JP2933832B2 - Track position deviation method of track deviation and vehicle up and down measurement data - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a measuring position collating method which makes it possible to collate measurement data of vertical and vertical sway of a vehicle with a traveling position on a track.

[0002]

2. Description of the Related Art In the field of track maintenance and management, track irregularities and irregularities of a track are regularly measured by a track inspection vehicle, but intervals of measurement are determined in accordance with the importance of the track. Since it is various, to supplement the regular inspection by the track inspection car, in order to maintain a good track condition, measure various data such as sway, wheel load, lateral pressure, etc.
It is being considered that this should be actively used, and in order to process measurement data (especially sway data) on the car, it is essential to accurately grasp the corresponding position on the track. is there.

[0003] On the other hand, measurement data on commercial vehicles is necessary not only for track maintenance and management but also for vehicle improvement and the like. However, analysis of data measured on vehicles is often required. In the case of (1), it is necessary to take into account the track conditions on the ground (track shape, track structure, track deviation, etc.), and thus it is also necessary to accurately grasp the corresponding position on the running track.

[0004]

However, regarding the on-track position verification of the on-vehicle measurement data as described above, accurate position verification can be expected by a method of manually inputting a kilometer along a track while traveling. As a result, there is a problem that it takes time to specify the data position. On the other hand, as a measure to input continuously and automatically while traveling about a kilometer,
There is a method of optically detecting the passage of about a kilometer. However, this method requires installation of new equipment on both the track side and the vehicle side, which is not advantageous in terms of cost. Therefore, there is a demand for the development of a new method capable of accurately performing position matching in terms of both accuracy and efficiency of data processing.

[0005] By the way, the position collation is basically based on the correspondence of kilometers, but in a track inspection vehicle, the diameter of the wheel for measuring the traveling distance, the pulse signal of the ATS installed along the track and the installation position thereof It is considered that the position correspondence with the kilometer as much as possible, such as recording of information, can be taken, so the track deviation measured by the track inspection vehicle is data that accurately corresponds to the kilometer, It is direct data that continuously represents the running position on the track. In addition, the orbital disorder generally progresses in an orbital manner, and its state gradually changes due to track maintenance. However, as a whole, the change is gradual, and the state once measured does not suddenly change thereafter.

Further, the present inventor (Yoshimura) has conducted a number of data analyzes on the dynamic causal relationship between the actual track deviation and the railcar sway. He discovered that there was a very strong linear dependence between the vertical motion and the vertical motion acceleration, and announced it earlier (Yoshimura, "Analysis of vertical motion of vehicles in response to ups and downs of track," RTRI report, Vol.1, No. .3, 1987.11).

[0007] Therefore, if the correspondence between the data of the irregularity of the track and the vertical swing acceleration data of the vehicle with respect to the kilometer can be clarified, the position of the measured data on the vehicle can be compared with the track, but the vertical swing acceleration of the vehicle can be verified. There are several problems with acquiring data.

As data used for position verification, digital data obtained by A / D-converting measurement data using a distance signal (pulse signal) from the speed generator as a sample pulse is used. And the interval is calculated from the wheel diameter values. Therefore, if the wheels have wear or the like, an error occurs in the calculation of the distance, and an error in the sampling interval causes a position matching error. For example, if the error of the sampling interval is 1%, a position collation error of as much as 30 m occurs when the track length is 3 km.

Also, if the wheel detecting the distance signal has a gliding or slip, the sampling of the A / D conversion may be omitted or duplicated, which also causes a position collation error. Further, the distance signal has a smaller waveform as the train speed decreases,
The signal may not be detected.

[0010] Here, the data of the irregularity of the track is managed by the diameter of the wheel for measuring the traveling distance as described above.
Since it is formed by being associated with the ground ATS pulse signal, the above-described problem of the position matching error and the like does not occur, and the data is sufficiently reliable. The vertical sway acceleration data is obtained simply by using a distance signal from a speed generator provided in a vehicle such as a track inspection vehicle that does not take care of the vehicle. It is desired that errors, sampling omissions and duplications due to wheel sliding / slip, and the relationship with the train speed can be solved in the actual data processing process of position verification.

