping (DTW) of the track gauge data. The complexity of the algorithm of dynamic time warping increases exponentially with the expansion of the matching range; the calculation amount can be reduced by adopting the correlation function to determine the search space, but the correlation function is suitable for measuring linear similarity, and the track gauge data have nonlinearity due to slippage/slip of the mileage wheel and sampling interval errors.
Disclosure of Invention
In view of the above, it is desirable to find a method for accurately and rapidly matching historical mileage data under the conditions of noise, asynchrony, drift and scaling of static inspection data.
A historical data matching method based on DTW and robust estimation comprises the following steps:
selecting [ t ]0,tend]The current static inspection data of the mileage range is used as a query sequence Q, and the same mileage t0,tend]The same of the rangeNamed prior static inspection data as a matching sequence C;
calculating Q and C regular paths P by using a dynamic time warping algorithm (DTW), calculating mileage offset delta by using robust estimation according to the regular paths P and calculating the mileage offset delta according to t0-δ,tend-δ]Reselecting a query sequence Q;
calculating the Q and C matching paths P again by using the dynamic time warping algorithm (DTW)bestNamely track static check historical data matching relation.
Further, the above track static inspection historical data matching method based on dynamic time warping and robust estimation, wherein the dynamic time warping algorithm step includes:
the distance matrix D between the two sequences is calculated:
d(i,j)＝||qi-cj||w
wherein i is the query sequence Q index, i is 1,2 …, n; j is the matching sequence C index, j is 1,2, …, m. When w is 1, manhattan distance; when w is 2, the Euclidean distance is obtained;
and D, searching a regular path P by adopting a dynamic programming method:
P＝{p1,p2,...,pk,...,pK}
wherein p iskIndicating the position of the regular path, i.e. pk＝(i,j)kDenotes qiAnd cjThe matching relationship between the two;
determination of regular path position p by dynamic programmingkWhen calculating DTW distance, a cost matrix is required to be constructed, and matrix elements gamma (i, j) are defined as
The dynamic time warping distance DTW (Q, C) for Q and C is calculated such that the cumulative distance value is minimized. The dynamic time warping distance is calculated as follows
Further, in the above track static inspection historical data matching method based on dynamic time warping and robust estimation, the warping path P should have a linear relationship between the data index i of Q and the data index j of C in the warping path P according to the static inspection sampling interval of 0.125m
j＝αi+Δ
Wherein, alpha is the slope, and alpha is approximately equal to 1.0; delta is an intercept, and alpha and delta are calculated by adopting a steady estimation; the mileage offset δ is Δ × 0.25 m.
Further, according to the track static inspection historical data matching method based on dynamic time warping and robust estimation, the robust estimation is adopted to calculate the mileage offset δ, in order to improve the estimation accuracy, the maximum allowable error of the offset δ can be set as a cycle termination condition, and multiple times of robust estimation are obtained.
Further, according to the track static inspection historical data matching method based on dynamic time warping and robust estimation, the track static inspection historical data is collected through a track inspection instrument.
The method realizes the track static inspection historical data matching based on the dynamic time warping and the robust estimation. Compared with the prior art, the method can realize accurate and quick matching of multiple static inspection data of the same mileage under the conditions of noise, asynchronism, deviation, drift and extension of the static inspection data. The method has high matching precision and high matching speed, and can be used for improving the accuracy of analysis of the track static inspection data.
Drawings
FIG. 1 is a flowchart of a track static inspection history data matching method based on dynamic time warping and robust estimation according to a first embodiment of the present invention;
FIG. 2 is a flowchart of a track static inspection history data matching method based on dynamic time warping and robust estimation according to a second embodiment of the present invention;
FIG. 3 is a diagram illustrating data samples of a query sequence Q and a matching sequence C according to a second embodiment of the present invention, wherein FIG. 3a) is the query sequence Q, and FIG. 3b) is the matching sequence C;
FIG. 4 is a schematic diagram illustrating a dynamic programming method used in the step D to find a regular path P according to a second embodiment of the present invention;
FIG. 5 shows a second embodiment of the Q and C matching paths PbestSchematic representation.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the accompanying drawings are illustrative only for the purpose of explaining the present invention, and are not to be construed as limiting the present invention.
