Information processing device, sensor system, information processing method, and program

Abstract

It accurately detects vehicles passing over bridges. [Solution] The information processing device comprises a first preprocessing unit, a second preprocessing unit, a candidate waveform detection unit, an axle waveform detection unit, and a vehicle passage determination unit. The first preprocessing unit generates first time series data by low-pass filtering displacement amount time series data representing the time change of the bridge displacement amount with a first low-pass filter having a first cutoff frequency. The second preprocessing unit generates second time series data by low-pass filtering the displacement amount time series data with a second low-pass filter having a second cutoff frequency higher than the first cutoff frequency. The candidate waveform detection unit detects a candidate vehicle waveform from the first time series data. The axle waveform detection unit detects one or more axle waveforms from the time range of the candidate vehicle waveforms in the second time series data. The vehicle passage determination unit determines whether or not the candidate vehicle waveform is a passing waveform based on the one or more axle waveforms.

Description

Technical Field

[0001] The present invention relates to an information processing apparatus, a sensor system, an information processing method, and a program.

Background Art

[0002] There is known a system that provides a sensor device for measuring a displacement amount in a structure of infrastructure such as a bridge, and manages the structure based on the displacement amount acquired from the sensor device. For example, there is known a system that detects the date and time and weight when a vehicle passes over a bridge. Such a system can detect the date and time and weight when a vehicle with a suspected overloading passes, and estimate the degree of deterioration of the bridge.

[0003] Patent Document 1 describes a technique for performing preprocessing on time-series data of displacement amounts acquired from a sensor device installed on a bridge, and detecting the passing date and time and weight of a vehicle that has passed over the bridge based on the maximum peak point of the preprocessed time-series data. Further, Patent Document 1 describes a vehicle detection system using a sensor device that measures the amount of expansion and contraction between two points in the traveling direction on a bridge.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] By the way, when a vehicle passes over a bridge, the weight of the vehicle is distributed over a plurality of axles and applied to the bridge. Therefore, even when one vehicle passes, a plurality of peaks occur in the time-series data of the displacement amount. Therefore, the technique of detecting the passing date and time and weight of a vehicle that has passed over the bridge based on the maximum peak point may not be able to accurately estimate the passing time and weight of the vehicle.

[0006] The present invention has been made in view of the above, and aims to provide an information processing device, a sensor system, an information processing method, and a program for accurately detecting vehicles passing over a bridge. [Means for solving the problem]

[0007] To solve the above-mentioned problems and achieve the objective, the information processing device according to the present invention comprises an acquisition unit, a first pre-processing unit, a second pre-processing unit, a candidate waveform detection unit, an axle waveform detection unit, and a vehicle passage determination unit. The acquisition unit acquires displacement time-series data representing the time change of the displacement amount of the bridge detected by a sensor device installed on the bridge. The first pre-processing unit generates first time-series data by low-pass filtering the displacement time-series data with a first low-pass filter having a first cutoff frequency. The second pre-processing unit generates second time-series data by low-pass filtering the displacement time-series data with a second low-pass filter having a second cutoff frequency higher than the first cutoff frequency. The candidate waveform detection unit detects a candidate vehicle waveform from the first time-series data that is estimated to be a waveform that occurs during the period when a vehicle is passing over the bridge. The axle waveform detection unit detects one or more axle waveforms from the time range of the candidate vehicle waveform in the second time series data that are estimated to occur at the time when the axle of the vehicle passes the position on the bridge where the sensor device is installed. The vehicle passage determination unit determines, based on the one or more axle waveforms, whether or not the candidate vehicle waveform is a passage waveform generated when the vehicle passes over the bridge. [Effects of the Invention]

[0008] According to the present invention, vehicles passing over a beam can be detected with high accuracy. [Brief explanation of the drawing]

[0009] [Figure 1] Figure 1 is a diagram showing the configuration of the sensor system according to the embodiment. [Figure 2] Figure 2 shows the arrangement of the sensor devices when viewed from the side of the bridge. [Figure 3] Figure 3 shows the arrangement of the sensor devices when viewed from above on the bridge. [Figure 4] Figure 4 shows an example of time-series data of displacement during vehicle passage, where the sensor device detects the displacement at the exit end in the direction of travel. [Figure 5] Figure 5 shows an example of time-series data of displacement during vehicle passage, where the sensor device detects the displacement at the entrance end in the direction of travel. [Figure 6] Figure 6 shows the functional configuration of the information processing device. [Figure 7] Figure 7 is a flowchart showing the processing flow of an information processing device. [Figure 8] Figure 8 shows the functional configuration of the first preprocessor unit. [Figure 9] Figure 9 shows the functional configuration of the second preprocessor unit. [Figure 10] Figure 10 shows an example of the waveform of time-series data of displacement. [Figure 11] Figure 11 shows examples of waveforms for the second filtered time series data and the second moving average time series data. [Figure 12] Figure 12 shows an example of the waveform of the second time series data. [Figure 13] Figure 13 shows an example of the waveform of the first time series data. [Figure 14] Figure 14 shows an example of the waveform of the second time series data. [Figure 15] Figure 15 is a flowchart showing the processing flow of the candidate waveform detection unit. [Figure 16] Figure 16 illustrates a method for identifying the maximum elevation point and the non-sensor boundary point when a sensor device detects the displacement of a bridge at the exit end in the direction of travel. [Figure 17]FIG. 17 is a diagram for explaining a method of specifying a maximum rise point and a non-sensor side boundary point when the sensor device detects the displacement amount of a bridge at the entrance side end in the traveling direction. [Figure 18] FIG. 18 is a flowchart showing an example of a specific process of a maximum rise point and a non-sensor side boundary point by a candidate waveform detection unit. [Figure 19] FIG. 19 is a flowchart showing the flow of processing of an axle waveform detection unit. [Figure 20] FIG. 20 is a diagram showing an example of second time-series data and the positions of one or more axle waveforms included in the second time-series data. [Figure 21] FIG. 21 is a flowchart showing the flow of processing of a vehicle passage determination unit. [Figure 22] FIG. 22 is a diagram showing a comparison example between the estimation result of the weight of a vehicle by a conventional method and the estimation result of the weight by the information processing apparatus according to the embodiment. [Figure 23] FIG. 23 is a diagram showing the hardware configuration of the information processing apparatus.

MODE FOR CARRYING OUT THE INVENTION

[0010] Hereinafter, embodiments will be described with reference to the drawings.

[0011] FIG. 1 is a diagram showing the configuration of a sensor system 10 according to an embodiment. The sensor system 10 includes a sensor device 20, an edge device 22, and an information processing apparatus 30.

[0012] The sensor device 20 is provided on a bridge through which a vehicle passes. The sensor device 20 detects a displacement amount representing the displacement of the bridge at the position where the sensor device 20 is provided on the bridge. More specifically, the sensor device 20 detects a displacement amount representing the displacement in the downward direction and the upward direction at the position where the sensor device 20 of the bridge is provided.

[0013] The sensor device 20 may indirectly, rather than directly, detect downward and upward displacements at the location on the bridge where the sensor device 20 is installed. For example, the space between two points on the underside of the superstructure, such as the main girder, in the direction of travel expands and contracts as a vehicle passes over the bridge. Therefore, for example, the sensor device 20 may detect the amount of expansion and contraction of the main girder in the direction of travel as the displacement. Alternatively, for example, the sensor device 20 may detect the amount of pressure at a predetermined position on the main girder of the bridge as the displacement.

[0014] The sensor device 20 detects the amount of displacement at predetermined sampling intervals. The sensor device 20 transmits the detected amount of displacement at predetermined sampling intervals to the edge device 22.

[0015] The edge device 22 is connected to the sensor device 20 and performs information processing and communication processing. For example, the edge device 22 is connected to the sensor device 20 by a wired connection via a LAN (Local Area Network).

