Determination device

Abstract

A determination device 229 according to the present disclosure classifies an equipment environment around an optical cable 21 in the cable longitudinal direction, and includes: a feature extraction unit 213 that extracts a feature amount for each position of the optical cable 21 in the longitudinal direction from the vibration distribution of the optical cable 21 in the longitudinal direction; an identification function value calculation unit 220 that calculates an identification function value y for identifying equipment around the optical cable 21 on the basis of the feature amount and a weight vector corresponding to the natural vibration of the equipment around the optical cable 21; and a class determination unit 222 that identifies the equipment on the basis of the result of calculating the identification function value y.

Description

Judgment device

[0001] The present disclosure relates to a determination device, a method, and a program for identifying facilities around an optical communication cable.

[0002] To continue providing safe and secure optical fiber communication networks, it is important to prevent failures in the networks that extend over a wide area. Optical cables buried near the ground surface are particularly susceptible to influences from the surrounding environment, and the possibility of unintended optical cable failures causing communication failures cannot be denied. It is therefore important to predict the occurrence of such optical cable failures and prevent communication failures before they occur.

[0003] However, while the current OTDR (Optical Time Domain Reflectometry) method (see, for example, Patent Document 1), which is an optical evaluation method for remote monitoring and testing of optical cables, can identify the location of a fault (disconnection) in an optical cable after the fact by measuring distance loss, it cannot predict the occurrence of a fault in the optical cable, and it is difficult to prevent communication failures from occurring.

[0004] In recent years, a distributed acoustic sensing (DAS) method has emerged that uses a C-OTDR (Coherent Optical Time Domain Reflectometry) method, which applies a coherent detection method to the OTDR method, to measure vibration distribution from phase changes in the waveform of backscattered light in the longitudinal direction of a continuous optical fiber by narrowing the laser linewidth (see, for example, Non-Patent Document 1). According to this method, optical fibers are used as sensors to detect vibrations applied from the environment surrounding the optical cable, making it possible to predict the occurrence of a failure in the optical cable.

[0005] Special Publication No. 7-28266

[0006] Y. Wakisaka, D. Iida, H. Oshida and N. Honda, “Fading Suppression of fwai-OTDR With the New Signal Processing Methodology of Complex Vectors Across Time and Frequency Domains,” in Journal of Lightwave Technology, vol. 39, no. 13, pp. 4279-4293, July 1, 2021, doi: 10.1109 / JLT. 2021.3071159.

[0007] If it were possible to identify the equipment around the optical cable based on the vibration distribution applied from the environment around the optical cable, it would be possible to identify the location of the vibration source by referring to information about the location of the equipment. However, the method disclosed in Non-Patent Document 1 alone can grasp that the vibration applied to the optical cable affects the quality of communication services, but it cannot identify the equipment around the optical cable, and ultimately cannot lead to identifying the location of the vibration source.

[0008] Therefore, an object of the present disclosure is to provide a determination device that can identify equipment around an optical cable from the vibration distribution (vibration distribution waveform) in the longitudinal direction of the optical cable.

[0009] In order to achieve the above object, the determination device of the present disclosure employs a method of extracting features from a vibration distribution, calculating a discrimination function value based on the features and a weight vector, and then identifying equipment based on the result of the calculation of the discrimination function value.

[0010] Specifically, the determination device of the present disclosure is a determination device that classifies the equipment environment in the longitudinal direction of an optical cable, and includes: a feature extraction unit that extracts feature values ​​for each position in the longitudinal direction of the optical cable from the vibration distribution in the longitudinal direction of the optical cable; a calculation unit that calculates a discrimination function value for identifying equipment around the optical cable based on the feature values ​​and a weight vector corresponding to the natural vibration of the equipment around the optical cable; and a determination unit that identifies the equipment based on the calculation result of the discrimination function value.

[0011] The present disclosure also provides a determination method for classifying an equipment environment in the longitudinal direction of an optical cable, the determination method including: extracting feature values ​​for each position in the longitudinal direction of the optical cable from the vibration distribution in the longitudinal direction of the optical cable; calculating a discrimination function value for identifying equipment around the optical cable based on the feature values ​​and a weight vector corresponding to the natural vibration of the equipment around the optical cable; and identifying the equipment based on the result of calculating the discrimination function value.

[0012] This makes it possible to identify the facility (manhole, etc.) by capturing the characteristic of the natural vibration of the facility from the vibration distribution in the longitudinal direction of the optical cable.

[0013] The system may further include an identification dictionary that holds a weight vector that is corrected based on an identification function value calculated from a linear sum of a first feature amount related to a vibration distribution of a first facility and a second feature amount related to a vibration distribution of a second facility different from the first facility and a weight vector, wherein the calculation unit calculates an identification function value for identifying the first facility and the second facility from a third feature amount related to the vibration distribution when the first facility and the second facility are mixed and the weight vector held in the identification dictionary, and the determination unit may identify the first facility and the second facility based on a result of calculation of the identification function value by the calculation unit.

