Structural displacement estimation method and system

Abstract

A structure displacement estimation method is performed by a computer program running on a computing device, and the program is configured to cause a processor of the device to perform automated initial calibration that can collect measurements, respectively, from a radar and an accelerometer mounted directly at a structure, an optimal target among candidate targets detected by the radar is automatically determined, and a final conversion factor for converting a displacement in a line-of-sight direction to a displacement in an actual vibration direction for the optimal target is automatically calculated, structural displacement monitoring may be performed by fusing a radar-based displacement obtained via application of the final conversion factor to a phase extracted from a radar measurement of the optimal target with an accelerometer-based displacement obtained via double integration of the accelerometer measurement based on an FIR filter, therefore, the accuracy of the structural displacement is improved.

Description

Technical Field

[0001] Embodiments relate to a structural displacement estimation method, and more particularly, embodiments relate to a method of estimating structural displacement at a desired point by installing a radar and an accelerometer at the desired structural displacement measurement point and fusing the measured values ​​to estimate the displacement at the desired point. Background Art

[0002] Displacements are crucial to understanding the behavior and current condition of a structure, and are therefore directly considered as safety indicators of the structure in the structural design codes of many countries. Excessive displacements caused by external loads can be used as a primary indicator that the stability of a structure is at risk, and can also be used for maintenance, management, and repair, such as improving numerical models or estimating damage.

[0003] However, unlike physical quantities such as acceleration, displacement is difficult to measure in actual structures, so various methods can be used to measure displacement, such as linear variable differential transformer ("LVDT"), laser Doppler velocimeter ("LDV"), indirect methods, etc.

[0004] The most commonly used LVDT is designed to be proportional to the mechanical displacement of the core and the output voltage, resulting in high accuracy and reliability. However, since the LVDT requires a fixed point to mount the sensor, an additional temporary structure is required to measure the displacement of the actual structure, and it is difficult to measure the displacement of bridges built on rivers and oceans. LDV is a non-contact sensor that can measure the displacement of a structure using the phase difference of reflected laser light. When measuring the center of a bridge, it is difficult to find the sensor installation point because the laser must be shot perpendicular to the surface of the measurement point because the bottom of the measurement point is in the ocean, and the equipment is very expensive, so it is not suitable for measuring displacement at multiple points at the same time.

[0005] The indirect method is a method of converting a physical quantity, such as acceleration, into displacement. An accelerometer (a sensor that measures acceleration) is mounted on the structure, and the displacement can be easily estimated from the acceleration via double integration. Although it has the advantage of not requiring a fixed point for sensor mounting and is relatively inexpensive, it has the problem of greatly amplifying low-frequency components due to measurement noise. To solve this problem, finite impulse response ("FIR") filter methods have been proposed, however, these methods cannot clearly distinguish between structural displacement and noise in the actual low-frequency band, so a method of fusing multiple sensor data is proposed.

[0006] On the other hand, it is also known how to use radar sensors to measure the displacement of structures. In this method, the radar sensor transmits a frequency modulated signal, receives a reflected signal from an object, and estimates the displacement in the line of sight (“LOS”) direction based on the time delay between the transmitted and received signals. However, conventional radar-based displacement estimation methods have the following two problems. First, the radar sensor is mounted at a fixed point outside the structure to be measured and detects multiple targets on the structure. At this time, the position of each target detected by the radar sensor must be determined to select the best target that is most suitable for estimating the structural displacement, and an initial calibration is required to estimate the final conversion factor for converting the displacement in the line of sight (“LOS”) between the radar and the target to the displacement in the actual vibration direction. However, in the past, these tasks were not automated and were done manually by the user. Manual initial calibration is cumbersome and reduces the accuracy and speed of displacement estimation. Second, if the raw phase extracted from the radar measurement exceeds the range of [-π, π], inaccurate displacement estimation will occur due to phase wrapping. Phase wrapping can be a serious problem, especially when the structural displacement is larger than the wavelength of the radar. Summary of the invention [Technical issues]

[0007] The embodiment provides a structural displacement estimation method and a displacement estimation system thereof, which can improve the accuracy of displacement measurement by using information of an accelerometer and a radar sensor installed in a structure to be measured together.

[0008] The embodiments are not limited to the above-mentioned embodiments, and other embodiments that are not mentioned can be clearly understood by a person of ordinary skill in the art from the following description. [Technical solution]

[0009] In order to achieve the above-mentioned purpose, a structural displacement estimation method according to one embodiment is performed by a computer program running on a computing device, and the computer program can be configured to cause a processor of the computing device to perform an automated initial calibration step and a structural displacement monitoring step. The automated initial calibration step can be configured to collect measurement values ​​from a radar and an accelerometer directly installed at a displacement measurement point of the structure, respectively, automatically determine an optimal target among multiple candidate targets detected by the radar, and automatically calculate a final conversion factor for converting the displacement of the optimal target in the line of sight direction of the displacement measurement point into a displacement in the actual vibration direction. The structural displacement monitoring step can be configured to calculate the final displacement by fusing the radar-based displacement obtained by applying the final conversion factor to the phase extracted from the radar measurement value of the optimal target and the accelerometer-based displacement obtained by double integration of the measurement value from the accelerometer based on an FIR filter.

[0010] In one embodiment, the measurements using the radar and the measurements using the accelerometer may be performed during the same time period.

[0011] In one embodiment, the measurements using the radar and the measurements using the accelerometer may be performed in less than one minute.

[0012] In one embodiment, radars and accelerometers, which may be mounted in close proximity to each other at displacement measurement points on the structure, may collect measurements.

[0013] In one embodiment, the automated initial calibration step may include measuring an initial displacement in a line-of-sight direction of a displacement measurement point for each of a plurality of candidate targets using a radar, calculating a plurality of first displacements in a vibration direction by applying a plurality of conversion factor values ​​to the initial displacement for each of the plurality of candidate targets, measuring an acceleration of the displacement measurement point using an accelerometer, calculating a second displacement by doubly integrating the acceleration, calculating an RMSE between each of the plurality of first displacements calculated for each of the plurality of candidate targets and the second displacement, and determining a minimum value among the calculated RMSE values ​​as a minimum RMSE value of the candidate targets, automatically determining a candidate target having a minimum minimum RMSE value among the plurality of minimum RMSE values ​​determined for each of the plurality of candidate targets as an optimal target, and automatically calculating a conversion factor applied to obtain the minimum RMSE value of the optimal target as a final conversion factor of the optimal target.

[0014] In one embodiment, the plurality of conversion factor values ​​may be in the range of 0.5 to 2.0.

[0015] In one embodiment, in the structure displacement estimation method, the structure displacement monitoring step may be performed periodically by periodically performing displacement measurements using radar and accelerometer for the best target automatically determined in the automated initial calibration step.

[0016] In one embodiment, the structural displacement monitoring step may include extracting a raw phase by performing radar measurement for an optimal target using a radar, measuring the acceleration of a displacement measurement point using an accelerometer, and calculating an acceleration-based displacement by double integrating the measured acceleration, selecting an unwrapped phase close to a predicted phase using the measured acceleration when a phase wrapping problem occurs in the raw phase due to the displacement of the structural displacement measurement point being greater than a wavelength of a radar signal of the radar, calculating a third displacement in a line-of-sight direction using the raw phase without the phase wrapping problem or the unwrapped processed phase due to the phase wrapping problem, calculating a radar-based displacement in a vibration direction by applying a final conversion factor to the third displacement in the line-of-sight direction, and calculating a final displacement by fusing the acceleration-based displacement and the radar-based displacement using a finite impulse response (FIR) filter.