[0016] The edge device 22 collects displacement amounts detected at predetermined sampling intervals from the sensor device 20. The edge device 22 transmits multiple displacement amounts collected during a transfer interval, for example, longer than the sampling interval, to the information processing device 30 via the network.

[0017] The network may be wired, wireless, or a mixture of wired and wireless. For example, the network may be a LAN, PAN (Personal Area Network), or WAN (Wide Area Network), or a mixed network of PAN, LAN, and WAN. The network may also include cellular communication lines such as LTE (Long Term Evolution).

[0018] The information processing device 30 is a device that performs information processing. The information processing device 30 may be a single computer, or it may be composed of multiple computers, such as in a cloud system.

[0019] The information processing device 30 is connected to the edge device 22 via a network. The information processing device 30 receives multiple displacement values ​​from the edge device 22 and stores each of the received displacement values ​​in association with the sampling time. The information processing device 30 processes the stored displacement values ​​as time-series data of displacement values ​​representing the time change of displacement values ​​representing the displacement of the bridge. By analyzing the time-series data of displacement values, the information processing device 30 identifies the time when a vehicle passed over the bridge. Furthermore, the information processing device 30 estimates the speed of the vehicle that passed over the bridge and the weight of the vehicle that passed over the bridge.

[0020] Figure 2 shows the arrangement of the sensor device 20 when the bridge is viewed from the side. Figure 3 shows the arrangement of the sensor device 20 when the bridge is viewed from above.

[0021] The sensor device 20 detects, for example, the amount of displacement at a position on the end side of the center in the direction of travel on a bridge. Furthermore, the sensor device 20 detects the amount of displacement at a position approximately at the center in the width direction of the vehicle lane on the bridge. The sensor device 20 is installed, for example, on the main girder of the bridge.

[0022] As shown in Figures 2 and 3, the sensor device 20 detects the displacement at the exit end in the direction of vehicle travel, approximately at the center of the vehicle lane of the bridge. Alternatively, the sensor device 20 may detect the displacement at the entrance end in the direction of vehicle travel, approximately at the center of the vehicle lane of the bridge.

[0023] Figure 4 shows an example of time-series data of displacement when a vehicle passes over a bridge, where the sensor device 20 detects the displacement at the exit end in the direction of travel. When a vehicle passes over a bridge, the time-series data of displacement, which represents the change in the displacement of the position at the exit end in the direction of travel on the bridge, changes as shown in Figure 4, for example.

[0024] When the sensor device 20 detects the displacement at the exit end in the direction of travel, the displacement time series data begins to change in the upward direction from the non-sensor side boundary point (A). The non-sensor side boundary point (A) represents the point in the displacement time series data when the vehicle passes over the boundary on the end side of the bridge where the sensor device 20 is not installed.

[0025] The displacement time series data continues to change in an upward direction from the non-sensor boundary point (A), including small fluctuations due to noise, until it reaches the maximum upward point (B) where the amount of increase is greatest. Subsequently, the displacement time series data reverses direction from the maximum upward point (B) and begins to change in a downward direction, continuing to change in a downward direction, including small fluctuations due to noise.

[0026] Subsequently, the displacement time series data reaches the maximum descent point (C) where the amount of descent is greatest. Near the maximum descent point (C), the displacement time series data contains one or more axle waveforms. Each of the one or more axle waveforms is a bell-shaped waveform in which the amount of displacement changes in the downward direction and then reverses to change in the upward direction. Each of the one or more axle waveforms is estimated to occur at the timing when the vehicle's axle passes the position of the sensor device 20.

[0027] For example, when a vehicle with two axles passes by, the displacement time series data will include two axle waveforms near the point of maximum descent (C). For example, when a vehicle with three axles passes by, the displacement time series data will include three axle waveforms near the point of maximum descent (C).

[0028] Next, the displacement time series data reverses direction from the point of maximum descent (C) and begins to change in the upward direction, continuing to change in the upward direction while including small fluctuations due to noise. Then, the displacement time series data ends its change in the upward direction at the sensor-side boundary point (D). The sensor-side boundary point (D) represents the point in the displacement time series data when a vehicle passes over the boundary on the end side of the bridge where the sensor device 20 is installed.

[0029] Figure 5 shows an example of time-series data of displacement when a vehicle passes over a bridge, where the sensor device 20 detects the displacement at the entrance end in the direction of travel. When a vehicle passes over a bridge, the time-series data of displacement, which represents the change in the displacement of the entrance end in the direction of travel on the bridge over time, changes as shown in Figure 5, for example.

[0030] When the sensor device 20 detects the displacement at the entrance end in the direction of travel, the displacement time series data begins to change downward from the sensor boundary point (D). The displacement time series data continues to change downward from the sensor boundary point (D), including minute fluctuations due to noise, until it reaches the maximum downward point (C) where the downward amount is greatest. The displacement time series data representing the time change in the displacement at the entrance end in the direction of travel of the bridge also includes one or more axle waveforms near the maximum downward point (C).

[0031] Next, the displacement time series data reverses direction of change from the maximum downward point (C) and begins to change upward, continuing to change upward while including small fluctuations due to noise. Subsequently, the displacement time series data reaches the maximum upward point (B) where the amount of increase is greatest. Subsequently, the displacement time series data reverses direction of change from the maximum upward point (B) and begins to change downward, continuing to change downward while including small fluctuations due to noise. Finally, the displacement time series data ends its downward change at the non-sensor boundary point (A).

[0032] Thus, even when the sensor device 20 detects the displacement amount at either the exit end or the inlet end in the direction of travel, the displacement amount time series data changes in one direction, the first direction, while including minute fluctuation components due to noise and one or more axle waveforms as the vehicle passes through. After changing in the first direction, it changes in the second direction opposite to the first direction, and after changing in the second direction, it changes back to the first direction.

[0033] Figure 6 shows the functional configuration of the information processing device 30.

[0034] The information processing device 30 includes an acquisition unit 42, a storage unit 44, a first preprocessing unit 46, a second preprocessing unit 48, a candidate waveform detection unit 50, an extraction unit 52, an axle waveform detection unit 54, a vehicle passage determination unit 56, a passage time determination unit 58, a speed estimation unit 60, and a weight estimation unit 62.

[0035] The acquisition unit 42 acquires time-series data of displacement from the edge device 22, which represents the time change in the displacement of the bridge detected by the sensor device 20 installed on the bridge. The storage unit 44 stores the time-series data of displacement acquired by the acquisition unit 42.

[0036] The first preprocessor unit 46 reads the displacement time series data from the storage unit 44. The first preprocessor unit 46 generates first time series data by low-pass filtering the read displacement time series data with a first low-pass filter having a first cutoff frequency. For example, the first cutoff frequency is lower than the frequency corresponding to the natural frequency of the primary mode in the bridge.

[0037] For example, the first preprocessor 46 generates first filtered time series data by applying a first low-pass filter to the displacement time series data. Furthermore, the first preprocessor 46 generates first moving average time series data by applying a moving average to the first filtered time series data over a predetermined time window. Then, the first preprocessor 46 generates first time series data by subtracting the first moving average time series data from the first filtered time series data.

[0038] The second preprocessor unit 48 reads the displacement time series data from the storage unit 44. The second preprocessor unit 48 generates second time series data by low-pass filtering the read displacement time series data with a second low-pass filter having a second cutoff frequency. The second cutoff frequency is higher than the first cutoff frequency. For example, the second cutoff frequency is higher than the natural frequency of the primary mode in the bridge and lower than the frequency corresponding to the natural frequency of the tertiary mode in the bridge.

[0039] For example, the second preprocessor 48 generates second filtered time series data by applying a second low-pass filter to the displacement time series data. Furthermore, the second preprocessor 48 generates second moving average time series data by applying a moving average over a predetermined time window to the second filtered time series data. Then, the second preprocessor 48 generates second time series data by subtracting the second moving average time series data from the second filtered time series data.