These and other aspects of embodiments of the invention will be apparent with reference to the following description and attached drawings. In the description and drawings, particular embodiments of the invention have been disclosed in detail as being indicative of some of the ways in which the principles of the embodiments of the invention may be practiced, but it is understood that the scope of the embodiments of the invention is not limited correspondingly. On the contrary, the embodiments of the invention include all changes, modifications and equivalents coming within the spirit and terms of the claims appended hereto.
Referring to fig. 1, a track static inspection history data matching method based on dynamic time warping and robust estimation according to a first embodiment of the present invention is applied to track static inspection history data matching to determine the relationship between mileage and amplitude among history data. The matching method includes steps S11-S13.
Step S11, selecting mileage range [ t0,tend]And determining the query sequence Q and the matching sequence C.
Mileage range [ t ]0,tend]Can be set by comprehensively considering the track management requirements and the data matching efficiency, t0And tendRespectively a starting point mileage value and a terminal point mileage value. In particular, matched static checklistsThe mileage error range between history data is generally not more than 100 m. In order to ensure that the sequences are overlapped and the subsequent mileage deviation calculation is convenient, the difference t between the starting mileage and the end mileage of the query sequence and the matching sequenceend-t0Is more than or equal to 100m, such as taking tend-t01000 m. And the query sequence Q and the matching sequence C are collected by an orbit inspection instrument.
Step S12, calculating Q and C regular paths P by using DTW algorithm, calculating mileage offset delta by using robust estimation from P, and calculating mileage offset delta according to [ t [ [ t ]0-δ,tend-δ]The query sequence Q is reselected.
The measurement distance determines the similarity degree between data, and the similarity measurement mode determines the measurement effect. DTW is a dynamic programming-based measure of elasticity. Unlike the euclidean distance metric, the DTW calculates the similarity between two time sequences by normalizing the time sequences, so it has significant advantages in dealing with time sequence matching that has problems of non-equal length, offset, or stretching of amplitude and time axis.
The step of calculating the Q and C regular paths P by using the DTW algorithm comprises the following steps:
step S121: the distance D (i, j) between the query sequence Q and the matching sequence C and the distance matrix D will be calculated.
Wherein
d(i,j)＝||qi-cj||w
Wherein i is the index of the query sequence Q, i is 1,2 …, n; j is the matching sequence C index, j is 1,2, …, m. When w is 1, manhattan distance; when w is 2, it is an euclidean distance. Obtaining a distance matrix
Step 122: in D, a dynamic planning method is adopted to search for regular paths
P＝{p1,p2,...,pk,...,pK}
Wherein p iskIndicating the position of the regular path, i.e. pk＝(i,j)kDenotes qiAnd cjThe matching relationship between the two;
specifically, the regular path position p is determined by adopting the idea of dynamic programmingkWhen calculating DTW distance, a cost matrix is required to be constructed, and matrix elements gamma (i, j) are defined as
The dynamic time warping distance DTW (Q, C) for Q and C is calculated such that the cumulative distance value is minimized. The dynamic time warping distance is calculated as follows
The presence of mileage drift directly affects matching accuracy and indirectly affects matching efficiency. The step of eliminating the mileage offset includes S123-S124.
Step S123: computing mileage offset δ from P using robust estimation
Specifically, according to the static check sampling interval of 0.125m, the data index i of Q in the regular path P and the data index j of C should be in a linear relationship
j＝αi+Δ
Wherein, alpha is the slope, and alpha is approximately equal to 1.0; delta is an intercept, and alpha and delta are calculated by adopting a steady estimation; the mileage offset δ is Δ × 0.125 m.
Further, according to the track static inspection historical data matching method based on dynamic time warping and robust estimation, the robust estimation is adopted to calculate the mileage offset δ, in order to improve the estimation accuracy, the maximum allowable error of the offset δ can be set as a cycle termination condition, and multiple times of robust estimation are obtained.
Step S124: according to [ t ]0-δ,tend-δ]Reselecting a query sequence Q
Specifically, the original mileage range [ t ] is divided according to the mileage offset delta0,tend]Is corrected to [ t0-δ,tend-δ]And selects the corresponding query sequence Q.
In the step of S13,calculating the matching paths P of Q and C by using the DTW algorithm againbestNamely track static check historical data matching relation.
Specifically, the query sequence Q and the matching sequence C after mileage correction are used as input, and step S121 and step S122 are repeated, and the dynamic time warping distance is calculated and the dynamic warping path P is determinedbestI.e. is the matching path Pbest，
I.e. the matching relationship between index i and index j under the path.