[0014] This makes it possible to grasp the characteristics of the natural vibrations of a plurality of pieces of equipment from the vibration distribution in the longitudinal direction of the optical cable, and to identify the plurality of pieces of equipment as being different from one another.

[0015] The calculation unit may also include a sequence generation unit that calculates the classification function value at predetermined time intervals, replaces the classification function value with a corresponding teacher signal value, and generates a sequence of the teacher signal in the time direction, and the determination unit may identify the equipment based on the length of the sequence.

[0016] This allows us to accurately identify the source of vibration by dividing the vibration distribution data into time series and repeating class determination.

[0017] The determination device may also include a vibration coordinate estimation unit that stores coordinate information of the equipment to be identified and compares the identification result of the equipment with the coordinate information to estimate the position of the vibration source.

[0018] This allows the location of the vibration source to be suitably estimated by comparing the equipment identification result with the coordinate information.

[0019] The present disclosure also provides a program for causing a computer to function as the above-described determination device.

[0020] The determination device may also be configured as a system in which some or all of its components are mounted on different devices.

[0021] The above disclosures can be combined as much as possible.

[0022] According to the determination device of the present disclosure, it is possible to identify the equipment around the optical cable from the vibration distribution in the longitudinal direction of the optical cable.

[0023] FIG. 1 is a diagram illustrating a threshold logic unit that identifies equipment around an optical cable from the vibration distribution around the optical fiber. FIG. 2 is a diagram illustrating a configuration for equipment identification. FIG. 3 is a diagram illustrating detection of vibration distribution by optical testing performed by a vibration distribution measuring instrument. FIG. 4 is a diagram illustrating the Fourier transform performed by a feature extraction unit and the definition of feature vectors. FIG. 5 is a diagram illustrating an example of class determination based on a generated sequence. FIG. 6 is a block diagram of a system according to an embodiment.

[0024] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. Note that the present disclosure is not limited to the embodiments shown below. These implementation examples are merely illustrative, and the present disclosure can be implemented in various forms with various modifications and improvements based on the knowledge of those skilled in the art. Note that components with the same reference numerals in this specification and drawings indicate the same components.

[0025] (Summary of the Invention) An embodiment according to the present disclosure will be described below. Fig. 1 is a diagram illustrating a threshold logic unit that is the basis of an identification algorithm for identifying equipment around an optical cable from a vibration distribution according to the present disclosure. Here, 11: the peak frequency (f １ , f ２ , ..., f ｋ , ..., f ｌ , ..., f ｎ－１ , f ｎ ), 12: Initial value, 13: Linear sum, 14: Variance value σ 2 , 15: average value f ａｖｇ． , 16: identification dictionary, 17: identification function value y, and 18: filter. Here, the surroundings of the optical cable are not limited to a narrow range along the optical cable. In other words, the present disclosure includes identifying equipment distant from the optical cable based on vibration distribution.

[0026] The vibration distribution facility identification algorithm using a threshold logic unit is established for the following two purposes: (1) To identify the facility (e.g., manhole) by grasping the characteristic vibration of the facility from the vibration distribution in the longitudinal direction of the optical cable, and (2) To establish a class determination algorithm that identifies the vibration source in real time.

[0027] The basis of the threshold logic unit is to calculate the linear sum of multiple input signals and output the result. Then, the output value from the multiple input signals is compared with a prepared teacher signal, and the input signal is classified into the class corresponding to the teacher signal that is closest to the output value. In this invention, the peak frequencies (f １ , f ２ , ..., f ｋ , ..., f ｌ, ..., f ｎ－１ , f ｎ ) 11 is defined. These are then added with 1, which is the value of the initial value 12, and the resulting n+1 peak frequencies are sent to the linear sum 13 as peak frequencies. Ideally, the multiple peak frequencies would be uniquely determined by the object to be identified (here, a manhole), but the input signal is not always uniquely determined for each manhole of different size (the peak frequency value fluctuates). On the premise that the overall variation and average value do not change even if the peak frequency value fluctuates, the following variance value σ, which is selected from the peak frequencies of the n Fourier spectra and calculated at the same time as the n+1 peak frequencies, is calculated. 2 The following average value f was calculated by selecting 14 and 1 ａｖｇ． 15 is sent to the linear sum 13. Here, the number of elements for calculating the variance and the mean may be the same (k=l). In the linear sum 13, a weight vector (ω) having the same number of elements (n+3) as the number of input signals is calculated. １， ω ２， ..., ω ｎ， ω ｎ＋１， ω ｎ＋２， ω ｎ＋３ ) is obtained from the identification dictionary 16 and the linear sum Σ(=ω １ f １ +ω ２ f ２ +...+ω ｎ f ｎ +ω ｎ＋１ +ω ｎ＋２ σ 2 +ω ｎ＋３ f ａｖｇ． ) is calculated. The calculated value is set as the discrimination function value y17. The discrimination function value y17 is sent to the filter 18, which outputs a discrimination function value y17 in a range close to the teacher signal (here, 1 or -1). Thereafter, the facility (manhole) is determined based on the output discrimination function value y17.