[0040] The candidate waveform detection unit 50 acquires first time-series data from the first preprocessing unit 46. From the first time-series data, the candidate waveform detection unit 50 detects a candidate vehicle waveform that is estimated to occur during the period when the vehicle is passing over the bridge.

[0041] The vehicle candidate waveform is a waveform in which the displacement changes from a reference value in a first direction beyond a first threshold, then changes in a second direction opposite to the first direction beyond a second threshold, and then changes back to the reference value in the first direction. The first threshold is a preset value on the side of the reference value in the first direction. The second threshold is a preset value on the side of the reference value in the second direction. The reference value is the displacement amount when the vehicle is not present on the bridge.

[0042] For example, when the sensor device 20 detects the amount of displacement at the exit end in the direction of travel, the candidate waveform detection unit 50 detects the waveform shown in Figure 4, which is included in the first time series data, as the candidate vehicle waveform. That is, when the sensor device 20 detects the amount of displacement at the exit end in the direction of travel, the candidate vehicle waveform is a waveform in which the amount of displacement changes from the reference value upwards beyond a first threshold, then changes downwards beyond a second threshold after changing upwards, and then changes upwards back to the reference value after changing downwards.

[0043] Furthermore, for example, when the sensor device 20 detects a displacement at the inlet end in the direction of travel, the candidate waveform detection unit 50 detects the waveform shown in Figure 5 included in the first time series data as a candidate vehicle waveform. That is, when the sensor device 20 detects a displacement at the inlet end in the direction of travel, the candidate vehicle waveform is a waveform in which the displacement changes from a reference value in the downward direction beyond a first threshold, changes in the upward direction beyond a second threshold after changing in the downward direction, and changes in the downward direction back to the reference value after changing in the upward direction.

[0044] The candidate waveform detection unit 50 then identifies the time of the non-sensor side boundary point (A), the maximum rise point (B), the maximum fall point (C), and the sensor side boundary point (D) for each of the detected vehicle candidate waveforms. Furthermore, the candidate waveform detection unit 50 identifies the displacement amount of the maximum rise point (B) and the displacement amount of the maximum fall point (C) for each of the detected vehicle candidate waveforms.

[0045] The extraction unit 52 extracts partial time-series data from the second time-series data for each of the candidate vehicle waveforms detected by the candidate waveform detection unit 50, corresponding to the time range of the candidate vehicle waveform. In this embodiment, the extraction unit 52 extracts partial time-series data from the second time-series data for each of the candidate vehicle waveforms detected by the candidate waveform detection unit 50, corresponding to the settlement period. The settlement period is the period within the time range of the candidate vehicle waveform during which the bridge is displaced downwards from a reference value.

[0046] The settlement period is the range from the reference point to the sensor-side boundary point (D) when the sensor device 20 detects the displacement at the exit end in the direction of travel. The settlement period is the range from the sensor-side boundary point (D) to the reference point when the sensor device 20 detects the displacement at the inlet end in the direction of travel.

[0047] The axle waveform detection unit 54 detects, for each of the candidate vehicle waveforms detected by the candidate waveform detection unit 50, one or more axle waveforms estimated to occur at the time the vehicle axle passes the position on the bridge where the sensor device 20 is installed, from the time range of the vehicle candidate waveform in the second time series data extracted by the extraction unit 52. In this embodiment, the axle waveform detection unit 54 detects, for each of the candidate vehicle waveforms detected by the candidate waveform detection unit 50, one or more axle waveforms from the time range of the settlement period in the second time series data.

[0048] Each of the one or more axle waveforms has a shorter period than the vehicle candidate waveform, and the displacement changes downward before changing upward as the bridge rises.

[0049] Furthermore, the axle waveform detection unit 54 identifies the peak time and peak displacement amount at the point of change where the waveform changes from a downward direction to an upward direction for each of the one or more axle waveforms. The peak time is the time at the point of change where the waveform changes from a downward direction to an upward direction in the target axle waveform. The peak displacement amount is the displacement amount at the point of change where the waveform changes from a downward direction to an upward direction in the target axle waveform.

[0050] Furthermore, the axle waveform detection unit 54 identifies axle waveforms that are considered noise from one or more axle waveforms and removes those axle waveforms that are considered noise from one or more axle waveforms.

[0051] The vehicle passage determination unit 56 determines, for each of the candidate vehicle waveforms detected by the candidate waveform detection unit 50, whether or not it is a passing waveform generated when a vehicle passes over a bridge, based on one or more detected axle waveforms.

[0052] For example, the vehicle passage determination unit 56 determines whether the vehicle candidate waveform to be judged is a passing waveform based on the number of axle waveforms (one or more) detected from the target vehicle candidate waveform, the peak time of each of the one or more axle waveforms, and the peak displacement amount of each of the one or more axle waveforms. If the vehicle passage determination unit 56 determines that it is a passing waveform, it outputs the peak time and peak displacement amount of each of the one or more axle waveforms detected from the vehicle candidate waveform determined to be a passing waveform, and the time of the non-sensor side boundary point (A).

[0053] The passage time identification unit 58 identifies the time when the vehicle passed over the bridge for each of the passage waveforms. The passage time identification unit 58 identifies the peak time of the axle waveforms in a predetermined time order from among the one or more axle waveforms detected from the time range of the passage waveforms as the time when the vehicle passed over the bridge.

[0054] For example, the passage time identification unit 58 identifies the peak time of the earliest axle waveform among one or more axle waveforms as the time the vehicle crossed the bridge. This allows the passage time identification unit 58 to identify the time when the vehicle's front axle passed over the bridge.

[0055] The speed estimation unit 60 estimates the speed of a vehicle that has crossed the bridge for each of the passing waveforms. The speed estimation unit 60 estimates the vehicle's speed based on the difference between the peak time of an axle waveform in a predetermined time order from among one or more axle waveforms detected from the time range of the passing waveforms, and the time of the non-sensor side boundary point (A).

[0056] For example, when the speed estimation unit 60 detects a displacement at the exit end in the direction of travel, the speed estimation unit 60 calculates the difference between the time at the non-sensor side boundary point (A) and the peak time of the earliest axle waveform among one or more axle waveforms as the difference time. In this case, the speed estimation unit 60 calculates the vehicle speed by dividing the first distance, which is the distance from the boundary at the end of the bridge where the sensor device 20 is not installed to the position where the sensor device 20 is installed, by the difference time. As a result, the speed estimation unit 60 can calculate the speed based on the time from when the vehicle's front axle enters the bridge until it passes over the sensor device 20.

[0057] Furthermore, for example, when the sensor device 20 detects a displacement at the entrance end in the direction of travel, the speed estimation unit 60 calculates the difference between the peak time of the slowest axle waveform among one or more axle waveforms and the time at the non-sensor side boundary point (A) as the difference time. In this case, the speed estimation unit 60 calculates the vehicle speed by dividing the first distance, which is the distance from the boundary of the end of the bridge where the sensor device 20 is not installed to the position where the sensor device 20 is installed, by the difference time. As a result, the speed estimation unit 60 can calculate the speed based on the time from when the rear axle of the vehicle passes over the sensor device 20 until it exits the bridge.

[0058] The weight estimation unit 62 estimates the weight of a vehicle that has passed over the bridge for each of the passing waveforms. The weight estimation unit 62 estimates the weight of the vehicle based on a cumulative value obtained by accumulating the difference values ​​for each of the one or more axle waveforms detected from the time range of the passing waveforms. The difference value for each of the one or more axle waveforms is the absolute value of the difference between the peak displacement and the displacement at the maximum rise point (B). For example, the weight estimation unit 62 calculates the weight of the vehicle by multiplying the cumulative value by a predetermined correction coefficient. In this way, the weight estimation unit 62 can sum the loads applied to the bridge that are distributed across multiple axles, and can estimate the weight of the vehicle with high accuracy.

[0059] Figure 7 is a flowchart showing the processing flow of the information processing device 30. The information processing device 30 processes the time-series data of displacement acquired during a predetermined period, for example, according to the flow shown in Figure 7.