[0017] In one embodiment, the step of calculating the final displacement may include obtaining a radar-based low-frequency displacement by performing a low-pass filter on the radar-based displacement, obtaining an acceleration-based high-frequency displacement by performing a high-pass filter on the acceleration-based displacement, and calculating the final displacement by fusing the radar-based low-frequency displacement and the acceleration-based high-frequency displacement.

[0018] In one embodiment, the step of selecting the unwrapped phase may include using the displacement at the (k-1)th and (k-2)th time steps and the (k-1)th acceleration, by equation and equation To calculate the predicted displacement at the kth time step and predicted phase And through the equation The unwrapped phase that is within ±2pπ (where p is an integer) of the original phase and closest to the predicted phase is found and selected as the phase to be used for estimating the radar-based displacement.

[0019] In one embodiment, the final displacement can be calculated using the formula Calculate (C H : Double integration and (2M+1)th order high-pass filter, a: Measured acceleration vector, C L : (2M+1)th order low-pass filter, u: radar-based displacement vector).

[0020] In one embodiment, the radar may be a frequency modulated continuous wave (FMCW) millimeter wave radar, and the FMCW may transmit a chirp signal, receive signals reflected from a target candidate group, and use the signal round trip time between the transmitted and received signals to estimate the displacement in the line of sight direction.

[0021] A structural displacement estimation system according to another embodiment includes a radar directly installed at a displacement measurement point of the structure, the radar being configured to transmit radar signals toward a plurality of candidate targets that do not change position, and being configured to receive reflection signals reflected from the plurality of candidate targets; an accelerometer installed at the displacement measurement point of the structure and being configured to measure acceleration at the displacement measurement point of the structure; and a displacement estimator, the displacement estimator being configured to: perform an automated initial calibration function configured to collect measurement values ​​from the radar and the accelerometer, respectively, automatically determine an optimal target among the plurality of candidate targets detected by the radar, automatically calculate a final conversion factor for converting a displacement of the optimal target in a line-of-sight direction of the displacement measurement point into a displacement in an actual vibration direction, and perform a structural displacement monitoring function configured to calculate a final displacement by fusing, based on an FIR filter, a radar-based displacement obtained by applying the final conversion factor to a phase extracted from a measurement value from the radar of the optimal target and an accelerometer-based displacement obtained by double integration of a measurement value from the accelerometer.

[0022] In one embodiment, the displacement estimator may include a computer program written to perform an automated initial calibration function and a structure displacement monitoring function; and a processor that executes the computer program.

[0023] In one embodiment, when a phase wrapping problem occurs because the displacement of the displacement measurement point of the structure is greater than the wavelength of the radar signal, the displacement estimator may be further configured to perform a phase unwrapping processing function to estimate the radar-based displacement by selecting an unwrapped phase that is closest to the predicted phase using the acceleration measured at the structure.

[0024] In one embodiment, the automated initial calibration function may include: a function of measuring an initial displacement of each of a plurality of candidate targets in a line-of-sight direction of a displacement measurement point using a radar, a function of calculating a plurality of first displacements in a vibration direction by applying a plurality of conversion factor values ​​to the initial displacement for each of the plurality of candidate targets, a function of measuring an acceleration of a displacement measurement point using an accelerometer, a function of calculating a second displacement by doubly integrating the acceleration, a function of calculating an RMSE between each of a plurality of radar-based first displacements calculated for each of the plurality of candidate targets and the second displacement, and determining a minimum value among the calculated RMSE values ​​as a minimum RMSE value of the candidate targets, a function of automatically determining a candidate target having a minimum minimum RMSE value among a plurality of minimum RMSE values ​​determined for each of the plurality of candidate targets as an optimal target, and a function of automatically calculating a conversion factor applied to obtain a minimum RMSE value of the optimal target as a final conversion factor of the optimal target.

[0025] In one embodiment, the structural displacement monitoring function may include a function of extracting a raw phase by performing radar measurement for an optimal target using a radar, a function of measuring the acceleration of a displacement measurement point with an accelerometer and calculating an acceleration-based displacement by doubly integrating the measured acceleration, a function of selecting an unwrapped phase close to a predicted phase using the measured acceleration when a phase wrapping problem occurs in the raw phase due to the displacement of the displacement measurement point of the structure being greater than a wavelength of a radar signal, a function of calculating a third displacement in a line-of-sight direction using the raw phase without the phase wrapping problem or the unwrapped processed phase due to the phase wrapping problem, a function of calculating a radar-based displacement in a vibration direction by applying a final conversion factor to the third displacement in the line-of-sight direction, and a function of calculating a final displacement by fusing the acceleration-based displacement and the radar-based displacement by using a finite impulse response (FIR) filter.

[0026] In one embodiment, the function of calculating the final displacement may include the function of obtaining a radar-based low-frequency displacement by performing a low-pass filter on the radar-based displacement, the function of obtaining an acceleration-based high-frequency displacement by performing a high-pass filter on the acceleration-based displacement, and the function of calculating the final displacement by fusing the radar-based low-frequency displacement and the acceleration-based high-frequency displacement. [Beneficial Effects]

[0027] According to a structural displacement estimation method of an exemplary embodiment of the present invention, radar and accelerometers can be directly mounted on the structure, so there is no need to locate additional fixed points outside the structure; initial calibration including selecting the best target required for estimating radar-based displacement and estimating the final conversion factor can be performed automatically, thereby reducing time, human resources and costs.

[0028] Furthermore, the phase wrapping problem is solved using accelerometers, and the estimated displacement based on the unwrapped phase can be fused with the measured acceleration based on the FIR filter to more accurately calculate the final displacement, thereby allowing continuous structural displacement estimation monitoring.

[0029] However, the effects of the embodiments are not limited to the above-described effects and can be extended in various ways within the scope of the concept and the scope of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 An embodiment is schematically shown in which a structural displacement estimation system according to an embodiment of the present invention is installed in a bridge structure.

[0031] Figure 2 yes Figure 1 Magnified view of portion A.

[0032] Figure 3 (A) and (B) respectively show the Figure 1 1 and 1 are examples of signal waveforms of the radar 20 in the structural displacement estimation system, which are represented as time functions of frequency and amplitude, respectively.

[0033] Figure 4 It shows the use of Figure 1 A flow chart of a method for estimating structural displacement using a structural displacement estimation system.

[0034] Figure 5 A detailed description is included in Figure 4 Flowchart of the automated initial calibration step in the method for estimating structural displacements.

[0035] Figure 6 It is shown that the Figure 4 Flow chart of the structural displacement monitoring step in the method for estimating structural displacement.

[0036] Figure 7 It is shown that the application Figure 4 A view of the accelerometer 30 assisted phase unwrapping algorithm of the method for estimating structural displacement.

[0037] Figure 8 (A) and (B) are shown in Figure 4 A view of the steps of automatically determining the optimal target and automatically calculating the final conversion factor in a method of estimating the displacement of a structure.

[0038] Fig. 9 It is shown based on Figure 6 A view of the steps in which the FIR filters are fused to calculate the final displacement.

[0039] Fig.10 It is the configuration layout for short-range simulation.

[0040] Fig.11 is a diagram showing estimated displacement according to vibration amplitude.

[0041] Fig.12 is the result of calculating the final conversion factor and the error in the displacement estimated by applying the final conversion factor.

[0042] Fig.13 Is the configuration layout for remote simulation.

[0043] Fig.14 is a diagram showing an automated initial calibration step.