The embodiment of the invention realizes the track static inspection historical data matching based on the dynamic time warping and the robust estimation. Compared with the prior art, the method can realize accurate and quick matching of multiple static inspection data of the same mileage under the conditions of noise, asynchronism, deviation, drift and extension of the static inspection data. The method has high matching precision and high matching speed, and can be used for improving the accuracy of analysis of the track static inspection data.
Referring to fig. 2, a track static inspection history data matching method based on dynamic time warping and robust estimation according to a second embodiment of the present invention is shown. The matching method includes steps S21-S23.
Step S21, selecting mileage range [ t0,tend]And determining the query sequence Q and the matching sequence C.
Specifically, as shown in FIG. 3, a mileage range [ t ] is selected0,tend]The left high-low data of the static inspection of the current orbit inspection tester is used as a query sequence Q (see figure 3a)), and the same mileage is t0,tend]Left high and low data of the previous static inspection of the range orbit inspector are taken as a matching sequence C (see FIG. 3 b)). Wherein, t0＝K1224.200km，tend＝K1225.200km。
Step S22, calculating Q and C regular paths P by using DTW algorithm, calculating mileage offset delta by using robust estimation from P, and calculating mileage offset delta according to [ t [ [ t ]0-δ,tend-δ]The query sequence Q is reselected.
Step S221: the distance D (i, j) between the query sequence Q and the matching sequence C and the distance matrix D will be calculated. Wherein d (i, j) is Euclidean distance
d(i,j)＝||qi-cj||2
Wherein i is the index of the query sequence Q, i is 1,2 …, n; j is the matching sequence C index, j is 1,2, …, m. Obtain a distance matrix D
Step S222: in D, a dynamic programming method is used to find the regular path P, as shown in fig. 4.
Step S223: computing mileage offset δ from P using robust estimation
The estimated Δ is 319, and the mileage offset δ is 39.9 m.
Step S224: according to [ t ]0-δ,tend-δ]Reselecting a query sequence Q
Specifically, according to the mileage offset delta, the original mileage range [ K1224.2, K1225.2] is corrected to [ K1224.1601, K1225.1601] and a corresponding query sequence Q is selected.
Step S23, calculating Q and C matching paths P again by using the DTW algorithmbestNamely track static check historical data matching relation.
Specifically, the query sequence Q and the matching sequence C after mileage correction are used as input, and the steps S221 to S222 are repeated, the dynamic time warping distance is calculated, and the dynamic warping path P is determinedbestAs can be seen in the figure 5,
i.e. the matching relationship between index i and index j under the path.
It will be appreciated that as an implementable approach, static checks of left and right rail direction, right elevation, and level, warp, and gauge may also be combined to construct the query sequence and the matching sequence to further mention detection accuracy.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The above-mentioned embodiments only express several embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (5)
1. A track static inspection historical data matching method based on DTW and robust estimation is characterized by comprising the following steps:
selecting [ t ]0,tend]The current static inspection data of the mileage range is used as a query sequence Q, and the same mileage t0,tend]The homonymous past static inspection data of the range is used as a matching sequence C;
calculating a Q warping path P and a C warping path P by using a dynamic time warping algorithm, calculating a mileage offset delta by using robust estimation according to the warping path P and calculating the mileage offset delta according to [ t [ [ t ]0-δ,tend-δ]Reselecting a query sequence Q;
calculating the matching paths P of Q and C by using the dynamic time warping algorithm againbestNamely track static check historical data matching relation.
2. The DTW and robust estimation based track static inspection historical data matching method according to claim 1, wherein said dynamic time warping algorithm step comprises:
calculating a distance matrix D between the two sequences;
and D, searching a regular path P by adopting a dynamic programming method.
3. The DTW and robust estimation based track static inspection history data matching method according to claim 1, wherein the data index i of Q and the data index j of C in the warping path P are in a linear relationship according to the static inspection sampling interval 0.125 m:
j＝αi+Δ
wherein α is a slope, Δ is an intercept, α and Δ are calculated by using a robust estimation, and the mileage offset δ is Δ × 0.125 m.
4. The method for matching historical data of static orbit inspection based on DTW and robust estimation according to claim 1, wherein in the step of calculating the mileage offset δ by using robust estimation, the maximum allowable error of the mileage offset δ is set as a cycle termination condition, and the maximum allowable error is obtained by multiple times of robust estimation.