[0060] First, in S11, the information processing device 30 performs a first preprocessing on the displacement time series data acquired over a predetermined period to generate first time series data. More specifically, the information processing device 30 generates first filtered time series data by low-pass filtering the displacement time series data acquired over a predetermined period using a first low-pass filter with a first cutoff frequency. Furthermore, the information processing device 30 generates first moving average time series data by moving average the first filtered time series data over a predetermined time window. Then, the information processing device 30 generates first time series data by subtracting the first moving average time series data from the first filtered time series data.

[0061] Next, in S12, the information processing device 30 performs a second preprocessing on the displacement time series data acquired over a predetermined period to generate second time series data. More specifically, the information processing device 30 generates second filtered time series data by low-pass filtering the displacement time series data acquired over a predetermined period using a second low-pass filter with a second cutoff frequency higher than the first cutoff frequency. Furthermore, the information processing device 30 generates second moving average time series data by moving average the second filtered time series data over a predetermined time window. Finally, the information processing device 30 generates second time series data by subtracting the second moving average time series data from the second filtered time series data.

[0062] The information processing device 30 may execute processes S11 and S12 in parallel.

[0063] Next, in S13, the information processing device 30 detects candidate vehicle waveforms from the first time-series data that are estimated to occur during the period when the vehicle is passing over the bridge. The information processing device 30 then identifies the times of the non-sensor-side boundary point (A), the maximum rise point (B), the maximum descend point (C), and the sensor-side boundary point (D) for each of the detected candidate vehicle waveforms. Furthermore, the information processing device 30 identifies the displacement amount at the maximum rise point (B) and the displacement amount at the maximum descend point (C) for each of the detected candidate vehicle waveforms.

[0064] Next, the information processing device 30 executes the processes from S15 to S20 for each of the detected vehicle candidate waveforms (loop processing between S14 and S21).

[0065] In S15, the information processing device 30 extracts partial time series data from the second time series data, specifically the time range of the candidate vehicle waveform to be processed. In this embodiment, the information processing device 30 extracts partial time series data from the second time series data, specifically the subsidence period of the candidate vehicle waveform to be processed.

[0066] The information processing device 30 may perform the second preprocessing in S12 after S15. In this case, in S15, the information processing device 30 extracts partial time series data including the settlement period of the candidate vehicle waveform to be processed from the displacement amount time series data before the second preprocessing is performed. In this case, the information processing device 30 extracts the partial time series data from a sample timing that is a moving average time window earlier than the sample timing at the beginning of the settlement period. The information processing device 30 then performs the second preprocessing on the extracted partial time series data to generate the second time series data for the settlement period.

[0067] Next, in S16, the information processing device 30 detects one or more axle waveforms from the time range of the vehicle candidate waveforms in the extracted second time series data. In this embodiment, the information processing device 30 detects one or more axle waveforms from the time range of the sinking period in the second time series data.

[0068] Next, in S17, the information processing device 30 determines, based on the detected one or more axle waveforms, whether the candidate vehicle waveform to be judged is a passing waveform generated when a vehicle passes over a bridge. Based on the number of detected one or more axle waveforms, the peak time and peak displacement amount of each of the outputted one or more axle waveforms, the information processing device 30 determines whether the candidate vehicle waveform to be judged is a correct passing waveform generated when a vehicle actually passes over a bridge.

[0069] If the information processing device 30 determines that the waveform is a passing waveform (Yes in S17), it proceeds to process S18. If the information processing device 30 determines that the waveform is not a passing waveform (No in S17), it proceeds to process S21 without executing processes S18 to S20.

[0070] In S18, the information processing device 30 identifies the peak time of the axle waveforms in a predetermined time order from among the one or more axle waveforms detected from the time range of the passing waveform as the time when the vehicle passed over the bridge.

[0071] Next, in S19, the information processing device 30 estimates the vehicle speed based on the difference between the peak time of one or more axle waveforms in a predetermined time order from among the time range of the passing waveforms, and the time of the non-sensor side boundary point (A).

[0072] Next, in S20, the information processing device 30 estimates the vehicle weight based on a cumulative value obtained by accumulating the difference values ​​for each of the one or more axle waveforms detected from the time range of the passing waveform. The difference value for each of the one or more axle waveforms is the absolute value of the difference between the peak displacement and the displacement at the maximum rise point (B). After completing the processing in S20, the information processing device 30 proceeds to processing in S21.

[0073] In S21, the information processing device 30 determines whether it has performed the processes from S15 to S20 for all of the detected vehicle candidate waveforms. If the information processing device 30 has not performed the processes from S15 to S20 for at least one of the detected vehicle candidate waveforms, it returns to S14 and performs the processes from S15 for the next vehicle candidate waveform. If the information processing device 30 has performed the processes from S15 to S20 for all of the detected vehicle candidate waveforms, it terminates this flow.

[0074] Figure 8 shows the functional configuration of the first preprocessing unit 46. The first preprocessing unit 46 includes a first low-pass filter unit 72, a first moving average unit 74, and a first difference calculation unit 76.

[0075] The first low-pass filter section 72 is set to a first cutoff frequency. The first cutoff frequency is lower than the frequency corresponding to the natural frequency of the primary mode in the bridge. For example, the first cutoff frequency is 2 Hz.

[0076] The first low-pass filter unit 72 generates first filtered time-series data by low-pass filtering the displacement time-series data using a first low-pass filter with a first cutoff frequency. In other words, the first low-pass filter unit 72 performs a filtering process that removes frequency components higher than the first cutoff frequency and allows frequency components below the first cutoff frequency to pass through.

[0077] The first moving average unit 74 generates first moving average time series data by moving the first filtered time series data over a predetermined time window. The predetermined time window is appropriately adjusted so that, for example, frequency components lower than the frequency components of the change in displacement caused by the passage of a vehicle are removed.

[0078] The first difference calculation unit 76 generates the first time series data by subtracting the first moving average time series data from the first filtered time series data.

[0079] Figure 9 shows the functional configuration of the second preprocessing unit 48. The second preprocessing unit 48 includes a second low-pass filter unit 78, a second moving average unit 80, and a second difference calculation unit 82.

[0080] The second low-pass filter section 78 is set to a second cutoff frequency. The second cutoff frequency is higher than the natural frequency of the primary mode in the bridge and lower than the frequency corresponding to the natural frequency of the third mode in the bridge. For example, the second cutoff frequency is 10 Hz.

[0081] The second low-pass filter unit 78 generates second filtered time-series data by low-pass filtering the displacement time-series data using a second low-pass filter with a second cutoff frequency. In other words, the second low-pass filter unit 78 performs a filtering process that removes frequency components higher than the second cutoff frequency and allows frequency components below the second cutoff frequency to pass through.

[0082] The second moving average unit 80 generates second moving average time series data by applying a moving average to the second filtered time series data over a predetermined time window. The predetermined time window is appropriately adjusted so that, for example, frequency components lower than the frequency components of the change in displacement caused by the passage of a vehicle are removed.

[0083] The second difference calculation unit 82 generates the second time series data by subtracting the second moving average time series data from the second filtered time series data.

[0084] Figure 10 shows an example of the waveform of the displacement time series data. Figure 11 shows an example of the waveform of the second filtering time series data and the second moving average time series data. Figure 12 shows an example of the waveform of the second time series data.

[0085] Displacement time series data, for example, contains high-frequency noise components that are higher than the frequency of the axle waveform. The second filtered time series data has had the high-frequency noise components present in the displacement time series data removed by low-pass filtering.

[0086] Furthermore, bridges physically fluctuate at frequencies lower than the frequency components of the displacement changes caused by passing vehicles, due to changes in environmental parameters such as temperature or humidity. The second moving average time series data represents these changes in displacement in accordance with changes in environmental parameters.