[0011] Further, when acquiring the vertical swaying acceleration data in a commercial vehicle, it is difficult to know the measurement start kilometer accurately, so it is necessary to use a method that does not require the measurement start position to be specified. is there.

The present invention has been made based on the finding that there is a very strong linear dependence between the elevation of the track and the vertical motion acceleration of the vehicle. It is an object of the present invention to provide a method of verifying a track position and a vertical movement measurement data of a vehicle, which can perform verification with a measurement position on a track with high accuracy.

[0013]

Means for Solving the Problems To achieve the above-mentioned object, a method for verifying the position of a track and a vertical movement of a vehicle according to the present invention has the following configuration. That is, in the first invention, the measurement position collation method of the measurement data of the track deviation and the vertical motion of the vehicle is stored in a first memory for storing data of the deviation of the track height; Means for averaging; means for performing low-pass filtering processing on the averaged out-of-order data; second memory for storing up-and-down swing data; and trajectory of the sampling interval of up-and-down swing data stored in the second memory. Means for converting the data up and down of the sampling interval into sampling intervals; means for performing low-pass filtering on the up-down sway data converted at the sampling interval; Means for performing a correlation process with the sway data to detect a maximum value; and detecting a positional shift amount of the up and down sway data from both sampling data having the maximum value. Means for setting, in a table, the relationship between the data number of the sampling data of the up-and-down swing data corrected by the amount of positional deviation and the kilometer; Means for correcting the interval.

In the first embodiment of the present invention, when the measurement section of the track is long, etc., the track section is divided into a plurality of check sections, and a maximum is determined for each predetermined check section. Each means from the value detection to the table setting is performed; Further, track map data may be used in place of the data of the irregularity of the track.

[0015] In the second invention, the measurement position collation method of the track deviation and vehicle up-and-down movement measurement data according to the second invention includes: a first memory for storing data on the deviation of the track; and a height of the left and right tracks stored in the first memory. Means for averaging out-of-order data;
Means for performing a low-pass filtering process on the averaged deviation data; a second memory for storing the vertical fluctuation data; and a sampling interval of the vertical fluctuation data stored in the second memory. Means for converting the data into sampling intervals; means for performing low-pass filtering on the up-and-down oscillating data converted for sampling intervals; and, for each predetermined first matching section, high-low filtering performed for low-pass filtering in the first matching section. The maximum value is detected by performing a correlation process between the out-of-order data and the vertical sway data, the positional deviation amount of the vertical sway data is detected from both sampling data having the maximum value, and the data number of the sampling data of the vertical sway data is determined. Means for setting, in the first table, the relationship between the value corrected by the displacement and the distance in kilometer; the data number and the kilometer set in the first table Means for correcting the sampling interval of the up-and-down motion data based on the relationship between; and for each second collation section shorter than the first collation section, referencing the averaged deviation data in the second collation section and the first table. Using the vertical sway data obtained by correcting the sampling interval taken out of the second memory, analysis of the vertical sway response to elevation deviation and matching processing between the obtained frequency response characteristics and ideal frequency response characteristics are performed by using the orbit deviation data. Performing by changing one or both sampling data of the vertical fluctuation data back and forth, and detecting the positional relationship between the vertical fluctuation data and the vertical fluctuation data that gives the frequency response characteristic that best matches the ideal frequency response characteristic, Means for setting, in a second table, the relationship between the data number of the sampling data of the detected vertical sway data and the kilometer; It is characterized in further comprising a; the relationship between higher data number and kilometers set in the table means for correcting the sampling interval of the vertical motion data.

According to the second aspect of the present invention, at least one of the following methods is used to verify the track position and the vertical movement measurement data of the vehicle.
When the processing in one first collation section is completed, the processing in the second collation section is performed, and the processing is repeated. That is, real-time processing can be performed.

In the first and second aspects of the present invention, the reading of the data having the irregularity in the collation section in which the correlation processing is performed is performed by referring to a table which defines the relationship between the data number of the sampling data and the kilometer. be able to.