[0028] (First embodiment) Hereinafter, a method for extracting feature amounts (feature vectors), a method for discriminative calculation, and class determination will be described in detail.

[0029] FIG. 2 is a diagram illustrating a configuration for identifying equipment. Here, the components are: 21: optical cable, 22: ground, 23: conduit, 24 and 25, 26: manhole (MH), 27: vibration distribution measuring instrument, 28: vibration source, 29: first vibration distribution data, 210: second vibration distribution data, 211: third vibration distribution data, 212: memory unit, 213: feature extraction unit, 214: data reading unit, 215: time-frequency conversion unit, 216: peak frequency extraction unit, 217: peak frequency variance / average calculation unit, 218: identification dictionary, 219: identification calculation unit, 220: identification function value calculation unit, 221: sequence generation unit, 222: class determination unit, 223: result display unit, 224: identification function value calculation unit, 225: error calculation unit, 226: teacher signal, 227: squared error calculation unit, 228: weight vector update unit, 229: determination device, and 230: vibration coordinate estimation unit. The discriminant function value calculation unit 220 functions as a "calculation unit." The class determination unit 222 functions as a "determination unit."

[0030] Specifically, the determination device 229 is a determination device 229 that classifies the equipment environment in the longitudinal direction of the optical cable 21, and includes: a feature extraction unit 213 that extracts features for each position in the longitudinal direction of the optical cable 21 from the vibration distribution in the longitudinal direction of the optical cable 21; a discrimination function value calculation unit 220 that calculates a discrimination function value y for identifying equipment around the optical cable 21 based on the features and a weight vector corresponding to the natural vibration of the equipment around the optical cable 21; and a class determination unit 222 that identifies the equipment based on the calculation result of the discrimination function value y.

[0031] One end of an optical cable 21 is connected to a vibration distribution measuring instrument 27, and the optical cable 21 is inserted into a conduit 23 buried in the ground 22. Manholes (MH) 24, 25, and 26 are provided along the conduit 23 for maintenance purposes. Coordinate information for the manholes is stored in a database in advance. The vibration distribution measuring instrument 27 detects the vibration distribution in the longitudinal direction of the optical cable by a distributed acoustic sensing (DAS) method (see, for example, Non-Patent Document 1) using optical time domain reflectometry (OTDR) or coherent optical time domain reflectometry (C-OTDR). Here, when the vibration distribution measuring instrument 27 detects vibrations from a vibration source 28 that affect the provision of communication services on the optical cable 21, the only available location information is distance information from the vibration distribution measuring instrument 27. On the other hand, to dispatch a maintenance personnel to the site, direct location information such as coordinates is required. Therefore, in this embodiment, the natural vibration of the MH is identified from the vibration distribution obtained by the vibration distribution measuring instrument 27, and the coordinates of the vibration source 28 are estimated by comparing the MH position information (=distance information) with MH coordinate information in a database. Specifically, the MH affected by vibrations from the vibration source 28 is identified by comparing the MH distance information with the MH coordinate information in the database, and the position of the vibration source 28 is estimated based on the results of identifying the MH. For example, if the vibration source 28 is near the MH 25, the location is estimated from the coordinate information of the MH 25.

[0032] Thus, the objective of the present invention is to extract and learn the characteristics of the natural vibration of MH from the vibration distribution and identify MH locations along the length of an optical cable. The vibration distribution is acquired by installing a vibration distribution measuring device 27 at one end of the optical cable and using the distributed acoustic sensing (DAS) method (Non-Patent Document 1) that uses coherent optical time domain reflectometry (C-OTDR). Continuously detecting the vibration distribution allows acquisition of three-dimensional data of optical cable distance, time, and vibration magnitude. Hereinafter, this three-dimensional data will be referred to as vibration distribution data. The vibration distribution data includes vibration components due to events on the ground 22 (vehicle traffic, road construction) as well as equipment environments such as MH and conduits 23. While an example of identifying MH is described here, the present disclosure can also be applied to identifying equipment other than MH. The present disclosure also includes identifying two or more types of point equipment as different point equipment by using multiple teacher signals. The present disclosure also includes identifying two or more types of communication equipment as different communication equipment by using multiple teacher signals. The MH is an example of a "first equipment," and the pipeline 23 is an example of a "second equipment."

[0033] [Detection of vibration distribution by optical testing performed by vibration distribution measuring instrument] Here, with reference to Fig. 3, a supplementary explanation will be given of the measurement of vibration distribution by optical testing (C-OFDR) performed by the vibration distribution measuring instrument 27. In Fig. 3, the respective symbols indicate the following: 41: light intensity distribution, 42: local section, 43: waveform pattern, 44: waveform pattern after Δt seconds, 45: change Δν. Fig. 3(A) shows the light intensity distribution over the entire longitudinal direction of the optical fiber, and Fig. 3(B) shows an enlarged view of the light intensity of local section 42.