[0087] The second time series data is generated by subtracting the second moving average time series data from the second filtered time series data. As a result, high-frequency noise components are removed from the displacement time series data, as well as components representing changes in displacement due to changes in environmental parameters. Therefore, the information processing device 30 can accurately determine the time of passage of a vehicle, estimate the vehicle's speed, and estimate the vehicle's weight by detecting the passing waveform based on this second time series data. The same applies to the first time series data.

[0088] Figure 13 shows an example of the waveform of the first time series data. The first time series data is filtered by a first low-pass filter with a first cutoff frequency lower than the second cutoff frequency. For example, the first time series data is filtered so that the frequency components of the axle waveform are removed.

[0089] As a result, the first time-series data does not include the axle waveform, but the waveform is such that the non-sensor boundary point (A), the maximum rise point (B), the maximum fall point (C), and the sensor boundary point (D) are clearly visible without being obscured by noise. Therefore, the information processing device 30 can accurately identify the time and displacement amount of each of the non-sensor boundary point (A), the maximum rise point (B), the maximum fall point (C), and the sensor boundary point (D) based on the first time-series data.

[0090] Figure 14 shows an example of the waveform of the second time series data. The second time series data is filtered by a second low-pass filter with a second cutoff frequency higher than the first cutoff frequency. For example, the second time series data is filtered so that the frequency components of the axle waveform are passed through.

[0091] As a result, the second time-series data may be obscured by noise of similar frequency components to the axle waveform, obscuring the non-sensor boundary point (A), the maximum rise point (B), the maximum fall point (C), and the sensor boundary point (D). However, the second time-series data includes peaks (E, F) due to the axle waveform. Therefore, the information processing device 30 can accurately detect the peak time and peak displacement of one or more axle waveforms based on the second time-series data.

[0092] Figure 15 is a flowchart showing the processing flow of the candidate waveform detection unit 50. The candidate waveform detection unit 50 performs processing according to the flow shown in Figure 15.

[0093] First, in S31, the candidate waveform detection unit 50 detects one or more candidate peak points from the first time series data.

[0094] Each of the one or more candidate peak points is a point of change where the displacement changes from a downward to an upward direction. Furthermore, each of the one or more candidate peak points is a point where the displacement is lower than the reference value, and the absolute value of the difference of the displacement from the reference value is greater than a preset threshold.

[0095] Next, the candidate waveform detection unit 50 executes the processes from S33 to S37 for each of the one or more candidate peak points (loop processing between S32 and S38).

[0096] In S33, the candidate waveform detection unit 50 searches for the first time series data from the target candidate peak point to be processed among one or more candidate peak points, and identifies the sensor-side boundary point (D) in the first time series data.

[0097] When the sensor device 20 detects the displacement of the bridge at the exit end in the direction of travel, the candidate waveform detection unit 50 searches for the first time series data in the direction advancing in time from the target candidate peak point and identifies the point where the displacement first returns to the reference value without changing in the downward direction as the sensor-side boundary point (D).

[0098] Furthermore, when the sensor device 20 detects the amount of bridge displacement at the entrance end in the direction of travel, the candidate waveform detection unit 50 searches for the first time series data in the direction backward in time from the target candidate peak point and identifies the point where the displacement amount first returns to the reference value without changing in the downward direction as the sensor-side boundary point (D).

[0099] The candidate waveform detection unit 50 may not be able to identify the sensor-side boundary point (D) with respect to the target candidate peak point.

[0100] Next, in S34, the candidate waveform detection unit 50 searches from the target candidate peak point to identify a reference point in the first time series data. The reference point is the point of reference value closest to the target candidate peak point, which is temporally opposite to the sensor-side boundary point (D).

[0101] When the sensor device 20 detects the displacement of the bridge at the exit end in the direction of travel, the reference point is the reference value in the first time series data that is temporally immediately preceding the target candidate peak point. Furthermore, when the sensor device 20 detects the displacement of the bridge at the entrance end in the direction of travel, the reference point is the reference value in the first time series data that is temporally immediately following the target candidate peak point.

[0102] Next, in S35, the candidate waveform detection unit 50 searches from the reference point to identify the point of maximum rise (B) and the non-sensor side boundary point (A) in the first time series data.

[0103] The method for identifying the maximum rise point (B) and the non-sensor side boundary point (A) will be described in detail later with reference to Figures 16 and 17. Furthermore, the candidate waveform detection unit 50 may not be able to identify the maximum rise point (B) or the non-sensor side boundary point (A) for the target candidate peak point.

[0104] Next, in S36, the candidate waveform detection unit 50 determines whether it was able to identify all of the non-sensor side boundary point (A), the maximum rise point (B), and the sensor side boundary point (D).

[0105] If all of the non-sensor side boundary point (A), the maximum rise point (B), and the sensor side boundary point (D) can be identified (Yes in S36), the candidate waveform detection unit 50 proceeds to S37. If even one of the non-sensor side boundary point (A), the maximum rise point (B), and the sensor side boundary point (D) cannot be identified (No in S36), the candidate waveform detection unit 50 proceeds to S38 without executing the process in S37.

[0106] In S37, the candidate waveform detection unit 50 identifies the range between the non-sensor side boundary point (A) and the sensor side boundary point (D) in the first time series data as a candidate vehicle waveform.

[0107] Furthermore, the candidate waveform detection unit 50 identifies the target candidate peak point being processed as the maximum descent point (C). The candidate waveform detection unit 50 also identifies the times of the non-sensor side boundary point (A), the maximum rise point (B), the maximum descent point (C), and the sensor side boundary point (D). In addition, the candidate waveform detection unit 50 identifies the displacement amounts of the maximum rise point (B) and the maximum descent point (C).

[0108] When the candidate waveform detection unit 50 finishes processing in S37, it proceeds to processing in S38. In S38, if the candidate waveform detection unit 50 has not yet processed all of the detected candidate peak points, it returns to processing in S32 and starts processing in S33 for the next candidate peak point. When the candidate waveform detection unit 50 has processed all of the detected candidate peak points, it terminates this flow.

[0109] Figure 16 is a diagram illustrating how to identify the maximum elevation point (B) and the non-sensor side boundary point (A) when the sensor device 20 detects the amount of bridge displacement at the exit end in the direction of travel.

[0110] When the sensor device 20 detects the amount of bridge displacement at the exit end in the direction of travel, the candidate waveform detection unit 50 identifies the maximum rise point (B) and the non-sensor side boundary point (A) as follows.

[0111] First, the candidate waveform detection unit 50 searches for the first time-series data in a time-backward direction from the reference point. Then, the candidate waveform detection unit 50 identifies the point where the displacement amount changes in the downward direction for the first time after it has changed in the upward direction beyond a first threshold as the maximum upward point (B). In this case, the first threshold is a preset value on the upward side from the reference value. Note that, for example, if the displacement amount does not exceed the first threshold and returns to the reference value, the candidate waveform detection unit 50 cannot identify the maximum upward point (B).

[0112] Next, the candidate waveform detection unit 50 identifies the maximum rise point (B) and then searches for the first time series data in a direction that is temporally backward from the maximum rise point (B). Then, the candidate waveform detection unit 50 identifies the point where the displacement changes to an upward direction for the first time after it has changed to a downward direction at the maximum rise point (B) as the non-sensor side boundary point (A).

[0113] When the sensor device 20 detects the amount of bridge displacement at the exit end in the direction of travel, the candidate waveform detection unit 50 can identify the maximum rise point (B) and the non-sensor side boundary point (A) by performing the above processing.

[0114] Figure 17 is a diagram illustrating how to identify the maximum elevation point (B) and the non-sensor side boundary point (A) when the sensor device 20 detects the amount of bridge displacement at the entrance end in the direction of travel.

[0115] When the sensor device 20 detects the amount of bridge displacement at the entrance end in the direction of travel, the candidate waveform detection unit 50 identifies the maximum rise point (B) and the non-sensor side boundary point (A) as follows.