[0018]

A description will now be given of the operation of the measurement position collating method of the track deviation and vehicle vertical movement measurement data of the present invention configured as described above. As described above, there is a very strong linear dependency between the height of the track and the vertical motion of the vehicle. Therefore, the first invention focuses on the point that the waveforms of the low-frequency components of both data are similar, performs low-pass processing, performs correlation processing, detects the maximum value, and determines the position of the vertical sway data. A deviation is detected, and a table of data numbers and kilometers is corrected. Since the correction is based on data that is clearly up and down in correspondence with kilometers on the ground, in the process of this correction, errors in sampling intervals due to wheel wear, etc. The problem goes away. In relation to the train speed, it can be avoided by not performing the position matching process in the matching section where the speed is equal to or less than the predetermined speed. Further, according to the present invention, it is not necessary to specify the measurement start position on the vehicle.

Here, in the first invention, the matching section is, for example, a section of several km because it is necessary to include as many slope change points as possible on the track, so that the maximum value obtained by the correlation processing is sufficiently large. Although it is not sharp, it does not change much even if the positions of both data are slightly shifted. In other words, it cannot be said that sufficiently high-precision position collation was performed, but position collation data that can be used for a while is obtained.

According to the second invention, in order to enhance the accuracy of the position collation data obtained in the first invention, the effect of the positional shift appears most sensitively on the phase characteristic. Focusing on the fact that the wavelength greatly changes, the first
A second matching section corresponding to the subdivision of the matching section (first matching section) of the present invention is set, and a frequency response analysis and a matching process are performed for each of the second matching sections while changing sampling data back and forth. Then, the positional relationship between the orbital deviation and the vertical sway data indicating the best match is detected, and a position collation table is created.

According to the second aspect of the present invention, a considerably fine adjustment can be performed, so that a position matching data with a considerably high accuracy can be obtained.

[0022]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a principle configuration of a measurement position collation method of measurement data of track deviation and vertical movement of a vehicle according to the present invention. In FIG. 1, this measurement position collation method basically includes a first memory 1, a second memory 2, a third memory 3,
U4.

The first memory 1 stores data on the irregularity of the track according to the periodic inspection of the track inspection vehicle (for example, FIGS. 2 and 3). As described above, in the track inspection vehicle, the data corresponding to the level of kilometers on the ground is acquired sufficiently accurately to obtain data that is out of order.

Specifically, the LABOCS system developed by the present inventors (Yoshimura, Hosokawa) and the like uses a signal of an ATS grounding element installed along the track and information of its installation position to provide a highly reliable position. Checked (Yoshimura, Yoshida, Hosokawa, Kikuchi “Track maintenance database system:
Development of Micro LABOCS-II + "RTRI Research Report, Vo
l.6, No. 11, 1992.11).

The high and low data is obtained by sampling the distance signal from the speed generator.
In the OCS system, the irregular data is provided along with a table called a “km / data number comparison table” which is a position collation result. In other words, the first memory 1 stores “high and low level data” and “table (km / data number comparison table)”. In addition, as described above, in the case of the height deviation, the state does not suddenly change after the measurement, but the latest inspection data is used.

This "kilogram / data number comparison table" records, for example, as shown in FIG. 4, a kilometer of an ATS ground child and a corresponding data number (a number indicating a sampling order) every 2-3 km. The sampling data of the irregular data can be specified by the kilometer and the data number. Since this “km / data number comparison table” is very effective, in the present invention, the position comparison result of the target vertical motion measurement data is also expressed in the “km / data number comparison table”. I decided to earn. Therefore, in order to distinguish between the two, if the “km / data number comparison table” is referred to as KDT, the former is referred to as “high / low KD”.
T ", and the latter is referred to as" shaking KDT ".

The second memory 2 stores data on vertical sway acceleration and train speed measured by a high-speed commercial vehicle (for example, FIGS. 2 and 3). This is a commercially available portable oscillometer ("Bound Storage CDH-30" manufactured by Eikura Tsushin Co., Ltd.).
1)). This oscillometer has a built-in acceleration sensor, A / D-converts the vertical and horizontal sway on the vehicle using a distance signal supplied from the vehicle as a sample clock, and, after calibration, records it on a memory card together with the train speed Having. Therefore, in the present embodiment, the second memory 2 refers to this memory card, but the data used in the present invention is the data of the vertical swing acceleration and the train speed.

The third memory 3 has two types of “shaking KD”.
T "is set as described later.