[0044] Fig.15 is the error between the calculation result of the unwrapping algorithm and the displacement estimated by applying the unwrapping algorithm.

[0045] Fig.16 It is used to simulate the configuration layout of the structure caused by pedestrians walking.

[0046] Fig.17 (A) is a view showing an automated initial calibration, and Fig.17 (B) is the result of calculating the estimated displacement error for each target.

[0047] Fig.18 is a diagram showing estimated displacement according to vibration amplitude.

[0048] Fig.19 It is shown Fig.18 (B) View of the original phase and the unwrapped phase. DETAILED DESCRIPTION

[0049] The preferred embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. In the accompanying drawings, the same reference numerals are used for the same components, and redundant descriptions of the same components will be omitted.

[0050] Figure 1 An embodiment is schematically shown in which a structural displacement estimation system according to an embodiment of the present invention is installed in a bridge structure.

[0051] refer to Figure 1 A structural displacement estimation system 1000 according to an embodiment of the present invention may include a radar (radio detection and ranging) 20 and an accelerometer 30 as measurement devices to measure the displacement of a structure 10, which is a target of the displacement measurement. In addition, the structural displacement estimation system 1000 may also include a displacement estimator 50 that receives measurement signals from the radar 20 and the accelerometer 30 and performs operations to estimate the displacement of the structure 10.

[0052] exist Figure 1 In the present invention, a bridge is shown as a structure 10 whose displacement is to be estimated, and there is no particular limitation on the type of structure to which the present invention can be applied.

[0053] The radar 20 may be directly installed at the displacement measurement point P of the structure 10 , transmit radar 20 signals toward a plurality of candidate targets t1 . . . tn whose positions do not change, and receive reflection signals reflected from the plurality of candidate targets.

[0054] The accelerometer 30 may be installed at the displacement measurement point P of the structure 10 and measure the acceleration at the displacement measurement point P of the structure 10. In other words, the radar 20 and the accelerometer 30 may be installed at the displacement measurement point P of the structure 10 in close proximity to each other so that each of the radar 20 and the accelerometer 30 may collect measurement values.

[0055] The displacement estimator 50 may be communicatively connected to the radar 20 and the accelerometer 30 by wired, wireless, or both. The displacement estimator 50 may be configured to calculate an estimate of the displacement of the structure 10 by providing radar measurement data provided from the radar 20 (obtained by the radar processing a transmission signal reflected from the structure 10) and acceleration data of the structure 10 provided by the accelerometer 30 via communication.

[0056] The displacement estimator 50 may include a computer program written to perform an automated initial calibration function (S100) and a structural displacement monitoring function (S200); and a processor 52 that executes the computer program. In addition, when a phase wrapping problem occurs because the displacement of the displacement measurement point of the structure is greater than the wavelength of the radar 20 signal, the computer program of the displacement estimator 50 may also be configured to perform a phase unwrapping processing function to estimate the radar-based displacement by selecting an unwrapped phase closest to the predicted phase using the measured acceleration at the structure 10.

[0057] The hardware resources for the displacement estimator 50 may include a computing device including a processor 52. In addition to the processor 52, the computing device may include a memory 54, a data storage 136 of a non-volatile storage device, an input / output device 58, etc. For example, the hardware of the displacement estimator 50 may include a general-purpose computer (which includes the above-mentioned devices) or a computing device, a workstation device, etc. dedicated to the present invention.

[0058] The processor 52 may execute an automated initial calibration function configured to automatically determine an optimal target ( t1 . . . tn ) among a plurality of candidate targets t1 . . . tn detected by the radar 20. Figure 5 A), automatically calculating a final conversion factor B for converting the displacement in the line-of-sight direction of the displacement measurement point D for the best target A to the displacement in the actual vibration direction D; and also performing a structural displacement monitoring function, which is configured to calculate the final displacement by fusing, based on an FIR filter, a radar-based displacement obtained by applying the final conversion factor B to a phase extracted from the measurement value of the best target A from the radar 20 and an accelerometer-based displacement obtained by double integration of the measurement value from the accelerometer 30.

[0059] The automated initial calibration function may include the following functions: a function of measuring an initial displacement of each of a plurality of candidate targets in a line-of-sight direction of a displacement measurement point using the radar 20, a function of calculating a plurality of first displacements in a vibration direction by applying a plurality of conversion factor values ​​to the initial displacement of each of the plurality of candidate targets, a function of measuring an acceleration of a displacement measurement point using the accelerometer 30, a function of calculating a second displacement by doubly integrating the acceleration, a function of calculating an RMSE between each of the plurality of radar-based first displacements and the second displacement calculated for each of the plurality of candidate targets, and determining a minimum value among the calculated RMSE values ​​as a minimum RMSE value of the candidate targets, a function of automatically determining a candidate target having a minimum minimum RMSE value among a plurality of minimum RMSE values ​​determined for each of the plurality of candidate targets as an optimal target A, and a function of automatically calculating a conversion factor applied to obtain a minimum RMSE value of the optimal target A as a final conversion factor B of the optimal target A.

[0060] The structural displacement monitoring function may include the following functions: a function of extracting an original phase by performing radar 20 measurement of an optimal target A using the radar 20, a function of measuring the acceleration of a displacement measurement point using the accelerometer 30 and calculating an acceleration-based displacement by doubly integrating the measured acceleration, a function of selecting an unwrapped phase close to a predicted phase using the measured acceleration when a phase wrapping problem occurs in the original phase due to the displacement of the displacement measurement point of the structure 10 being greater than the wavelength of the radar 20 signal of the radar 20, a function of calculating a third displacement in a line-of-sight direction using the original phase without the phase wrapping problem or the unwrapped processed phase due to the phase wrapping problem, a function of calculating a radar-based displacement in a vibration direction by applying a final conversion factor B to the third displacement in the line-of-sight direction, and a function of calculating a final displacement by fusing the acceleration-based displacement and the radar-based displacement by using a finite impulse response (FIR) filter.

[0061] On the other hand, the function of selecting the unwrapping phase can be a function of obtaining a radar-based low-frequency displacement by performing a low-pass filter on the radar-based displacement, a function of obtaining an acceleration-based high-frequency displacement by performing a high-pass filter on the acceleration-based displacement, and a function of calculating a final displacement by fusing the radar-based low-frequency displacement and the acceleration-based high-frequency displacement.

[0062] The following will refer to Figures 4 to 9 The automatic initial calibration function, structural displacement monitoring function and phase unwrapping processing function are described in detail.

[0063] Figure 2 yes Figure 1 An enlarged view of part A of Figure 3(A) and (B) respectively show the Figure 1 1 and 1 are examples of signal waveforms of the radar 20 in the structural displacement estimation system, which are represented as time functions of frequency and amplitude, respectively.

[0064] refer to Figure 2 , the distance D between the radar 20 and any target can be expressed as follows:

[0065] [Equation 1]

[0066] In Equation 1, c represents the speed of light, and Δt represents the signal round trip time. Figure 3 (B) and Equation 2 and Equation 3 describe Δt in detail.

[0067] In an exemplary embodiment, the radar 20 may be an FMCW radar 20. In the structural displacement estimation system 1000, the FMCW radar 20 installed in the structure 10 may transmit a modulated frequency signal in the millimeter wave band, receive a signal reflected from a target (object), convert it into a digital signal through predetermined signal processing, and provide it to the displacement estimator 50. The displacement estimator 50 may process the digital signal, calculate the phase, and convert it into a displacement. That is, according to the time delay between the transmitted signal and the received signal, the displacement in the line of sight ("LOS") direction D may be estimated. At the same time, the accelerometer 30 may convert the acceleration signal into a displacement through double integration. Then, the displacement estimator 50 may fuse these displacements to estimate the final displacement of the structure 10.