[0116] First, the candidate waveform detection unit 50 searches for first time-series data in a direction that advances time from the reference point. Then, the candidate waveform detection unit 50 identifies the point where the displacement amount changes in the downward direction for the first time after it has changed in the upward direction beyond the second threshold as the maximum upward point (B). In this case, the second threshold is a preset value on the upward side from the reference value. Note that, for example, if the displacement amount does not exceed the second threshold and returns to the reference value, the candidate waveform detection unit 50 cannot identify the maximum upward point (B).

[0117] Next, the candidate waveform detection unit 50 identifies the maximum rise point (B) and then searches for the first time series data in a direction that advances time from the maximum rise point (B). Then, the candidate waveform detection unit 50 identifies the point where the displacement changes to an upward direction for the first time after it has changed to a downward direction at the maximum rise point (B) as the non-sensor side boundary point (A).

[0118] When the sensor device 20 detects the amount of bridge displacement at the entrance end in the direction of travel, the candidate waveform detection unit 50 can identify the maximum rise point (B) and the non-sensor side boundary point (A) by performing the above processing.

[0119] Figure 18 is a flowchart showing an example of the process by which the candidate waveform detection unit 50 identifies the maximum rise point (B) and the non-sensor side boundary point (A) when the sensor device 20 detects the amount of bridge displacement at the exit end in the direction of travel.

[0120] The candidate waveform detection unit 50 may, for example, identify the maximum rise point (B) and the non-sensor side boundary point (A) in the flow shown in Figure 18 when the sensor device 20 detects the amount of bridge displacement at the exit end in the direction of travel.

[0121] First, in S41, the candidate waveform detection unit 50 assigns 0 to the variable representing the previous state, null to the variable representing the previous time, and null to the variable representing the previous displacement amount.

[0122] Next, in S42, the candidate waveform detection unit 50 substitutes the time of the reference point into the variable representing the processing target time, and substitutes the displacement amount of the reference point into the variable representing the processing target displacement amount.

[0123] Next, the candidate waveform detection unit 50 executes the processes from S44 to S56 for each sample of the first time series data (loop processing between S43 and S57).

[0124] In S44, the candidate waveform detection unit 50 determines whether the variable representing the previous state is 0 or not. If the variable representing the previous state is 0 (Yes in S44), the candidate waveform detection unit 50 proceeds to S45. If the variable representing the previous state is not 0 (No in S44), the candidate waveform detection unit 50 proceeds to S46.

[0125] In S45, the candidate waveform detection unit 50 assigns 1 to the variable representing the previous state. After completing S45, the candidate waveform detection unit 50 proceeds to S55.

[0126] In S46, the candidate waveform detection unit 50 performs a displacement amount increase / decrease determination process. In the increase / decrease determination process, the candidate waveform detection unit 50 determines whether the value of the variable representing the displacement amount to be processed has increased or decreased from the value of the variable representing the displacement amount immediately before. After completing S46, the candidate waveform detection unit 50 proceeds to S47.

[0127] In S47, the candidate waveform detection unit 50 determines whether the variable representing the previous state is 1 or not. If the variable representing the previous state is 1 (Yes in S47), the candidate waveform detection unit 50 proceeds to S48. If the variable representing the previous state is not 1 (No in S47), the candidate waveform detection unit 50 proceeds to S51.

[0128] In S48, the candidate waveform detection unit 50 determines whether the result of the increase / decrease determination process was an increase or not. If the result of the increase / decrease determination process is an increase (Yes in S48), the candidate waveform detection unit 50 proceeds to S49. If the result of the increase / decrease determination process is not an increase (No in S48), the candidate waveform detection unit 50 proceeds to S55.

[0129] In S49, the candidate waveform detection unit 50 assigns 2 to the variable representing the previous state. After completing S49, the candidate waveform detection unit 50 proceeds to S50.

[0130] In S50, the candidate waveform detection unit 50 assigns the value of the variable representing the previous time to the time of the maximum rise point (B). After completing S50, the candidate waveform detection unit 50 proceeds to S55.

[0131] In S51, the candidate waveform detection unit 50 determines whether the variable representing the previous state is 2 or not. If the variable representing the previous state is 2 (Yes in S51), the candidate waveform detection unit 50 proceeds to S52. If the variable representing the previous state is not 2 (No in S51), the candidate waveform detection unit 50 proceeds to S55.

[0132] In S52, the candidate waveform detection unit 50 determines whether the result of the increase / decrease judgment process was a decrease or not. If the result of the increase / decrease judgment process is a decrease (Yes in S52), the candidate waveform detection unit 50 proceeds to S53. If the result of the increase / decrease judgment process is not a decrease (No in S52), the candidate waveform detection unit 50 proceeds to S55.

[0133] In S53, the candidate waveform detection unit 50 assigns 3 to the variable representing the previous state. After completing S53, the candidate waveform detection unit 50 proceeds to S54.

[0134] In S54, the candidate waveform detection unit 50 assigns the value of the variable representing the previous time to the time of the non-sensor side boundary point (A). After completing S54, the candidate waveform detection unit 50 proceeds to S57.

[0135] In S55, the candidate waveform detection unit 50 assigns the value of the variable representing the processing target time to the variable representing the immediate past time, and assigns the value of the variable representing the processing target displacement amount to the variable representing the immediate past displacement amount. After completing S55, the candidate waveform detection unit 50 proceeds to S56.

[0136] In S56, the candidate waveform detection unit 50 assigns the time one sample prior to the processing target time to the variable representing the processing target time. The candidate waveform detection unit 50 also assigns the displacement amount one sample prior to the processing target time in the first time series data to the variable representing the displacement amount to be processed. After completing S56, the candidate waveform detection unit 50 proceeds to processing in S57.

[0137] In S57, the candidate waveform detection unit 50 determines whether the variable representing the previous state is 3 or not. If the variable representing the previous state is not 3, the process returns to S43 and repeats from S44. If the variable representing the previous state is 3, the candidate waveform detection unit 50 terminates this flow.

[0138] By performing the above processing, the candidate waveform detection unit 50 can identify the maximum rise point (B) and the non-sensor side boundary point (A) when the sensor device 20 detects the amount of bridge displacement at the exit end in the direction of travel. If the sensor device 20 detects the amount of bridge displacement at the entrance end in the direction of travel, in S56, the time one sample after the processing target time should be substituted into the variable representing the processing target time, and the amount of displacement one sample after the processing target time in the first time series data should be substituted into the variable representing the amount of displacement to be processed.

[0139] Figure 19 is a flowchart showing the processing flow of the axle waveform detection unit 54. Figure 20 is a diagram showing an example of the second time series data and the positions of one or more axle waveforms included in the second time series data.

[0140] The axle waveform detection unit 54 performs the processing shown in Figure 19 for each candidate vehicle waveform.

[0141] First, in S61, the axle waveform detection unit 54 detects one or more axle waveforms from the time range of the settlement period of the target vehicle candidate waveform to be processed in the second time series data.

[0142] The time range for the settlement period of the candidate vehicle waveform is between the time at the reference point and the time at the sensor-side boundary point (D). Furthermore, each of the one or more axle waveforms is a bell-shaped waveform in which the displacement changes from a downward direction to an upward direction.

[0143] Next, in S62, the axle waveform detection unit 54 identifies the peak time and peak displacement amount at the point of change where the waveform changes from a downward direction to an upward direction for each of the one or more axle waveforms.

[0144] Next, in S63, the axle waveform detection unit 54 detects two or more valley waveforms from the time range of the settlement period of the target vehicle candidate waveform in the second time series data. Each of the two valley waveforms is a waveform in which the displacement changes in the upward direction before changing in the downward direction.

[0145] Next, in S64, the axle waveform detection unit 54 identifies a target valley waveform for each of the one or more axle waveforms. The target valley waveform is the waveform on the downward side of the change point of two valley waveforms that are temporally adjacent to the axle waveform.