The CPU 4 includes a program memory, a working memory, and the like, and executes a predetermined program according to the processing procedure shown in FIG. 5 (first step) and FIG. 6 (second step) to realize the position collation of the present invention. The grounds for the processing procedure of the first step shown in FIG. 5 are as follows.

The basis of the present invention is, as described above, the linear dependency existing between the height of the track and the vertical motion of the vehicle. This can be known by analyzing the frequency response of the vertical sway acceleration with respect to the height deviation. In this frequency response analysis, as shown in FIG. 7, an estimation is performed by using the average value of the deviation of the height of the left and right lines as an input and the vertical vibration acceleration of the vehicle as an output. The estimation result is shown in FIG.
2 shows a very good correspondence with the simple two-degree-of-freedom vibration model composed of the bogie and the vehicle body.

FIG. 9 shows the frequency response estimation of the vertical sway acceleration with respect to the vertical sway and the vertical sway analyzed using the data of the vertical sway and the vertical sway acceleration shown in FIG. 2. FIG. 10 shows the frequency response of the vertical sway and the vertical sway shown in FIG. 9 shows the frequency response estimation of the vertical sway acceleration with respect to height fluctuation analyzed using the acceleration data. As shown in FIGS. 9 and 10 (c), the value of the coherence function indicating the degree of linear dependence shows a high value (0.8 to 0.9) in a wide frequency band, and a strong linear dependence relationship between input and output data. It is shown to be present. The good coherence was attributable to the averaging operation performed in a high-low level.

The present invention is based on the premise that such linear dependence can always be expected between the vertical vibration of a railroad vehicle and the track deviation. First, the position of the vertical sway measurement data is determined based on the track deviation data. Detect the deviation. In general, a correlation method is effective as a method for detecting the position shift. This is a method of calculating the correlation function shown in Expression 1 and obtaining a delay (lag number) τ _{max at} which the value is maximum.

[0033]

(Equation 1)

As is well known, this correlation method can be applied only when the waveforms of both data are very similar. However, according to FIGS. The waveform of the vertical sway acceleration is completely different and cannot be applied as it is.

However, paying attention to the characteristics shown in FIGS. 9 and 10, certain characteristics are found in the phase characteristics among the response characteristics. In other words, for a track deviation in a relatively long wavelength band exceeding 100 m, the phase lag of the vertical motion acceleration waveform of the vehicle is almost equal to 180 degrees (π radians).
That's what it means. This is because the acceleration is obtained by differentiating the displacement twice and there is a phase difference of 180 degrees with respect to the displacement.
The waveform of the displacement indicates that it follows the waveform of the out-of-orbit with almost no phase delay.

Therefore, if a long wavelength component of 100 m or more is extracted from the data of the trajectory deviation and the vertical movement of the vehicle, those waveforms should form very similar waveforms to each other. The cross-correlation can be calculated and the displacement can be known. The position collation processing of the first step shown in FIG. 5 is configured based on the above considerations.

In FIG. 5, the processing procedure of the first step is divided into a pre-processing step (S1 to S8) and a displacement detection step (S10 to S14). Basically, in a certain track section, pre-processing is performed on all data in the section, and then position shift detection is performed using the pre-processed data. The extension of the section is determined so as to include as many gradient change points as possible every 300 to 500 m. In this embodiment, the distance is 5 km.

Therefore, the length of the track section to be verified is 5 km.
In many cases, the line section is divided into, for example, 5 km units, and position deviation detection is performed for each of the divided sections. This unit section having an extension of 5 km is referred to as a “first collation section”. In this case as well, the pre-processing and the positional deviation detection may be performed for each first collation section in principle, but in the present embodiment (FIG. 5), the pre-processing is performed for all the measurements requiring the position collation. Data (for example, if you want to perform position matching for the track from Tokyo to Osaka,
Measurement data from Tokyo to Osaka) at once,
Thereafter, the position shift detection is performed for each first collation section.

By the way, in the pre-processing for height deviation, the first
The left and right high and low deviation data are sequentially taken out from the memory 1 (S1), and a high and low deviation average processing for obtaining a half of the right and left deviation is performed (S2), and a digital filter having a low-pass characteristic (FIG. 11). A filtering process is performed (S3), and the data is temporarily stored (S4).