[0068] The modulated frequency signal transmitted from the FMCW radar 20 may be reflected from multiple candidate targets t1 . . . tn, and an initial calibration process may be required to determine the best target A for estimating radar-based displacement.

[0069] The plurality of candidate targets t1 . . . tn may include targets that are fixed and do not change position. For example, objects such as large rocks, piers of bridges, etc. may be selected as targets.

[0070] The distance between the FMCW radar 20 and the target may include displacement information in the line of sight direction D of the displacement measurement point. However, due to external forces, the displacement (vibration) direction of the structure 10 may be arbitrary, so that the line of sight direction D and the actual vibration direction may be different. A conversion factor may be required to convert the displacement in the line of sight direction D of the structure 10 into the actual (vibration) displacement u.

[0071] The following will refer to Figure 5 The initial calibration steps of selecting the best target A and estimating the final conversion factor B are described in detail.

[0072] refer to Figure 3 (A), the FMCW radar 20 may use a chirp signal. For example, the chirp signal may have a waveform whose frequency increases as time increases.

[0073] refer to Figure 3 (B), the FMCW radar 20 can transmit a modulated frequency signal 21 and receive a reflected signal 22 from a target. The transmitted signal (T(t), 21) corresponds to a solid line, and the reflected signal (R(t), 22) corresponds to a dotted line.

[0074] The reflected signal 22 may be a delayed version of the transmitted signal 21, such as Figure 3 The transmission signal 21 and the reflection signal 22 for a single target can be expressed as follows:

[0075] [Equation 2]

[0076] [Equation 3]

[0077] In Equation 2 and Equation 3, f s , K, T c and δ represent the starting frequency (Fs), slope (K), chirp duration (Tc) and amplitude attenuation factor of the chirp signal, respectively.

[0078] The signal round trip time Δt required for calculating the displacement in the line-of-sight direction D can be extracted from the intermodulation frequency signal (intermodulation frequency (IF), w) or the phase (φ).

[0079] The signal round trip time Δt required for calculating the displacement in the line of sight direction D can be extracted from the intermodulation frequency (IF, w) or the phase (φ).

[0080] The intermodulation frequency signal may be generated by a mixer combining the transmission signal 21 and the reflection signal 22. The intermodulation frequency signal may be represented as follows:

[0081] [Equation 4]

[0082] In Equation 4, R * (t) represents the conjugate complex number of R(t).

[0083] By combining Equation 1 and Equation 4, the displacement of a single target in the line of sight direction of the displacement D measurement point can be estimated as follows:

[0084] [Equation 5]

[0085] like Figure 1 As shown, the radar 20 can detect a plurality of candidate targets t1 ... tn. Therefore, the displacement in the line of sight direction D of the displacement measurement point can be estimated to several values.

[0086] The reflection signal 22 and the intermodulation frequency signal used to estimate the displacement in the line-of-sight direction D of the displacement measurement point can be expressed as follows.

[0087] [Equation 6]

[0088] [Equation 7]

[0089] In Equation 6 and Equation 7, Q, m, δ m and Δt m They respectively represent the total number of targets included in the plurality of candidate targets t1 . . . tn, the target index, the amplitude attenuation factor of the mth target, and the signal round trip time Δt between the transmitted signal 21 and the signal reflected from the mth target.

[0090] The distances D between multiple candidate targets t1...tn and the radar 20 can be converted into IF tones, respectively. Referring to Equation 1, the distance D between the radar 20 and the target can be proportional to the signal round trip time Δt. Therefore, the signal round trip time Δt can be delayed in proportion to the distance, and an IF consisting of as many tones as the time delay can be obtained. It can be processed using a fast Fourier transform ("FFT") to distinguish the IF consisting of multiple tones. After processing the FFT, a spectrum can be generated in which single peaks of different tones appear, and each peak can represent a target at a certain distance.

[0091] In summary, the relationship between the displacement and the frequency or phase in the line of sight direction D of the displacement measurement point can be expressed as follows:

[0092] [Equation 8]

[0093] Reference below Figures 4 to 9 , describes in detail a structural displacement estimation method 2000 that can be executed by a computer program running on a computing device.

[0094] Figure 4 It shows the use of Figure 1 A flow chart of a method for estimating structural displacement using a structural displacement estimation system.

[0095] refer to Figure 4The computer program may cause the processor 52 of the computing device to execute the structural displacement estimation method 2000, which includes an automated initial calibration step (S100) and a structural displacement estimation step (S200).

[0096] The automated initial calibration step (S100) can collect measurement values ​​from the radar 20 and the accelerometer 30 respectively installed at the displacement measurement point of the structure 10, and automatically determine an optimal target A among multiple candidate targets t1...tn detected by the radar 20, and automatically calculate the final conversion factor B used to convert the displacement of the optimal target A in the line of sight direction of the displacement measurement point into the displacement in the actual vibration direction u.

[0097] The automated initial calibration step ( S100 ) may be performed by initially collecting measurements using the radar 20 and the accelerometer 30 within a predetermined time (eg, a short time of less than one minute), automatically selecting the best target, and automatically calculating the final conversion factor B.

[0098] The radar 20 can automatically detect multiple targets. Conventionally, the problem of large structural displacement error estimation may be caused by selecting the wrong target. However, the structural displacement estimation method 2000 according to an embodiment of the present invention can use the most suitable optimal target A among multiple candidate targets t1...tn, which can be automatically determined by radar-based displacement.

[0099] In addition, since the distance between the radar 20 and the optimal target A is the displacement in the line of sight direction D of the displacement measurement point, it is necessary to convert the displacement in the line of sight direction D into the displacement in the actual vibration direction u that causes the displacement of the structure 10. Traditionally, the conversion factor is manually calculated through geometric calculation, however, the structural displacement estimation method 2000 according to one embodiment of the present invention can automatically calculate the final conversion factor B in the automated initial calibration step (S100).

[0100] By periodically performing displacement measurements using the radar 20 and the accelerometer 30 in the automated initial calibration step ( S100 ) to automatically determine the optimal target A, the structural displacement monitoring step ( S200 ) may be performed periodically.

[0101] The structural displacement monitoring step (S200) may be performed to calculate a final displacement by fusing the following two displacements based on an FIR filter, namely, a radar-based displacement obtained by applying a final conversion factor B to a phase extracted from radar 20 measurements from the best target A and an accelerometer-based displacement obtained by double integrating measurements from the accelerometer 30.

[0102] The structural displacement monitoring step (S200) can obtain the displacement in the line of sight direction D at the phase extracted from the radar 20 measurement value of the automatically selected optimal target A, and then the final conversion factor B can be applied to the displacement to obtain the radar-based displacement in the vibration direction of the measurement point P, and the acceleration-based displacement can be obtained by double integration of the measurement value collected by the accelerometer 30. Then, by fusing the radar-based low-frequency displacement that performs low-pass filtering on the radar-based displacement and the acceleration-based high-frequency displacement that performs high-pass filtering on the acceleration-based displacement based on the FIR filter, the final displacement with higher accuracy than the displacement estimated using only the radar 20 or the accelerometer 30 can be calculated.

[0103] Figure 5 A detailed description is included in Figure 4 Flowchart of the automated initial calibration step in the method for estimating structural displacements.