[0034] Conventional C-OFDR measures the light intensity distribution 41 along the optical fiber's length. Laser light is incident on the optical fiber inside the optical cable, and changes in light intensity are observed by receiving the backward Rayleigh scattered light that propagates in the opposite direction to the incident direction. Focusing on the light intensity waveform in a localized section 42, a waveform corresponding to the inherent characteristics of the optical fiber can be observed, and this waveform shows the same waveform pattern 43 when the conditions of the optical fiber, laser light, etc. are unchanged. The central wavelength of laser light with a wide linewidth is constantly changing, and the waveform pattern also changes.

[0035] With the advancement of laser technology and the emergence of narrow linewidth lasers, the effect of changes in the central wavelength becomes less of a problem, and the waveform will have the same pattern if the optical fiber state is the same. If a vibration specific to the optical fiber is added, the waveform pattern 44 after Δt seconds will be different due to the influence of the vibration. The vibration is detected from the change Δν 45 between waveform pattern 43 and waveform pattern 44.

[0036] By moving the local section 42, the vibration distribution in the longitudinal direction of the optical fiber is detected. In addition, by continuously measuring the light intensity distribution and detecting vibrations each time, it is possible to observe the time change in vibration in the longitudinal direction of the optical fiber. In this way, three-dimensional vibration distribution data of optical cable distance - time - vibration magnitude is obtained. In the case of C-OTDR, the vibration distribution is detected by focusing on changes in phase information obtained from the waveform pattern.

[0037] [Feature Extraction Method] Returning to FIG. 2 , three pieces of vibration distribution data are acquired using a vibration distribution measuring instrument 27 to identify MHs. The first vibration distribution data 29 is vibration distribution data indicating the magnitude of vibration with respect to the time direction at the position of the MH. This data includes a collection of data obtained by striking the MH with a hammer or the like and dividing the time during which the vibration occurs into predetermined time intervals, and collecting the vibration distribution data for each different MH. The second vibration distribution data 210 is a collection of vibration distribution data indicating the magnitude of vibration with respect to the time direction at an arbitrary position in the conduit 23 other than the MH. The third vibration distribution data 211 is vibration distribution data in which it is unknown whether the equipment near the optical cable is an MH or a conduit 23. The first vibration distribution data 29 and the second vibration distribution data 210 are used as learning data. The three pieces of vibration distribution data are stored in a memory unit 212. Coordinate information for the MHs and the conduit 23 is also stored in advance in the memory unit 212.

[0038] The vibration distribution data 29, 210, and 211 are sent to a feature extraction unit 213. A data reading unit 214 reads each of the vibration distribution data 29, 210, and 211 independently and passes the data to the subsequent process, i.e., a time-frequency conversion unit 215. The time-frequency conversion unit 215 converts the amplitude of vibration continuously observed at each point in the longitudinal direction of the optical cable for the vibration distribution data 29, 210, and 211 passed from the data reading unit 214 from the time domain to the frequency domain. In other words, the vibration distribution data 29, 210, and 211 are three-dimensional data of optical cable distance-time-vibration amplitude, which are converted by the time-frequency conversion unit 215 into three-dimensional data of optical cable distance-frequency-amplitude. The time-frequency conversion may be performed using a Fourier transform or a cepstrum transform that extracts the envelope of a Fourier spectrum. The data after the time-frequency conversion is passed to a peak frequency extraction unit 216.

[0039] The peak frequency extraction unit 216 extracts the peak frequency (f ｄ，１ , f ｄ，２ , ..., f ｄ，ｋ , ..., f ｄ，ｌ , ..., f ｄ，ｎ－１ , f ｄ，ｎThe peak frequency is passed to the peak frequency variance / average calculation unit 217.

[0040] As described above, the peak frequency variance / average calculation unit 217 selects k or l peak frequencies from the n Fourier spectrum peak frequencies and calculates the variance value σ 2 and the average value f ａｖｇ． is calculated and passed to the subsequent process as a feature amount (feature vector) together with the peak frequency. The feature amounts (feature vectors) of the first vibration distribution data 29 and the second vibration distribution data 210 are passed to an identification dictionary 218. The feature amount (feature vector) of the third vibration distribution data is passed to an identification calculation unit 219.

[0041] [Fourier Transform Performed by Feature Extraction Unit and Definition of Feature Vector] Here, with reference to Fig. 4, a supplementary explanation will be given of the Fourier transform performed by the feature extraction unit 213 and the definition of the feature vector. In Fig. 4, the symbols indicate the following: 51: waveform g(t), 52: Fourier spectrum. Fig. 4(A) is a diagram explaining the waveform g(t) 51, and Fig. 4(B) is a diagram explaining the Fourier spectrum obtained by Fourier transforming the waveform g(t) 51.