Further, in the pre-processing for the vertical swing, the second
The vertical sway data is sequentially taken out from the memory 2 (S
5) Since the sampling interval between the data of the orbit and the height of the data is generally different, the sampling interval of the vertical motion data is made to match the sampling interval of the data of the uneven height (for example, 1 m) by linear interpolation (S6). Filtering is performed by a digital filter having a band-pass characteristic (FIG. 11) (S7), and temporarily stored (S8).

FIG. 12 shows the data shown in FIG. 3 which has been subjected to band extraction processing at a wavelength of 150 m or more. Very good similarities are recognized in the waveforms, indicating the possibility of correlation processing. The extracted wavelength band components are 15
Considering the result of the frequency response analysis of the Shinkansen (FIG. 9) and the margin for speed improvement in the future, the value of 0 m or more is determined as the wavelength whose phase delay is almost equal to 180 degrees.

Next, the detection of the displacement will be described. The elevation data temporarily stored in S4 corresponds to the data stored in the first memory 1 on a one-to-one basis, but as described above, the elevation data stored in the first memory 1 corresponds to kilograms on the ground. The position matching with the process is performed with high precision, and the required high and low sampling data can be extracted with reference to the high and low KDT.

On the other hand, the vertical sway data temporarily stored in S8 has a one-to-one correspondence with the data stored in the first memory 1, but the vertical sway data stored in the second memory 2 is simply a business train. Sampled at the measurement start point (data number 1)
The matching section (first matching section) where is present is clear, but the position (measurement start position) in the matching section is inaccurate, for example, near Shinagawa Station.

Then, out of the irregularity data temporarily stored in S4, the irregularity data for the verification section where the measurement is started on the vehicle is extracted with reference to the elevation KDT (S9).
The correlation processing with the up-and-down swing data temporarily stored in S8 is repeatedly executed until a maximum value having a correlation value of 0.7 or more can be detected (S10, S11). When the maximum value can be detected, the position deviation amount of the vertical sway data is detected from both sampling data giving the maximum value, and the data number of the sampling data of the vertical sway data is corrected by the position deviation amount and a kilometer. Is set in a table (first oscillation KDT) provided in the third memory 3 (S12).

For example, as shown in FIG. 13, a waveform having a peak value at the position of the displacement amount τ _max is obtained by the correlation processing. From the sampling data number (for example, 200) of the vertical sway, the sampling interval is set to 1 m, the amount of displacement 50 m is detected, the data number of the vertical sway is corrected to 150, and the corresponding kilometer is obtained from the high and low KDT. . In FIG. 13, the vertical motion data is acceleration data, and the phase of displacement is 1
Since it is shifted by 80 degrees, it is indicated by a negative maximum value.

When there are a plurality of first collation sections, S9 to S12 are repeatedly executed for each collation section via S13. At this time, when the value of the cross-correlation is 0.7 or more, it is determined that the correspondence of the position has been obtained. When the value is less than 0.7, it is determined that the correspondence cannot be obtained for some reason, and the matching section is excluded and the next matching section is excluded. Proceed to. When the train speed approaches 0, the vertical sway becomes smaller and affects the calculation result. Therefore, when the minimum train speed of the collation section becomes 30 km / h or less, the collation section is excluded and the process proceeds to the next collation section. I do.

When the processing for all collation sections is completed (S13), the relationship between the data number and the kilometer obtained in each collation section is not always the same, so that the data set in the first sway KDT to match them are matched. An error in the sampling interval of the vertical swing data is corrected from the relationship between the number and the kilometer (S14). Thus, the processing of the first step is completed, but in the present embodiment, the second step shown in FIG.
Move to step processing.

As is apparent from the above description, the correction is performed based on the data which is clearly corresponding to the level of kilometers on the ground, and the error in the sampling interval due to the wear of the wheels and the like in the course of this correction processing. The problem of missing sampling and duplication due to gliding and slipping of wheels and wheels is eliminated. In relation to the train speed, it can be avoided by not performing the position matching process in the matching section where the speed is equal to or less than the predetermined speed. Further, position matching can be performed without specifying the measurement start position on the vehicle.
This is because it is sufficient if there is a data amount capable of acquiring a correlation value of 0.7 or more. Specifically, the error of the measurement start kilometer is ±
If it is within 500 m, it can be automatically detected and corrected.