[0146] Next, in S65, the axle waveform detection unit 54 calculates the maximum amplitude. More specifically, the axle waveform detection unit 54 identifies the peak displacement amount in the maximum axle waveform, which is the one or more axle waveforms with the largest peak displacement amount. Furthermore, the axle waveform detection unit 54 calculates the maximum amplitude as the absolute value of the difference between the peak displacement amount in the maximum axle waveform and the displacement amount at the maximum rise point (B).

[0147] Next, the axle waveform detection unit 54 executes the processes from S67 to S70 for each of the one or more axle waveforms (loop processing between S66 and S71).

[0148] In S67, the axle waveform detection unit 54 determines whether the absolute value of the difference between the displacement amount of the target valley waveform and the peak displacement amount of the axle waveform to be processed from among one or more axle waveforms is less than or equal to a preset reference wave height value. If the axle waveform detection unit 54 determines that the absolute value of the difference between the displacement amount of the target valley waveform and the peak displacement amount is less than or equal to the reference wave height value (Yes in S67), it proceeds to S68.

[0149] In S68, the axle waveform detection unit 54 removes the axle waveform to be processed as noise from among one or more axle waveforms. That is, the axle waveform detection unit 54 removes as noise any axle waveform from one or more axle waveforms in which the absolute value of the difference between the displacement amount of the target trough waveform and the peak displacement amount is less than or equal to a preset reference wave height value. In this way, the axle waveform detection unit 54 can remove as noise any axle waveforms that do not have a sufficiently high peak shape, such as the black triangle mark in Figure 20. After completing the processing in S68, the axle waveform detection unit 54 proceeds to processing in S71.

[0150] If the axle waveform detection unit 54 determines that the absolute value of the difference between the displacement amount and the peak displacement amount of the target valley waveform is not less than or equal to the reference wave height (No. in S67), it proceeds to S69.

[0151] In S69, the axle waveform detection unit 54 determines whether the peak displacement amount of the axle waveform to be processed from among one or more axle waveforms is within an appropriate range relative to the maximum amplitude. More specifically, the axle waveform detection unit 54 determines that it is not within an appropriate range if the absolute value of the difference between the peak displacement amount and the displacement amount at the maximum rise point (B) is less than or equal to a preset reference ratio relative to the maximum amplitude. For example, the axle waveform detection unit 54 also determines that it is not within an appropriate range if the absolute value of the difference between the peak displacement amount and the maximum rise point (B) is less than or equal to a preset reference amplitude.

[0152] If the axle waveform detection unit 54 determines that the peak displacement amount of the axle waveform to be processed is not within an appropriate range relative to the maximum amplitude (Yes in S69), it proceeds to process S70. If the axle waveform detection unit 54 determines that the peak displacement amount of the axle waveform to be processed is within an appropriate range relative to the maximum amplitude (No in S69), it proceeds to process S71.

[0153] In S70, the axle waveform detection unit 54 removes the axle waveform to be processed as noise from among one or more axle waveforms. Specifically, the axle waveform detection unit 54 removes as noise any axle waveform from one or more axle waveforms if the absolute value of the difference between the peak displacement and the displacement at the maximum rise point is less than or equal to a preset reference ratio to the maximum amplitude, or if the absolute value of the difference between the peak displacement and the maximum rise point is less than or equal to a preset reference amplitude. In this way, the axle waveform detection unit 54 can remove as noise any axle waveform whose displacement is too small when compared to other axle waveforms, such as the white triangle mark in Figure 20. After completing the processing in S70, the axle waveform detection unit 54 proceeds to processing in S71.

[0154] In S71, the axle waveform detection unit 54 determines whether or not it has performed the processes from S67 to S70 for all of the one or more axle waveforms. If the axle waveform detection unit 54 has not yet performed any of the processes for one or more axle waveforms, it returns to S66 and performs the processes from S67 for the next axle waveform. If the axle waveform detection unit 54 has performed the processes for all of the one or more axle waveforms, it proceeds to S72.

[0155] In S72, the axle waveform detection unit 54 registers the peak time and peak displacement amount for each of the one or more axle waveforms after noise has been removed. After completing the process in S72, the axle waveform detection unit 54 terminates this flow.

[0156] Figure 21 is a flowchart showing the processing flow of the vehicle passage determination unit 56. The vehicle passage determination unit 56 executes the processing shown in Figure 21 for each candidate vehicle waveform.

[0157] First, in S81, the vehicle passage determination unit 56 acquires the peak time and peak displacement amount for each of the one or more axle waveforms registered by the axle waveform detection unit 54 after noise has been removed, for the vehicle candidate waveform to be processed.

[0158] Next, in S82, the vehicle passage determination unit 56 determines whether the number of axle waveforms (one or more) remaining after noise has been removed from the candidate vehicle waveform to be processed is within a preset threshold number.

[0159] If the number of axle waveforms (one or more) after noise removal is not within the threshold number (No. in S82), the vehicle passage determination unit 56 terminates this flow. This allows the vehicle passage determination unit 56 to determine that vehicle candidate waveforms that are greater than the number of axles a normal vehicle has are not waveforms generated by an actual vehicle passing through.

[0160] If the number of axle waveforms after noise removal is within a threshold number (Yes in S82), the vehicle passage determination unit 56 proceeds to processing in S83.

[0161] In S83, the vehicle passage determination unit 56 determines whether the absolute value of the difference between the peak displacement amount of the largest axle waveform and the displacement amount of the maximum rise point (B) among the one or more axle waveforms after noise has been removed is greater than or equal to a preset threshold amplitude.

[0162] If the absolute value of the difference between the peak displacement amount in the maximum axle waveform and the displacement amount at the maximum rise point (B) is not greater than or equal to the threshold amplitude (No. in S83), the vehicle passage determination unit 56 terminates this flow. As a result, the vehicle passage determination unit 56 can determine that vehicle candidate waveforms with an estimated weight lighter than the actual vehicle are not waveforms generated by the actual passage of a vehicle.

[0163] If the absolute value of the difference between the peak displacement amount in the maximum axle waveform and the displacement amount at the maximum rise point (B) is greater than or equal to the threshold amplitude (Yes in S83), the vehicle passage determination unit 56 proceeds to processing in S84.

[0164] In S84, the vehicle passage determination unit 56 determines whether the time difference between the time of the non-sensor side boundary point (D) and the peak time of a predetermined axle waveform among one or more axle waveforms after noise removal is within a predetermined time range. For example, the vehicle passage determination unit 56 determines whether the time difference between the time of the non-sensor side boundary point (D) and the peak time of the first axle waveform among one or more axle waveforms after noise removal is within a predetermined time range.

[0165] If the time difference between the time at the non-sensor boundary point (D) and the peak time in the predetermined axle waveform is not within a predetermined time range (No. in S84), the vehicle passage determination unit 56 terminates this flow. As a result, the vehicle passage determination unit 56 can determine that for candidate vehicle waveforms where the estimated passing speed is not within the range of the actual speed at the time of the vehicle's passage, the waveform is not a passing waveform generated by an actual vehicle passing.

[0166] If the time difference between the time at the non-sensor boundary point (D) and the peak time in the waveform of a predetermined axle is within a predetermined time range (Yes in S84), the vehicle passage determination unit 56 proceeds to S85.

[0167] In S85, the vehicle passage determination unit 56 registers the candidate vehicle waveform to be processed as the passing waveform.

[0168] The vehicle passage determination unit 56 terminates this flow after completing the process in S85.

[0169] By performing the above processing, the vehicle passage determination unit 56 can determine that the vehicle candidate waveform is a passing waveform if the first determination condition is met. The first determination condition includes that the number of one or more axle waveforms after noise removal is less than a preset threshold number. Furthermore, the first determination condition includes that, among the one or more axle waveforms after noise removal, the absolute value of the difference between the peak displacement amount of the largest axle waveform and the displacement amount of the maximum rise point is greater than or equal to a preset second threshold amplitude. Furthermore, the first determination condition includes that the time difference between the time of the non-sensor side boundary point (D) and the peak time of a predetermined axle waveform among the one or more axle waveforms after noise removal is within a predetermined time range.