[0042] A waveform 41g(t) showing a continuous change in vibration in a local section 32 indicates the magnitude of vibration in the time direction over a plot interval Δt. An example of a transformation formula from the time domain to the frequency domain is shown below. where F(ω) is the waveform obtained by converting the waveform g(t) into the frequency domain, and f is the frequency [Hz].

[0043] The waveform g(t) is transformed into the frequency domain to become a Fourier spectrum 52. In the Fourier spectrum 52, peaks are observed at specific frequencies depending on the vibration components. The frequencies of the spectrum peaks are sorted in descending order of their peak values ​​(f １ , f ２ , ..., f ｋ , ..., f ｌ , ..., f ｎ－１ , f ｎ ) are assigned. These processes are performed by moving the local section 32 along the entire length of the optical fiber. １ , f ２ , ..., fｋ , ..., f ｌ , ..., f ｎ－１ , f ｎ ) and the peak amplitude are passed to the discrimination calculation unit 219 or the discrimination dictionary 218 as feature quantities.

[0044] [Discrimination Calculation Method] Returning to Fig. 2, the feature amount (feature vector) of the third vibration distribution data 211 passed to the discrimination calculation unit 219 is passed as an input signal to the discrimination function value calculation unit 220, which calculates one discrimination function value y.

[0045] Specifically, the discrimination function value calculation unit 220 calculates a discrimination function value y at predetermined time intervals, replaces the discrimination function value y with the value of a corresponding teacher signal, and includes a sequence generation unit 221 that generates a sequence of teacher signals in the time direction, and the class determination unit 222 identifies equipment based on the length of the sequence.

[0046] Here, a method for calculating the discriminant function value y will be described. The feature quantity (feature vector) of the third vibration distribution data 211 is calculated using n+3 signals (f ｄ，１ , f ｄ，２ , ..., f ｄ，ｎ－１ , f ｄ，ｎ , 1, σ 2 ,f ａｖg． ) and the n+3 coefficients (weight vectors) (ω ｄ，１， ω ｄ，２， ..., ω ｄ，ｎ－１， ω ｄ，ｎ， ω ｄ，ｎ＋１， ω ｄ，ｎ＋２， ω ｄ，ｎ＋３ ) is received from the discrimination dictionary 218, and the following linear sum (=discrimination function value y) is calculated.

[0047] The above is the discriminant function value at a distance d in the longitudinal direction of the optical cable. Also shown below is a matrix of discriminant function values ​​for the time length t obtained by dividing the time during which the MH is struck and the vibration is generated into predetermined time intervals from the start time of vibration distribution measurement. Note that it is up to the user to decide at what distance in the longitudinal direction of the optical cable the discriminant function value is calculated. Each row in the matrix below corresponds to a different time, and each column corresponds to a different distance. In this way, the vibration distribution measurement time is t ｍａｘIf the time is d×t seconds, the calculation unit calculates d×t from the feature quantities (feature vectors) of all the third vibration distribution data 211 that have been passed. ｍａｘ / t (y ０ , y １ , ..., y ｄ－１ ) is calculated. ｍａｘ It is not necessary to have 20 (y / t) pieces (for example, if it is a vibration distribution of 10 seconds and the time length t is 0.5 seconds, 20 (y ０ , y １ , ..., y ｄ－１ ) is output, but it is not necessary to divide the data into 0.5-second intervals without any gaps; calculations may be performed every 0.5 seconds while shifting the data in the time axis direction by 0.2 seconds. ) A classification function value close to the teacher signal set in the MH is extracted by filtering and passed to the sequence generation unit 221. For example, when the teacher signal of the MH is "1," the sequence generation unit 221 extracts a classification function value corresponding to 1.1 > y > 0.9. Here, the threshold value is determined arbitrarily. The sequence generation unit 221 replaces the extracted classification function value with the teacher signal, and a sequence of teacher signals in the distance direction and the time direction is generated. The generated sequence of values ​​is passed to the class determination unit 222. Note that the classification calculation unit 220 may store data related to the teacher signal in advance for sequence generation.

[0048] The class determination unit 222 determines the class based on the generated series model for each distance d point. For example, the class determination unit 222 determines that an MH is installed at a certain distance point when a series of 1s (1, 1, 1, 1) and 1s occur in the time direction at that distance point. The determination result is passed to the result display unit 223, which displays the MH equipment in the longitudinal direction of the optical cable. Specific examples of class determination will be described later.

[0049] For example, when the teacher signal of a pipeline 23 other than an MH is "-1," the series generation unit 221 may be configured to extract a discriminant function value in the range of -1.1 < y < -0.9. In this case, when a series of (-1, -1, -1, -1) and -1 continues in the time direction at a certain distance point, the class determination unit 222 may determine that a pipeline 23 other than an MH is present at that distance.