Here, the collation section in the first step is, for example, a section of 5 km since it is necessary to include as many gradient change points of the track as possible. Therefore, the maximum value obtained by the correlation processing is shown in FIG. As shown in FIG. 13, it is not sufficiently sharp and does not change much even if the positions of both data are slightly shifted. In other words, although it cannot be said that sufficiently high-precision position matching was performed, a position matching that can be used for a while Data is obtained.

The processing procedure of the second step shown in FIG. 6 is to further finely correct the positional deviation to the position collation data obtained in the first step so that sufficiently high precision position collation can be performed. is there. The correction of the minute displacement is performed by paying attention to the fact that the effect of the displacement appears most sensitively on the phase characteristic, and that it hardly changes in the long wavelength region but largely changes in the short wavelength region. This fact is evident from the fact that the influence of a small displacement of several meters is more affected as the wavelength is relatively shorter, considering the phase shift angle. For example, the displacement of 5 m is as large as 180 degrees for a wavelength of 10 m, but is only 18 degrees for a wavelength of 100 m.

According to the present invention, the positional relationship in the short wavelength band is detected by pattern matching between the frequency analysis result and the reference value, instead of extracting the phase characteristic in the short wavelength region by a filter. At this time, the reference value is strictly set to an “ideal value” so that incorrect association is not performed.

In the second step of the position matching process, since a short wavelength band such as a wavelength of 10 m is considered, for example, an error in a sampling interval, a fluctuation in a train speed, and a deviation in a track are measured, and then the fluctuation is measured. It is expected to be susceptible to various error factors, such as running out of track or performing track maintenance work during the period before running, aerodynamic sway not caused by height fluctuations in the tunnel, or swaying due to train passing. Is done.

Therefore, in the second step, the extension of the collation section (the second collation section) is set to a short distance (for example, 512 m) as if the first collation section was subdivided so as not to be influenced as much as possible. .

In FIG. 6, in S21, the first fluctuation KD
With reference to T, the vertical movement data for one second collation section is acquired from the second memory 2. The sampling interval of the up-and-down swing data in the second memory 2 is corrected by the process of S14.

In S22, since the result of the averaging process in S2 is temporarily stored, the height KDT of one second collation section is acquired from the result by referring to the height KDT.

In step S23, the data obtained in steps S21 and S22 are used to analyze the response of the vertical sway to fluctuations in height and to perform matching processing between the obtained frequency characteristics and the ideal characteristics. Execute while changing before and after the collation result obtained in
The data positional relationship between the most up-and-down incongruity and the up-down sway is detected.

When the frequency response analysis is executed by changing the positional deviation by a small amount before and after the collation result obtained in the first step, for example, as shown in FIG. , The pattern that best matches the phase characteristics of the correct position corresponding to the thick solid line in the center (ideal characteristics determined experimentally and shows good agreement with the theoretical values) is detected. At that time, the data positional relationship between the height fluctuation and the vertical movement is examined. It can be seen that this matching processing is equivalent to an operation of extracting the phase characteristics of a short wavelength band of about 10 m in wavelength.

In step S24, the data number of the sampling data of the up-down sway is corrected based on the detected data positional relationship between the ups and downs and the up-down sway, and is set in the second sway KDT of the third memory 3 together with the corresponding distance.

In S25, it is monitored whether or not the processing in S21 to S24 has been completed for all the second comparison sections.
Upon completion, the process advances to S26.

In step S26, similarly to step S14, the error in the sampling interval of the vertical fluctuation data is corrected again based on the relationship between the data number set in the obtained second fluctuation KDT and the kilometer.

FIG. 15 shows the distribution of the positioning error when the processing of the first step and the second step described above is actually performed using the measurement data of the 221 series (new rapid) on the Tokaido / Sanyo Main Line. Show. The error is ± 3 m per 2 km of the orbit, and it can be understood that the position matching can be performed with considerably high accuracy. The sampling interval had an error of about 0.5% in the first step, but the average value of the error was -0.15 m per 2 km due to the fine adjustment in the second step.
It was found that the error was further reduced to 0.008%, and that the position matching worked very accurately.