[0104] refer to Figure 5 , the automated initial calibration (S100) may include each of the radar 20 measurement and the accelerometer 30 measurement and the step of calculating a plurality of first displacements by applying the conversion factor β value to the measurement value of the radar 20 (S110, S115), the step of calculating the second displacement by double integrating the measurement value of the accelerometer 30 (S122, S124), and the step of automatically determining the optimal target A and automatically calculating the conversion factor (S130).

[0105] First, steps ( S110 , S115 ) may be performed to each of causing the radar 20 to measure and causing the accelerometer 30 to measure.

[0106] The initial displacement of each of the plurality of candidate targets t1 . . . tn on the line of sight of the displacement measurement point P may be measured using the radar 20 ( S110 ), and the acceleration of the displacement measurement point P may be measured using the accelerometer 30 ( S115 ).

[0107] There may be a plurality of targets within the detection range of the radar 20. Since the accuracy of the measurements provided by the respective targets may vary, it is preferable to select the best target A that guarantees the most accurate measurement. The displacement value of the measurement point measured by the accelerometer 30 may be a criterion for selecting the best target A. In other words, for each of the plurality of candidate targets, the displacement may be calculated using the radar 20 and the accelerometer 30, and the candidate target having the smallest difference between the two calculated displacements may be selected as the best target A. The signal processing process thereof will be described in detail.

[0108] During a predetermined measurement time, the radar 20 may detect a plurality of target candidates t1 ... tn, select any one target (e.g., target 1), and perform measurement on the selected target by the radar 20 (S110). During the same measurement time in which the radar 20 transmits a radar 20 signal to the target 1 and receives a reflected signal to measure a phase, the accelerometer 30 may measure acceleration at the same measurement point during the same measurement time (S115). The displacement measurement using the radar 20 and the accelerometer 30 may be performed independently. As previously described with reference to Figure 1 As described, the radar 20 and the accelerometer 30 may be mounted close to each other at the displacement measurement points of the structure 10 and may collect measurements.

[0109] A plurality of initial conditions required for the automatic initial calibration step (S100) can be set. For example, the conversion factor ( Figure 8 The initial value of the conversion factor β) is set to a predetermined size. The variable range of the conversion factor β can also be set. In one embodiment, the conversion factor β can be set to a value in the range of 0.5 to 2. When the conversion factor β is 1, the accuracy of estimating the distance between the radar 20 and the target can be maximized. If the conversion factor β is outside the range of 0.5 to 2, the strength of the reflected signal of the radar 20 may be weakened, and the accuracy of the estimation may be reduced, so it is set to a value in the above range.

[0110] The time for performing one cycle of radar 20 measurements and acceleration measurements can also be set. Measurements using radar 20 and measurements using accelerometer 30 can be performed simultaneously. In one embodiment, measurements using radar 20 and measurements using accelerometer 30 can be performed in less than one minute. Up to one minute of measurement is sufficient to collect data to perform initial calibration.

[0111] Next, the step of applying the conversion factor β value to the measurement value of the radar 20 to calculate a plurality of first displacements, and calculating the second displacement by doubly integrating the measurement value of the accelerometer 30 may include the step of applying the plurality of conversion factor β values ​​to the initial displacement to calculate a plurality of first displacements in the vibration direction (S122) and calculating the second displacement by doubly integrating the acceleration (S124).

[0112] The radar 20 may be used to measure the displacement of each of the plurality of candidate targets t1 ... tn in the line of sight direction D of the displacement measurement point P. The displacement in the line of sight direction D may be defined as an initial displacement. Thereafter, a plurality of first displacements may be calculated by applying a plurality of transformation coefficient factors β to the initial displacement for each of the plurality of candidate targets t1 ... tn. The first displacement may be defined as a displacement u in the vibration direction.

[0113] On the other hand, the accelerometer 30 measurement can be calculated by double integration of the second displacement of the measuring point P.

[0114] Finally, the step of automatically determining the optimal target A and automatically calculating the conversion factor (S130) may include calculating the RMSE between each of the multiple first displacements calculated for each of the multiple candidate targets t1...tn and the second displacement, determining the minimum value among the calculated RMSE values ​​as the minimum RMSE value of the candidate targets (S132), automatically determining the candidate target having the smallest minimum RMSE value among the multiple minimum RMSE values ​​determined for each of the multiple candidate targets as the optimal target A (S134), and automatically calculating the conversion factor applied to obtain the minimum RSME value of the optimal target A as the final conversion factor B of the optimal target (S136).

[0115] High-pass filtering can be performed on the first and second displacements to obtain the first high-frequency displacement and the second high-frequency displacement. When the acceleration of the structure 10 is measured and integrated in the time domain to calculate the second displacement, the uncertainty of the initial velocity of the actual structure 10, the noise of the measurement signal are accumulated during the time integration process, and low-frequency drift occurs, and the calculated second displacement is offset. Therefore, high-pass filtering can be performed to offset the influence of these low-frequency drifts. During the high-pass filtering, the cutoff frequency can be set to about 0.5 Hz.

[0116] By calculating the RMSE between the high-pass filtered first and second high-frequency displacements (S132), the optimal target A can be automatically determined (S134), and the final conversion factor B can also be automatically calculated (S136). Figure 8 The steps of automatically determining the optimal target A and automatically calculating the final conversion factor B by calculating the RMSE (S132, S134, S136) are described in detail.

[0117] Figure 8 The (A) and (B) views show Figure 4 The steps of automatically determining the best target and automatically calculating the final conversion factor in the method for estimating structural displacements.

[0118] refer to Figure 8 (A), the RMSE between each of the first and second displacements may be calculated by gradually increasing the conversion factor β (S132).

[0119] For each of the plurality of candidate targets, RMSE between a first displacement calculated by applying a plurality of conversion factor β values ​​to an initial displacement measured using the radar 20 and a second displacement calculated by doubly integrating the measured value of the acceleration 30 may be calculated, and correlation between the RMSE and the conversion factor β may be obtained. The plurality of conversion factor values ​​may be selected within a range of 0.5 to 2.0.

[0120] The error function ("RMSE") can be defined as follows:

[0121] [Equation 9]

[0122] Specifically, the RMSE between the above-mentioned displacement based on the radar 20 and the displacement based on the accelerometer 30 may be calculated for a randomly selected target 1 .

[0123] In this case, the RMSE calculation may be repeatedly performed by gradually increasing the conversion factor β from an initial value of 0.5 to 2. In this way, the RMSE value for the target 1 as the conversion factor β value increases may be obtained.

[0124] like Figure 8 As shown in (A), a graph of RMSE values ​​as the conversion factor β values ​​increase can be plotted, and the conversion factor value E corresponding to the minimum RMSE value can be selected.

[0125] The above operation 1 for finding the conversion factor β value corresponding to the minimum RMSE E value for target 1t1 can be performed for multiple target candidates t1...tn.

[0126] Through this process, a minimum RMSE E value equal to the total number of targets detected by the radar 20 can be obtained.

[0127] refer to Figure 8 (B), the optimal target A can be automatically determined according to the minimum RMSE E value obtained for each target, and the final conversion factor B can be automatically calculated.

[0128] like Figure 8 As shown in (B), by comparing the minimum RMSE E values ​​corresponding to each of the total objectives, the objective corresponding to the smallest minimum RMSE E value can be automatically determined as the optimal objective ( Figure 5 A, Figure 8 (B) the best target) (S134).

[0129] Furthermore, when the optimal target A is selected, the conversion factor β corresponding to the minimum RMSE E value of the selected optimal target A may be automatically calculated as the final conversion factor B to be obtained ( S136 ).