[0050] [Vibration Coordinate Estimation] At the same time, the class determination unit 222 passes the determination result to the vibration coordinate estimation unit 230. The vibration coordinate estimation unit 230 identifies the MHs affected by vibration from the vibration source 28 by comparing the MH distance information with the MH coordinate information stored in advance in the storage unit 212, and estimates the position of the vibration source 28 based on the result of the identification of the MH. The result of the estimation is passed to the result display unit 223, and information regarding the position (coordinates) of the vibration source 28 is displayed.

[0051] Specifically, the determination device 229 of the present disclosure includes a vibration coordinate estimation unit 230 that stores coordinate information of the equipment to be identified and compares the equipment identification results with the coordinate information to estimate the position of the vibration source.

[0052] [Weight Vector Update Method] The accuracy of the discriminant function value depends on the weight vector received from the discriminant dictionary 218. The weight vector update method will be described below.

[0053] A set of feature quantities (feature vectors) of the first vibration distribution data 29 and the second vibration distribution data 210 passed to the discrimination dictionary 218 is passed as an input signal to the discrimination function value calculation unit 224. At this time, the feature quantities (feature vectors) of the first or second vibration distribution data, which indicate the magnitude of vibration with respect to the time direction at the position of the MH and at any position of the pipe 23 other than the MH, are expressed as n+3 signals (f １ , f ２ , ..., f ｎ－１ , f ｎ , 1, σ 2 , f ａｖg． Each of the n+3 signals is assigned a weight vector (ω) consisting of n+3 elements that are appropriately set as initial values ​​in advance. １， ω ２，・・・， ω ｎ－１， ω ｎ， ω ｎ＋１， ω ｎ＋２， ω ｎ＋３ ) are multiplied by the corresponding elements of the third vibration distribution data 211, and the discriminant function value y is calculated using the above-mentioned linear sum Σ in the same way as for the third vibration distribution data 211.

[0054] The calculated discriminant function value y is passed to the error calculation unit 225. The error calculation unit 225 calculates the error with respect to a teacher signal 226 set in advance for each category. (Equation 7) y-bi (7) where b ｉ : Teacher signal (for example, MH is "1" and pipeline is "-1"). The error is passed to the square error calculation unit 227. The square error calculation unit 227 executes the following calculation of the square error: (y-bi) (Equation 8) ２ (8)

[0055] The calculated squared error is passed to the weight vector update unit 228. Meanwhile, the error is also passed to the weight vector update unit 228, and the weight vector is updated by the following equation. where ρ is a learning coefficient (any value). Note that the weight vector may be updated based on a teacher signal corresponding to any one of the feature quantities of the plurality of pieces of equipment, or may be updated based on any or all of the plurality of teacher signals. In other words, the weight vector may be updated based on a linear sum of any one of the feature quantities of the plurality of pieces of equipment and the weight vector, or may be updated based on a linear sum of any or all of the feature quantities of the plurality of pieces of equipment and the weight vector. Furthermore, the identification dictionary 218 may store data related to the teacher signals in advance for updating the weight vector.

[0056] The updated weight vector (ω ’ １， ω ’ ２，・・・， ω ’ ｎ－１， ω ’ ｎ， ω ’ ｎ＋１， ω ’ ｎ＋２， ω ’ ｎ＋３ ) is passed to the discriminant function value calculation unit 224, and the weight vector (ω ｄ，１， ω ｄ，２，・・・， ω ｄ，ｎ－１， ω ｄ，ｎ， ω ｄ，ｎ＋１， ω ｄ，ｎ＋２， ω ｄ，ｎ＋３) and a discriminant function value is calculated as described above for the third vibration distribution data 211. The discrimination dictionary 218 repeatedly calculates equations (6) to (9). The repetition ends when the squared error calculated by equation (8) becomes equal to or less than a predetermined value (threshold) and stable. When the number of data sets of feature amounts (feature vectors) of the first vibration distribution data 29 and the second vibration distribution data 210 passed to the discrimination dictionary 218 is a, the squared error may be calculated using the following:

[0057] As mentioned above, the weight vector (ω ｄ，１， ω ｄ，２，・・・， ω ｄ，ｎ－１， ω ｄ，ｎ， ω ｄ，ｎ＋１， ω ｄ，ｎ＋２， ω ｄ，ｎ＋３ ) is used in the discriminant function value calculation unit 220.

[0058] In this way, the judgment device 229 includes an identification dictionary 218 that holds a weight vector corrected based on an identification function value calculated from a linear sum of a first feature related to the vibration distribution of a first equipment (e.g., MH) and a second feature related to the vibration distribution of a second equipment (e.g., pipeline 23) different from the first equipment and a weight vector; a discrimination function value calculation unit 220 calculates a discrimination function value for identifying the first equipment and the second equipment from a third feature related to the vibration distribution when the first equipment and the second equipment are mixed and the weight vector held by the identification dictionary 218; and a class judgment unit 222 identifies the first equipment and the second equipment based on the result of the calculation of the discrimination function value by the discrimination function value calculation unit 220.