In this embodiment, the case where the position collation is performed in a batch processing format from the viewpoint of positioning as an additional function to the “micro LABOCS-II +” developed by the present inventors has been described. When the processing in the verification section is completed, the processing in the second verification section may be performed, and the processing may be repeated. That is, it is possible to perform real-time processing in which position matching is immediately performed while acquiring vertical swing data on a vehicle.

Further, when a very low frequency component having a wavelength of 200 m or more is extracted by a digital filter from the measured data of the vertical swing acceleration of the vehicle and the deviation of the track, as shown in FIG. An extremely good correspondence was obtained with the basic line shape, that is, an interesting result was obtained that each gradient inflection point of the vertical alignment and the unevenness of the waveform corresponded very well including its positive and negative.

This fact indicates that the traveling position can be determined only from the information on the track diagram and the data of the vertical swaying acceleration of the vehicle without using the track deviation data.

[0065]

As described above, the method of verifying the measurement position of the track deviation and the vertical movement measurement data of the vehicle according to the first invention is as follows.
Based on the finding that there is a very strong linear dependence between the height of the trajectory and the vertical motion of the vehicle, the low-frequency components of both data with similar waveforms are extracted, and the correlation processing is performed. A value is detected, a positional deviation of the vertical sway data is detected, and a table of the data number and the kilometer corrected for the deviation is created. Since the correction is based on data that is clearly up and down in correspondence with kilometers on the ground, in the process of this correction, errors in sampling intervals due to wheel wear, etc. The problem goes away. In relation to the train speed, it can be avoided by not performing the position matching process in the matching section where the speed is equal to or less than the predetermined speed. Further, according to the present invention, it is not necessary to specify the measurement start position on the vehicle.

Here, in the first invention, the collation section is, for example, a section of several km since it is necessary to include as many gradient change points of the track as possible. Therefore, the maximum value obtained by the correlation processing is sufficiently large. It is not sharp, and it does not change much even if the positions of both data are slightly shifted. That is, it cannot be said that sufficiently high-accuracy position collation can be performed, but there is an effect that position collation data that can be used for a while is obtained.

In the second aspect of the present invention, the method of verifying the position of the track deviation and the vertical movement measurement of the vehicle according to the second invention is highly sensitive to the effect of the positional deviation in order to enhance the accuracy of the position verification data obtained in the first invention. Focusing on the fact that the phase characteristic appears, and that it hardly changes in the long wavelength range but changes greatly in the short wavelength range, the second section corresponding to subdivision of the matching section (first matching section) of the first invention. A collation section is set, and for each second collation section, the frequency response analysis and the matching process are performed while changing the sampling data back and forth, and the positional relationship between the orbital deviation and the vertical sway data indicating the best match is detected. Since the position matching table is created, there is an effect that considerably high-accuracy position matching data can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a principle configuration for implementing a measurement position collation method of measurement data of track deviation and vertical movement of a vehicle according to the present invention.

FIG. 2 is an example (Shinkansen) of measured data of elevation deviation and vertical sway acceleration.

FIG. 3 is an example (conventional line) of measurement data of elevation deviation and vertical sway acceleration.

FIG. 4 is an example of a kilometer / data number comparison table (KDT).

FIG. 5 is a procedure flowchart of a position matching process (first step) of the present invention.

FIG. 6 is a procedure flowchart of a position matching process (second step) of the present invention.

FIG. 7 is an explanatory diagram of a method of analyzing the frequency response of the vertical oscillating acceleration with respect to height deviation.

FIG. 8 is a vibration model diagram of the up-and-down swing acceleration of a vehicle that is excited due to height deviation.

9A and 9B are examples of frequency response estimation (shinkansen) of vertical sway acceleration with respect to fluctuations in elevation, wherein FIG. 9A is an amplitude characteristic diagram, FIG. 9B is a phase characteristic diagram, and FIG. 9C is a coherence characteristic diagram.

10A and 10B are examples of frequency response estimation (vertical line) of vertical sway acceleration with respect to height deviation, where FIG. 10A is an amplitude characteristic diagram, and FIG.
Is a phase characteristic diagram, and (c) is a coherence characteristic diagram.