[0130] Through this process, the optimal target A most suitable for estimating the structural displacement based on the radar 20 can be automatically selected through calculation processing performed by the program without human intervention, and the final conversion factor B can be automatically estimated to convert the displacement in the line of sight direction D of the structure 10 to the displacement in the actual vibration direction u of the structure 10.

[0131] Figure 6 The flowchart shows the Figure 4 The structural displacement monitoring step in the method for estimating structural displacement.

[0132] refer to Figure 6 , the structural displacement monitoring (S200) can first extract the original phase (S210) by making the radar measure the optimal target A. If a phase wrapping problem occurs in the extracted original phase, the unwrapped phase obtained by the phase unwrapping process can be selected as the actual phase (S212) by an auxiliary unwrapping algorithm using the acceleration 30 of the acceleration information. If there is no phase wrapping problem, the original phase can be used, and if a phase wrapping problem occurs, the unwrapped phase can be used to calculate the third displacement in the line of sight direction (S214), and the final conversion factor B obtained in the automated initial calibration step (S100) can be applied to the calculated third displacement to obtain the radar-based displacement in the vibration direction of the displacement measurement point (S216).

[0133] The following will refer to Figure 7 Describe the unwrapping algorithm in detail.

[0134] After calculating the acceleration-based displacement by double integration of the acceleration measured by the accelerometer 30 (S220), the final displacement may be calculated by fusing the radar-based displacement and the acceleration-based displacement based on an FIR filter (S230). Fig. 9 The final displacement calibration step (S230) is described in detail.

[0135] The structural displacement monitoring (S200) can estimate the structural displacement of the structure 10 at fixed intervals to monitor its stability, or if an immediate response is required, estimate the structural displacement in real time. At this point, the automatic selection of the best target A and its final conversion factor B in the automated initial calibration step (S100) can continue. In other words, the best target A can be selected in the automated initial calibration step (S100), the final conversion factor B for the target can be obtained, and then the structural displacement monitoring step can be performed to calculate the displacement of the structure 10 (S200).

[0136] Figure 7 It is shown that the application Figure 4A view of the accelerometer 30 assisted phase unwrapping algorithm of the method for estimating structural displacement.

[0137] like Figure 1 As shown, the radar 20 transmits a frequency modulated signal to a target and then receives a reflected signal from the target. The displacement estimator 50 may perform signal processing on the reflected signal and extract the original phase.

[0138] refer to Figure 7 , no matter how much the distance between the radar 20 and the target increases, the phase change corresponding to the distance change Only values ​​between -π and π may be output. Therefore, the original phase extracted by signal processing may not accurately estimate the distance between the radar 20 and the target, resulting in a large structural displacement estimation error.

[0139] Specifically, when the displacement of the displacement measurement point P of the structure 10 is greater than the wavelength of the radar 20 signal, a phase wrapping problem may occur. For example, the phase may gradually increase from zero, suddenly decrease to -π when it passes through π, and then increase toward π again, resulting in phase wrapping, and phase unwrapping is required to obtain meaningful information from the wrapped phase.

[0140] In the illustrated embodiment, an unwrapping algorithm based on acceleration 30 can be applied to the raw phase measured from radar 20 to further utilize the acceleration information. In this way, when the raw phase is an unwrapped phase, it can be restored to the actual phase.

[0141] refer to Figure 7 , the steps of selecting the unwrapping phase will be described in detail below.

[0142] First, at the kth time stage, the predicted displacement (u k ) and predicted phase It can be expressed as follows:

[0143] [Equation 10]

[0144] [Equation 11]

[0145] In Equation 10 and Equation 11, u k-1 、u k-2 and a k-1 are the displacements at the k-1th and k-2th time stages and the acceleration at the k-1th time phase, respectively, Δt is the signal round trip time, and f s , β and c represent the starting frequency, conversion factor and speed of light of the chirp signal respectively.

[0146] Next, the following circle function can be used to select the closest predicted phase Existing in the original phase The unwrapped phase in the range of ±2pπ

[0147] [Equation 12]

[0148] Finally, the displacement (u k ).

[0149] [Equation 13]

[0150] In Equation 13, Indicates the initial phase.

[0151] Using the acceleration information as described above, even if a structural displacement greater than the wavelength of the radar 20 occurs due to vibration, the displacement of the structure 10 can be accurately estimated by selecting the unwrapping phase.

[0152] Fig. 9 The view shows the Figure 6 The FIR filters are fused to calculate the final displacement.

[0153] refer to Fig. 9 , the step of calculating the final displacement (S230) may include: performing low-pass filtering (S214, S232) after obtaining the displacement based on the radar 20, performing high-pass filtering (S222, S234) after obtaining the displacement based on the acceleration, and calculating the final displacement (S236) by fusing the low-frequency displacement based on the radar 20 and the high-frequency displacement based on the acceleration.

[0154] As mentioned above Figure 7 As described above, the displacement based on the radar 20 and the displacement based on the accelerometer 30 may include a noise error component. A high-pass filter may be performed to eliminate low-frequency noise in the displacement based on the accelerometer 30 (S234). In this process, the low-frequency energy of the actual displacement data cannot be recovered, and in order to compensate for this, the displacement may be estimated by fusing the low-frequency displacement of the measured value based on the radar 20.

[0155] In the illustrated embodiment, a FIR filter can be used for such noise elimination. In the FIR filter, the displacement based on the radar 20 can be obtained by applying the final conversion factor B, and the low-frequency displacement based on the radar 20 can be obtained by low-pass filtering (S216). In addition, after double integration of the acceleration measurement (S222), the high-frequency displacement based on acceleration can be obtained by high-pass filtering (S234). By fusing the high-frequency displacement based on acceleration and the low-frequency displacement based on the radar (20) thus obtained, the final displacement with improved accuracy can be calculated (S236).

[0156] The minimization function can be used to combine two physical quantities.

[0157] [Equation 14]

[0158] In Equation 14, L a and L c denote the weighted diagonal matrix and the differential operator, respectively, u, u, a denote the vector representation of the estimated displacement, radar-based displacement, and acceleration-based displacement, and Δt denotes the signal round-trip time. The normalization constant λ is defined as 46.81(2N+1) -1.95 .

[0159] In summary, the final displacement It can be expressed as follows:

[0160] [Equation 15] u * =(Δt) 2 (L T L+λ 2 I) -1 L T L a a+λ 2 (L T L+λ 2 I) -1 u

[0161] [Equation 16]

[0162] In Equation 15 and Equation 16, L = L a L c , C H is as {(Δt) 2 (L T L+λ 2 I) -1 L T L a The filter factor applied to the acceleration measurements in row M+1 of}, C L is as {λ 2 (LT L+λ 2 I) -1} is the filter factor applied to the radar 20 measurements in the M+1th row, a is the measured acceleration vector, and u is the displacement vector based on the radar 20. By sacrificing only a small amount of time delay MΔt, we can more accurately estimate the displacement at the kth time step.

[0163] Figures 10 to 19 The actual experimental verification results of the structural displacement estimation method according to the embodiment of the present invention and the system using the method are shown.

[0164] The actual test measured the structural displacement of the bridge, which was caused by short-range, long-range and pedestrian-induced large vibrations. Based on the actual displacement (reference, solid line), the estimated structural displacement according to the embodiment of the present invention (hereinafter referred to as "(the present invention) proposed technology displacement") and the structural displacement estimated according to the conventional method (hereinafter referred to as "conventional technology displacement") were calculated together. The conventional technology displacement is estimated after randomly selecting a target and obtaining the conversion factor β2 using the geometric relationship between the radar 20 and the target.