[0059] [Class Determination Example] FIG. 5 is a diagram illustrating an example of class determination based on a generated sequence. Here, the process of passing the values ​​generated by the sequence generation unit 221 to the class determination unit 222 for processing will be described. FIG. 5(A) shows an example of determining that a MH is present at a given distance based on a sequence with only one teacher signal "1" in the time direction. FIG. 5(B) shows an example of determining that a MH is present at a given distance based on a sequence with three consecutive teacher signal "1"s in the time direction. Here, 31 is sequence (1), 32 and 36 are plots of distance-time-MH determination, 33 is the result of the MH determination (correct determination result), 34 and 37 are results of incorrect determination, and 35 is sequence (1, 1, 1).

[0060] As shown in FIG. 5A, when the model included in the class determination unit 222 determines that a MH is located at a given distance based on a sequence containing only one teacher signal in the time direction, if the value obtained by substituting the teacher signal for the filtered classification function value is 1, the model is determined to be class 1, i.e., MH. In this case, the result display unit 223 displays a distance-time-MH determination plot 32. The results include a determination result 33 in which the natural vibration of the MH is identified, but also many erroneous determination results 34. If the time length t of the vibration distribution data (the time during which the MH is struck and the vibration is generated, divided into predetermined time intervals) is shorter than the time during which the natural vibration of the MH occurs, the MH determination results are continuous. Therefore, as shown in FIG. 5B, when the model included in the class determination unit 222 determines that a MH is located at a given distance based on a sequence of three consecutive teacher signals (1, 1, 1) in the time direction, the distance-time-MH determination plot 36 will show erroneous determination results 37, which can be reduced to a smaller number than the erroneous determination results 34. This is expected to improve the accuracy of MH determination. Here, the model sequence is not limited to (1, 1, 1), and it may be determined that there is MH at the distance when, for example, 1 appears consecutively any number of times, for example, two or more times.

[0061] Second Embodiment The determination device 229 can also be realized by a computer and a program, and the program can be recorded on a recording medium or provided via a network. The program of the present disclosure is a program for causing a computer to realize each function of the device according to the present disclosure, and a program for causing a computer to execute each procedure of the method executed by the device according to the present disclosure. Figure 6 is a block diagram of a system 100 according to an embodiment. The system 100 includes a computer 105 connected to a network 135.

[0062] Network 135 is a data communications network. Network 135 may be a private or public network and may include, for example, any or all of the following: (a) a personal area network covering a room, (b) a local area network covering a building, (c) a campus area network covering a campus, (d) a metropolitan area network covering a city, (e) a wide area network covering an area spanning city, region, or country boundaries, or (f) the Internet. Communications are conducted over network 135 by electronic or optical signals.

[0063] The computer 105 includes a processor 110 and a memory 115 connected to the processor 110. Although the computer 105 is depicted as a standalone device in Fig. 6, it is not limited to such a configuration. A distributed processing system may be employed in which the computer 105 is connected to other devices not shown, and the distributed processing may be performed in each device.

[0064] Processor 110 is an electronic device made up of logic circuits that responds to and carries out instructions.

[0065] The memory 115 is a tangible computer-readable storage medium on which a computer program is encoded. Specifically, the memory 115 stores data and instructions, i.e., program code, that can be read and executed by the processor 110 to control the operation of the processor 110. The memory 115 can be realized as a random access memory (RAM), a hard drive, a read-only memory (ROM), or a combination thereof. One component of the memory 115 is a program module 120.

[0066] The program modules 120 include instructions for controlling the processor 110 to perform the processes described in this disclosure. Although operations are described in this disclosure as being performed by the computer 105 or a method or process or sub-process thereof, those operations are actually performed by the processor 110.

[0067] The term "module" is used in this disclosure to refer to a functional entity that may be embodied as either a stand-alone component or an integrated entity consisting of multiple sub-components. Thus, the program module 120 may be realized as a module consisting of a single component or as a module consisting of multiple components operating in cooperation with each other. Furthermore, the program module 120 is installed in the memory 115 and therefore may be realized as software, but it may also be realized as hardware (e.g., electronic circuitry), firmware, or any combination thereof (software, hardware, firmware).

[0068] The program module 120 may be already loaded into memory 115 or may be configured to reside on storage device 140 for later loading into memory 115. Storage device 140 is a tangible computer-readable storage medium that stores the program module 120. Examples of storage device 140 include compact discs, magnetic tape, read-only memory, optical storage media, hard drives, memory units consisting of multiple parallel hard drives, and universal serial bus (USB) flash drives. Storage device 140 may also be random access memory or other types of electronic storage devices located in a remote storage system (not shown) and connected to computer 105 via network 135.

[0069] The system 100 further includes databases 150A and 150B communicatively coupled to the network 135. However, the number of data sources is arbitrary and may be one, three, or more. The data sources 150A and 150B may include unstructured data and may include social media.