11A and 11B are frequency characteristics of a low-pass filter, wherein FIG. 11A is an amplitude characteristic diagram and FIG. 11B is a phase characteristic diagram.

FIG. 12 is a diagram (conventional line) showing a relationship between height deviation and vertical oscillation acceleration in a long wavelength band.

FIG. 13 is a cross-correlation waveform diagram (wavelength of 150 m or more) between vertical deviation and vertical oscillation acceleration.

FIG. 14 is an explanatory diagram of an influence of a minute positional deviation on a phase characteristic (a thick solid line at the center is an ideal characteristic indicating a correct positional correspondence).

FIG. 15 is a distribution diagram of a positioning error in position matching performed on the Tokaido / Sanyo Main Line 221 system.

FIG. 16 is a diagram showing the relationship between a vertical alignment, a deviation in elevation, and vertical acceleration.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 1st memory 2 2nd memory 3 3rd memory 4 CPU

Claims (6)

(57) [Claims] A first memory for storing orbital deviation data; a means for averaging left and right track deviation data stored in the first memory; and a low-pass filter for the averaged deviation data. Means for performing wave processing; second memory for storing up-and-down swing data; means for converting the sampling interval of up-and-down swing data stored in the second memory to the sampling interval for data of irregular orbits in the trajectory; Means for performing low-pass filtering processing on the up-and-down oscillating data; means for performing correlation processing between the up-and-down oscillating data subjected to the low-pass filtering processing in a predetermined collation section and the up-and-down oscillating data to detect a maximum value; Detects the amount of displacement of the vertical sway data from both sampling data giving the maximum value, and changes the data number of the sampling data of the vertical sway data to that position. Means for setting the relationship between the amount corrected by the quantity and the kilometer in a table; and means for correcting the sampling interval of the vertical sway data from the relationship between the data number and the kilometer set in the table. A method to verify the position of the track data and the vertical movement of the vehicle. 2. The method according to claim 1, wherein each step from detection of a maximum value to setting of a table is performed for each predetermined verification section. method. 3. The method according to claim 1, wherein the track diagram data is used in place of the track height error data. 4. A first memory for storing orbital up / down data; means for averaging left and right track up / down data stored in the first memory; and low-pass filtering of the averaged up / down data. Means for performing wave processing; second memory for storing up-and-down swing data; means for converting the sampling interval of up-and-down swing data stored in the second memory to the sampling interval for data of irregular orbits in the trajectory; Means for performing a low-pass filtering process on the up-and-down motion data;
The maximum value is detected by performing a correlation process between the up-and-down fluctuation data and the up-down fluctuation data subjected to the low-pass filtering process in the first collation section, and the displacement of the up-down fluctuation data from both sampling data giving the maximum value. Means for detecting the amount and correcting the data number of the sampling data of the up-and-down swing data by the positional deviation amount, and setting the relationship between the distance and the kilometer in the first table; the data number set in the first table and the kilometer Means for correcting the sampling interval of the up-and-down motion data based on the relationship between; and for each second collation section shorter than the first collation section, referencing the averaged deviation data in the second collation section and the first table. Using the vertical sway data obtained by correcting the sampling interval taken from the second memory, the vertical sway response analysis to the deviation of the height and the obtained frequency response characteristics and Matching process with the typical frequency response characteristics by changing one or both of the sampling data of the orbit deviation data and the vertical sway data back and forth, and giving the frequency response characteristics that best match the ideal frequency response characteristics Means for detecting the positional relationship between the height fluctuation data and the vertical sway data, and setting the relation between the data number of the sampling data of the detected vertical sway data and the kilometer in the second table; the data number set in the second table Means for correcting the sampling interval of the vertical sway data based on the relationship between the vertical skew data and the kilometer range. 5. The method according to claim 4, wherein when the processing in at least one first verification section is completed, the processing in the second verification section is performed, and the processing is repeated. Data measurement position collation method. 6. A method according to claim 1, wherein the reading of the data having the irregularity in the collation section in which the correlation processing is performed is performed by referring to a table which defines a relationship between a data number of the sampling data and a kilometer. The method of claim 1 or claim 4, wherein the measured position of the track deviation and the vertical movement of the vehicle is measured.