[0165] Fig.10 It is the configuration layout for short-range simulation. Fig.11 shows the displacement estimated from the vibration amplitude, and Fig.12 is the result of calculating the final conversion factor and the error in the displacement estimated by applying the final conversion factor.

[0166] refer to Fig.10 , the radar 20 and the accelerometer 30 are directly mounted on the structure 10, and the LDV is mounted on the ground to measure the structural displacement. The structure 10 is designed to vibrate horizontally by a shaker.

[0167] refer to Fig.11 , when the displacement of the displacement measurement point P of the structure 10 is greater than the wavelength of the radar 20 signal, a phase wrapping problem may occur. Fig.11 As shown in (B) and (C), in the bridge that induced large vibration (1 Hz sine wave) and actual vibration, the simulation results show that the conventional technology displacement (single-point dot-dash line) shows phase wrapping and diverges more than the actual displacement (solid line).

[0168] On the other hand, Fig.11 As shown in (A) to (C), it can be seen that the proposed technical displacement (double-dotted line) is almost the same as the actual displacement (solid line).

[0169] refer to Fig.12In (A), it can be seen that the RMSE error of the proposed technology shift (upward bias) is reduced by 38% compared to the conventional technology shift (downward bias). The time delay spent on the proposed technology shift is estimated to be only 0.5 seconds.

[0170] Fig.13 is the configuration layout for remote simulation, Fig.14 is a diagram showing an automated initial calibration step, and Fig.15 is the error between the calculation result of the unwrapping algorithm and the displacement estimated by applying the unwrapping algorithm.

[0171] refer to Fig.13 , the radar 20 and the accelerometer 30 are directly mounted on the structure 10, and the LDS is mounted on the ground to measure the structural displacement. The structure 10 is designed to vibrate horizontally by a shaker.

[0172] refer to Fig.14 , the best target with the smallest RMSE E value of 0.457 is automatically determined as the best target A, and the E value of the target 0.457 is automatically estimated as the final conversion factor B. (Since it is written only to three decimal places, it seems that there are targets with the same value in the view, however, if you can see more than four decimal places, you will see that the values ​​are different from each other.)

[0173] refer to Fig.15 (A), when the displacement of the displacement measurement point P of the structure 10 is greater than the wavelength of the radar 20 signal, a phase wrapping problem may occur. Therefore, the conventional technology displacement (downward displacement) can only estimate the structural displacement close to the actual displacement under weak vibration of about 0.3 Hz.

[0174] On the other hand, Fig.15 As shown in (A) and (B), the proposed displacement (upward spurious) can also reduce the displacement estimation error by unwrapping the wrapped phase using the auxiliary unwrapping algorithm of the accelerometer 30. Fig.15 As shown in (B), the final displacement calculated based on the FIR filter shows an additional 11% reduction in RMSE error.

[0175] Fig.16 It is used to simulate the configuration layout of the structure caused by pedestrians walking. Fig.17 (A) shows the automated initial calibration, Fig.17 (B) is the result of calculating the estimated displacement error for each target. Fig.18 shows the estimated displacement according to the vibration amplitude, and Fig.19 Shows Fig.18 (B) The original phase and unwrapped phase.

[0176] refer to Fig.16 , the radar 20 and the accelerometer 30 are directly mounted on the structure 10, and the LDV is mounted on the ground to measure the structural displacement.

[0177] refer to Fig.17 In (A), the best target is automatically determined to be the target located at a distance of 7m, rather than the target with the largest amplitude (target 1). Fig.17 In (B), the RMSE error of the structural displacement estimated by target 1 with the largest amplitude is as high as 0.208 mm, while the maximum RMSE error of the technology proposed in the present invention is only 0.030 mm. In other words, it can be confirmed that the target with the largest amplitude and the best target are irrelevant.

[0178] Fig.18 (A) is the simulation result of 14 people walking slowly on the bridge. Fig.18 (B) is the simulation result of 14 people walking slowly on the bridge and 5 people running at the 1 / 2 point of the bridge, and Fig.18 (C) is the simulation result of 6 people running at 1 / 2 of the bridge.

[0179] like Fig.18 As shown in (A) to (C), it can be seen that as the vibration increases, the RMSE error of the conventional technology displacement increases and the accuracy of the displacement estimation decreases. On the other hand, the maximum RMSE error of the proposed technology displacement is only 0.026mm, which confirms that the displacement can be estimated more accurately compared with the conventional technology.

[0180] refer to Fig.19 , the original phase (dashed line) is shown to have a phase wrapping problem after 35 seconds, however, after obtaining the unwrapped phase (solid line) by applying the unwrapping algorithm, the proposed technique displacement can be estimated based on the radar 20. Therefore, the proposed technique displacement can estimate the structural displacement more accurately than the conventional technique displacement that does not consider the phase wrapping problem.

[0181] The structural displacement estimation method according to the above-mentioned embodiment of the present invention can be software implemented in the form of program instructions, which can be performed by various computer devices. The software may include computer programs, codes, instructions or one or more combinations thereof, and the processing device may be configured to perform as required, or the processing device may be commanded independently or collectively. The program command may be recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., which may be combined individually or in combination. The program command recorded in the medium may be specifically designed and constructed for the embodiment, or it may be obtained by a computer software technician. Examples of computer-readable recording media may include magnetic media, such as hard disks, floppy disks and tapes, optical media, such as CD-ROMs and DVDs, magneto-optical media, such as optical magnetic floppy disks, and hardware devices, which are specifically configured to store and execute program instructions, such as ROMs, RAMs, flash memories, etc.

[0182] The computer device for implementing the structural displacement estimation method 1000 and the structural displacement estimation method 2000 according to the embodiment can be implemented using one or more general-purpose computers or special-purpose computers, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. [Industrial Applicability]

[0183] The present invention relates to a structural displacement estimation method and a displacement estimation system thereof, and can use measurement information of an accelerometer and a radar sensor installed in a structure to be measured together to improve the accuracy of displacement measurement.

[0184] Although the present invention has been described in detail above, the right scope of the present invention is not limited thereto, and various modifications and improvements to the basic concept of the present invention defined in the following claims also fall within the right scope of the present invention.

Claims

1. A method for estimating structural displacements executed by a computer program running on a computing device, the computer program being configured to cause a processor of the computing device to execute the method, the method comprising: an automated initial calibration step configured to collect measurements from a radar and an accelerometer, respectively, mounted directly at a displacement measurement point of the structure, automatically determine a best target among a plurality of candidate targets detected by the radar, and automatically calculate a final conversion factor for converting a displacement of the best target in a line-of-sight direction of the displacement measurement point to a displacement in an actual vibration direction; as well as a structural displacement monitoring step configured to calculate a final displacement by fusing, based on an FIR filter, two displacements, namely, a radar-based displacement obtained by applying the final conversion factor to a phase extracted from the measurement value of the radar of the optimal target and an accelerometer-based displacement obtained by double-integrating the measurement value from the accelerometer.

2. The structural displacement estimation method according to claim 1, wherein the measurement using the radar and the measurement using the accelerometer are performed within the same time period.

3. The structural displacement estimation method according to claim 1, wherein the measuring using the radar and the measuring using the accelerometer are performed in less than one minute. 4 . The structural displacement estimation method according to claim 1 , wherein the radar and the accelerometer are installed at the displacement measurement point of the structure in close proximity to each other and collect the measurement values.