5. The DTW and robust estimation based track static inspection history data matching method according to claim 1, wherein the track static inspection history data is collected by a track inspection instrument.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110167669.XA CN112949696A (en) | 2021-02-07 | 2021-02-07 | Track static inspection historical data matching method based on DTW and robust estimation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110167669.XA CN112949696A (en) | 2021-02-07 | 2021-02-07 | Track static inspection historical data matching method based on DTW and robust estimation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112949696A true CN112949696A (en) | 2021-06-11 |
Family
ID=76243170
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110167669.XA Withdrawn CN112949696A (en) | 2021-02-07 | 2021-02-07 | Track static inspection historical data matching method based on DTW and robust estimation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112949696A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114896803A (en) * | 2022-05-26 | 2022-08-12 | 江西科技学院 | Multi-parameter rail inspection data mileage positioning method |
| CN115292872A (en) * | 2022-05-30 | 2022-11-04 | 中国特种设备检测研究院 | Roller coaster track defect positioning method, system, medium and equipment |
| CN117312635A (en) * | 2023-11-30 | 2023-12-29 | 江西日月明测控科技股份有限公司 | On-line detection data analysis processing method, system, electronic equipment and storage medium |
-
2021
- 2021-02-07 CN CN202110167669.XA patent/CN112949696A/en not_active Withdrawn
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114896803A (en) * | 2022-05-26 | 2022-08-12 | 江西科技学院 | Multi-parameter rail inspection data mileage positioning method |
| CN115292872A (en) * | 2022-05-30 | 2022-11-04 | 中国特种设备检测研究院 | Roller coaster track defect positioning method, system, medium and equipment |
| CN117312635A (en) * | 2023-11-30 | 2023-12-29 | 江西日月明测控科技股份有限公司 | On-line detection data analysis processing method, system, electronic equipment and storage medium |
| CN117312635B (en) * | 2023-11-30 | 2024-02-02 | 江西日月明测控科技股份有限公司 | On-line detection data analysis processing method, system, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112949696A (en) | Track static inspection historical data matching method based on DTW and robust estimation | |
| CN112906782B (en) | Track static inspection historical data matching method based on DTW and least square estimation | |
| EP1886110B1 (en) | A method and a system for determining a plurality of load components on a wheel | |
| US4318617A (en) | DC Shift error correction for electro-optical measuring system | |
| CN109813269B (en) | On-line calibration data sequence matching method for structure monitoring sensor | |
| CN112883078B (en) | Track dynamic inspection historical data matching method based on DTW and least square estimation | |
| CN113032380B (en) | Method, system and equipment for aligning internal and external detection data of pipeline | |
| US11428590B2 (en) | Method for estimating a bearing load using strain parameters to account for contact angle variation | |
| CN103486971A (en) | Subway tunnel crack width detecting and correcting algorithm | |
| US7152015B2 (en) | Terminal with position-measuring functions | |
| CN112883079A (en) | Track dynamic inspection historical data matching method based on DTW and robust estimation | |
| CN117168337B (en) | OFDR strain edge optimization method and measurement method | |
| CN110672058B (en) | On-line calibration data sequence matching method and device of sensor for structure monitoring | |
| CN115754010B (en) | Quality control method for sealing end face of mechanical sealing ring | |
| CN116522085A (en) | Full-automatic inhaul cable frequency extraction, fixed-order and cable force identification method and application | |
| CN114896803B (en) | Multi-parameter rail inspection data mileage positioning method | |
| CN118444348B (en) | Track positioning data analysis and correction system and method | |
| CN117968992B (en) | Steel plate elasticity detection device for steel plate spring | |
| CN118177995B (en) | Intelligent calibration method for surgical laser display angle ruler | |
| CN117330604B (en) | Automatic temperature compensation method, device, computer equipment and storage medium | |
| CN115267331B (en) | Signal orientation positioning method and system based on multi-spectral analysis | |
| CN118734229B (en) | Abnormal identification method for motor operation data | |
| CN118882695A (en) | Track inspection data mileage deviation determining method based on DTW and Hough transformation | |
| CN116147833A (en) | Pressure sensor performance analysis method and system based on data mining | |
| CN113671435A (en) | Method and device for evaluating step response characteristics of direct current transformer |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20210611 |
|
| WW01 | Invention patent application withdrawn after publication |