[0070] System 100 further includes a user device 130 operated by a user and connected to computer 105 via network 135. Examples of user device 130 include input devices such as a keyboard or a voice recognition subsystem that allow the user to communicate information and command selections to processor 110. User device 130 may also include output devices such as a display device, a printer, or a voice synthesizer. In this case, a cursor control such as a mouse, trackball, or touch-sensitive screen allows the user to manipulate a cursor on a display device to communicate information and command selections to processor 110.

[0071] The processor 110 outputs the results 122 of the execution of the program modules 120 to the user device 130. The processor 110 may also provide the output to a storage device 125, such as a database or memory, or via a network 135 to a remote device not shown.

[0072] For example, a program that performs the operations of FIG.

[0073] The terms "comprising" or "comprising" should be interpreted as specifying the presence of the stated features, integers, steps or components, but not excluding the presence of one or more other features, integers, steps or components or groups thereof. The terms "a" and "an" are indefinite articles and therefore do not exclude embodiments having a plurality thereof.

[0074] (Other Embodiments) The scope of the present disclosure is not limited to the above-described embodiments, and various modifications can be made without departing from the spirit of the present disclosure. In other words, in the implementation stage, the components can be modified and embodied without departing from the spirit of the present disclosure.

[0075] Furthermore, various examples can be formed by appropriately combining multiple components disclosed in the above embodiments. For example, some components may be deleted from all the components shown in each embodiment and integrated with other components to form new components. Furthermore, components from different embodiments may be appropriately combined.

[0076] (Effects) The identification calculation device, method, and program for identifying facilities around an optical communication cable according to the present disclosure are believed to have the following advantages. First, it is possible to identify facilities (manholes, etc.) by capturing the characteristics of the natural vibration of the facilities (manholes, etc.) from the vibration distribution in the longitudinal direction of the optical cable. Second, it is possible to generate a series by dividing the vibration distribution data in the time direction and identify the vibration source in real time by repeating class determination.

[0077] 11: Peak frequency of Fourier spectrum 12: Initial value 13: Linear sum 14: Variance value 15: Average value 16: Identification dictionary 17: Identification function value 18: Filter 21: Optical cable 22: Ground 23: Pipe 24, 25, 26: Manhole 27: Vibration distribution measuring instrument 28: Vibration source 29: First vibration distribution data 31: Series (1) 32, 36: Distance-time-MH judgment plot 33: MH judgment result (correct judgment result) 34, 37: Incorrect judgment result 35: Series (1, 1, 1) 41: Light intensity distribution 42: Local section 43: Waveform pattern 44: Waveform pattern after Δt seconds 45: Change Δν 51: Waveform g(t) 52: Fourier spectrum 100: System 105: Computer 110: Processor 115: Memory 120: Program module 122: Result 125: Storage device 130: User device 135: Network 140: Storage device 150A, 150B: Data source 210: Second vibration distribution data 211: Third vibration distribution data 212: Memory unit 213: Feature extraction unit 214: Data reading unit 215: Time-frequency conversion unit 216: Peak frequency extraction unit 217: Peak frequency variance / average calculation unit 218: Discrimination dictionary 219: Discrimination calculation unit 220: Discrimination function value calculation unit 221: Sequence generation unit 222: Class determination unit 223: Result display unit 224: Discrimination function value calculation unit 225: Error calculation unit 226: Teacher signal 227: Squared error calculation unit 228: Weight vector update unit 229: Determination device 230: Vibration coordinate estimation unit

Claims

1. A determination device for classifying equipment environments in the longitudinal direction of an optical cable, comprising: a feature extraction unit that extracts feature values ​​for each position in the longitudinal direction of the optical cable from the vibration distribution in the longitudinal direction of the optical cable; a calculation unit that calculates a discrimination function value for identifying equipment around the optical cable based on the feature values ​​and a weight vector corresponding to the natural vibration of the equipment around the optical cable; and a determination unit that identifies the equipment based on the result of calculating the discrimination function value.

2. The determination device according to claim 1, further comprising an identification dictionary that holds a weight vector corrected based on an identification function value calculated from a linear sum of a first feature amount related to the vibration distribution of a first facility and a second feature amount related to the vibration distribution of a second facility different from the first facility, and a weight vector; the calculation unit calculates an identification function value for identifying the first facility and the second facility from a third feature amount related to the vibration distribution when the first facility and the second facility are mixed, and the weight vector held in the identification dictionary; and the determination unit identifies the first facility and the second facility based on the result of the calculation of the identification function value by the calculation unit.

3. The determination device according to claim 1, wherein the calculation unit calculates the discrimination function value at each predetermined time, and includes a sequence generation unit that replaces the discrimination function value with a corresponding teacher signal value and generates a sequence of the teacher signal in the time direction, and the determination unit identifies the equipment based on the length of the sequence.

4. The determination device according to claim 1, further comprising a vibration coordinate estimation unit that stores coordinate information of equipment to be identified and compares the identification results of the equipment with the coordinate information to estimate the position of the vibration source.

Images