5. The method for estimating structural displacement according to claim 1, wherein the automated initial calibration step comprises: measuring, using the radar, an initial displacement of each of the plurality of candidate targets in the line of sight direction of the displacement measurement point; calculating a plurality of first displacements in the vibration direction by applying a plurality of conversion factor values ​​to the initial displacement of each of the plurality of candidate targets; Measuring the acceleration of the displacement measurement point with the accelerometer; calculating a second displacement by double integrating the acceleration; calculating an RMSE between each of the plurality of first displacements calculated for each of the plurality of candidate targets and the second displacement, and determining a minimum value among the calculated RMSE values ​​as a minimum RMSE value of the candidate target; automatically determining a candidate target having a smallest minimum RMSE value as the optimal target, the smallest minimum RMSE value from the plurality of minimum RMSE values ​​determined for each of the plurality of candidate targets; as well as The conversion factor used to obtain the minimum RSME value of the optimal target is automatically calculated as a final conversion factor of the optimal target. 6 . The structural displacement estimation method according to claim 5 , wherein the plurality of conversion factor values ​​are in the range of 0.5 to 2.

0.

7. The structural displacement estimation method according to claim 1, wherein the structural displacement monitoring step is periodically performed by periodically performing displacement measurement of the optimal target automatically determined in the automated initial calibration step using the radar and the accelerometer.

8. The method for estimating structural displacement according to claim 1, wherein the structural displacement monitoring step comprises: extracting a raw phase by performing the radar measurement on the optimal target using the radar; measuring the acceleration of the displacement measurement point with the accelerometer, and calculating the acceleration-based displacement by double integrating the measured acceleration; When a phase wrapping problem occurs in the original phase because the displacement of the displacement measurement point of the structure is greater than a wavelength of a radar signal of the radar, selecting an unwrapped phase close to a predicted phase using the measured acceleration; Calculating a third displacement in the line of sight direction using the original phase without the phase wrapping problem or the unwrapped phase due to the phase wrapping problem; calculating the radar-based displacement in the vibration direction by applying the final conversion factor to the third displacement in the line-of-sight direction; as well as The final displacement is calculated by fusing the acceleration-based displacement and the radar-based displacement using a finite impulse response (FIR) filter.

9. The method for estimating structural displacement according to claim 8, wherein the step of calculating the final displacement comprises: A radar-based low-frequency displacement is obtained by performing a low-pass filter on the radar-based displacement, an acceleration-based high-frequency displacement is obtained by performing a high-pass filter on the acceleration-based displacement, and the final displacement is calculated by fusing the radar-based low-frequency displacement and the acceleration-based high-frequency displacement.

10. The method for estimating structural displacement according to claim 8, wherein the step of selecting the unwrapped phase comprises: Using the displacements at the (k-1)th and (k-2)th time stages and the (k-1th)th acceleration, we can use the equation and equation To calculate the predicted displacement at the kth time stage and predicted phase as well as By equation The unwrapped phase that is within the range of ±2pπ (where p is an integer) of the original phase and closest to the predicted phase is found and selected as the phase to be used for estimating the radar-based displacement.

11. The method for estimating structural displacement according to claim 8, wherein the final displacement is calculated using the formula Calculated, where C H : Double integration and (2M+1)th order high-pass filter factor, a: Measured acceleration vector, C L : (2M+1)th order low-pass filter factor, u: radar-based displacement vector.

12. The method for estimating structural displacement according to claim 1, The radar is a frequency modulated continuous wave (FMCW) millimeter wave radar, and the FMCW transmits a chirp signal, receives signals reflected from a target candidate group, and uses a signal round trip time between the transmitted and received signals to estimate the displacement in the line of sight direction.

13. A structural displacement estimation system, comprising: a radar mounted directly at a displacement measurement point of the structure, the radar being configured to transmit radar signals toward a plurality of candidate targets that do not change position, and being configured to receive reflection signals reflected from the plurality of candidate targets; an accelerometer mounted at the displacement measurement point of the structure and configured to measure acceleration at the displacement measurement point of the structure, and a displacement estimator configured to perform: an automated initial calibration function configured to collect measurement values ​​from the radar and the accelerometer, respectively, automatically determine a best target among a plurality of candidate targets detected by the radar, and automatically calculate a final conversion factor for converting a displacement of the best target in a line-of-sight direction of the displacement measurement point into a displacement in an actual vibration direction; and a structural displacement monitoring function configured to calculate a final displacement by fusing, based on an FIR filter, two displacements, namely, a radar-based displacement obtained by applying the final conversion factor to a phase extracted from the measurement value of the radar of the optimal target and an accelerometer-based displacement obtained by double integrating the measurement value from the accelerometer.

14. The structural displacement estimation system according to claim 13, wherein the displacement estimator comprises: a computer program written to perform said automated initial calibration function and said structural displacement monitoring function; as well as A processor for executing said computer program.

15. The structural displacement estimation system according to claim 13, wherein when a wrapping problem occurs in the phase due to the displacement of the displacement measurement point of the structure being greater than the wavelength of the radar signal, the displacement estimator is further configured to perform a phase unwrapping processing function to estimate the radar-based displacement by selecting an unwrapped phase closest to a predicted phase using the acceleration measured at the structure.

16. The structural displacement estimation system according to claim 13, The automated initial calibration features include: a function of measuring, using the radar, an initial displacement of each of the plurality of candidate targets in the line of sight direction of the displacement measurement point; a function to calculate a plurality of first displacements in the vibration direction by applying a plurality of conversion factor values ​​to the initial displacement of each of the plurality of candidate targets; A function of measuring the acceleration of the displacement measurement point using the accelerometer; a function for calculating a second displacement by double integrating said acceleration; calculating an RMSE between each of a plurality of radar-based first displacements calculated for each of the plurality of candidate targets and the second displacement, and determining a minimum value among the calculated RMSE values ​​as a function of a minimum RMSE value of the candidate target; automatically determining a candidate target having a smallest minimum RMSE value as a function of the optimal target, the smallest minimum RMSE value from a plurality of the minimum RMSE values ​​determined for each of the plurality of candidate targets; as well as A conversion factor used to obtain the minimum RSME value of the optimal target is automatically calculated as a function of the final conversion factor of the optimal target.

17. The structural displacement estimation system according to claim 13, wherein the structural displacement monitoring function comprises: A function of extracting a raw phase by performing radar measurement on the optimal target using the radar; measuring the acceleration of the displacement measurement point with the accelerometer, and calculating the acceleration-based displacement function by double integrating the measured acceleration; a function of selecting an unwrapped phase close to a predicted phase using the measured acceleration when a phase wrapping problem occurs in the original phase due to the displacement of the displacement measurement point of the structure being greater than the wavelength of the radar signal; A function for calculating a third displacement in the line of sight direction using an original phase without the phase wrapping problem or an unwrapped phase resulting from the phase wrapping problem; a function for calculating the radar-based displacement in the vibration direction by applying the final conversion factor to the third displacement in the line-of-sight direction; as well as A function of calculating the final displacement by fusing the acceleration-based displacement and the radar-based displacement using a finite impulse response (FIR) filter.

18. The structural displacement estimation system according to claim 13, wherein the function of calculating the final displacement comprises: a function of obtaining a low-frequency radar-based displacement by performing a low-pass filtering on the radar-based displacement; A function of obtaining a high frequency acceleration-based displacement by performing a high pass filter on the acceleration-based displacement; as well as A function to calculate the final displacement by fusing the radar-based low-frequency displacement and the acceleration-based high-frequency displacement.

19. A computer executable program stored in a computer readable recording medium to execute the structural displacement estimating method according to any one of claims 1 to 12.

20. A computer-readable recording medium recording a computer-executable program for executing the structural displacement estimating method according to any one of claims 1 to 12.

Images