[0170] Figure 22 shows a comparison of the results of estimating the weight of a vehicle using a conventional method and the results of estimating the weight using the information processing device 30 according to the embodiment. Note that the conventional method estimates the weight of the vehicle based only on the maximum peak value.

[0171] As shown in Figure 22, the weight estimated by the information processing device 30 according to this embodiment has a smaller error rate than the estimation results obtained by conventional methods. Thus, the information processing device 30 according to this embodiment can accurately estimate the weight of a vehicle that has passed over a bridge.

[0172] As described above, the information processing device 30 according to this embodiment can accurately detect vehicles passing over a bridge. For example, the information processing device 30 can accurately determine the time when a vehicle passed over the bridge. Also, for example, the information processing device 30 can accurately estimate the speed of a vehicle that has passed over a bridge. Also, for example, the information processing device 30 can accurately estimate the weight of a vehicle that has passed over a bridge.

[0173] Figure 23 shows the hardware configuration of the information processing device 30. The information processing device 30 can be implemented, for example, by a device with a hardware configuration similar to that of a general-purpose computer, as shown in Figure 23. The information processing device 30 comprises a CPU (Central Processing Unit) 301, an operating device 302, a display device 303, a main memory 305, an auxiliary memory 306, a communication device 307, and a bus 309. Each component is connected by the bus 309.

[0174] The CPU 301 uses a predetermined area of ​​the main memory 305 as a working area and executes various processes in cooperation with various programs pre-stored in the auxiliary memory 306, etc., and comprehensively controls the operation of each part that constitutes the server device 26. In addition, the CPU 301 operates the operating device 302, the display device 303, the communication device 307, etc., in cooperation with the programs.

[0175] The operating device 302 is an input device such as a touch panel, mouse, or keyboard, which receives information input from the user as an instruction signal and outputs that instruction signal to the CPU 301. The display device 303 displays various information based on the display signal from the CPU 301.

[0176] The main memory 305 is a volatile storage medium such as SDRAM (Synchronous Dynamic Random Access Memory). The main memory 305 functions as a workspace for the CPU 301.

[0177] The auxiliary storage device 306 is a rewritable recording device such as a semiconductor storage medium like flash memory, or a magnetically or optically recordable storage medium. The auxiliary storage device 306 stores the program used for control. The communication device 307 transmits and receives data with other devices.

[0178] The program executed by the information processing device 30 may be stored on a computer connected to a network such as the Internet and provided by downloading it via the network. Alternatively, the program executed by the information processing device 30 may be pre-installed on a portable storage medium or the like and provided in that form.

[0179] The program executed by the information processing device 30 has a module configuration that includes an acquisition module, a first preprocessing module, a second preprocessing module, a candidate waveform detection module, an extraction module, an axle waveform detection module, a vehicle passage determination module, a passage time determination module, a speed estimation module, and a weight estimation module. The CPU 301 reads such a program from a storage medium or the like and loads each of the above modules into the main memory 305. Then, by executing such a program, the CPU 301 functions as an acquisition unit 42, a first preprocessing unit 46, a second preprocessing unit 48, a candidate waveform detection unit 50, an extraction unit 52, an axle waveform detection unit 54, a vehicle passage determination unit 56, a passage time determination unit 58, a speed estimation unit 60, and a weight estimation unit 62. In addition, by executing such a program, the CPU 301 causes the main memory 305 or auxiliary memory 306 to function as a storage unit 44. Furthermore, some or all of the acquisition unit 42, the first preprocessing unit 46, the second preprocessing unit 48, the candidate waveform detection unit 50, the extraction unit 52, the axle waveform detection unit 54, the vehicle passage determination unit 56, the passage time determination unit 58, the speed estimation unit 60, and the weight estimation unit 62 may be configured by hardware.

[0180] Although embodiments of the present invention have been described above, these embodiments are presented as examples only and are not intended to limit the scope of the invention. Various modifications can be made to the embodiments. [Explanation of Symbols]

[0181] 10 Sensor Systems 20 Sensor device 22 Edge Devices 30 Information Processing Devices 42 Acquisition Department 44 Memory section 46 First pre-processing unit 48 Second pre-processing unit 50 Candidate waveform detection unit 52 Cutout part 54 Axle waveform detection unit 56 Vehicle passage determination unit 58 Passage time identification section 60 Speed ​​estimation part 62 Weight estimation section 72 First Low-Pass Filter Section 74 1st moving average part 76 1st difference operation section 78 Second Low-Pass Filter Section 80 2nd moving average part 82 Second difference operation section

Claims

1. An acquisition unit that acquires time-series data of displacement amount representing the time change of the displacement amount of the bridge detected by a sensor device installed on the bridge, A first preprocessor generates first time series data by low-pass filtering the displacement time series data with a first low-pass filter of a first cutoff frequency, A second preprocessor generates second time-series data by low-pass filtering the aforementioned displacement time-series data with a second low-pass filter having a second cutoff frequency higher than the first cutoff frequency. A candidate waveform detection unit detects candidate vehicle waveforms from the first time-series data that are estimated to occur during the period when the vehicle is passing over the bridge, An axle waveform detection unit detects one or more axle waveforms that are estimated to occur at the time the axle of the vehicle passes the position on the bridge where the sensor device is installed, based on the time range of the candidate vehicle waveform in the second time series data. A vehicle passage determination unit determines whether the candidate vehicle waveform is a passage waveform generated when the vehicle passes over the bridge, based on the one or more axle waveforms mentioned above. An information processing device equipped with the following features.

2. The axle waveform detection unit detects one or more axle waveforms from the time range of the sinking period in the second time series data. The aforementioned settlement period is the period within the time range of the candidate vehicle waveform during which the bridge is displaced downwards compared to the reference value. The aforementioned reference value is the amount of displacement when the vehicle is not present on the bridge. The information processing apparatus according to claim 1.

3. The vehicle candidate waveform is a waveform in which the displacement amount changes from the reference value in a first direction beyond a first threshold, then changes in a second direction opposite to the first direction beyond a second threshold after changing in the first direction, and then changes in the first direction back to the reference value after changing in the second direction. The first threshold is a preset value on the side of the reference value in the first direction, The second threshold is a preset value on the second side of the reference value, The information processing apparatus according to claim 2.

4. Each of the one or more axle waveforms is a waveform with a shorter period than the vehicle candidate waveform, in which the displacement changes in the downward direction before changing in the upward direction as the bridge rises. The information processing apparatus according to claim 3.

5. The first preprocessing unit generates the first time series data by subtracting the first moving average time series data, which is obtained by moving the first filtered time series data, from the first filtered time series data obtained by low-pass filtering the displacement time series data with the first low-pass filter, The second preprocessing unit generates the second time series data by subtracting the second moving average time series data, which is obtained by moving the second filtered time series data, from the second filtered time series data obtained by low-pass filtering the displacement time series data using the second low-pass filter. The information processing apparatus according to claim 4.

6. The candidate waveform detection unit detects one or more candidate peak points from the first time series data. Each of the one or more candidate peak points is a point of change where the displacement changes from the downward direction to the upward direction, where the displacement is greater than the reference value in the downward direction, and where the absolute value of the difference of the displacement from the reference value is greater than a preset threshold. The information processing apparatus according to claim 4.

7. The sensor device detects the amount of displacement at the exit end in the direction of travel of the vehicle. For each of the one or more candidate peak points, the candidate waveform detection unit: The first time-series data is searched in a direction that advances time from the target candidate peak point among the one or more candidate peak points, and the point where the displacement amount first returns to the reference value without changing in the downward direction is identified as the sensor-side boundary point. The first time-series data is searched in a direction backward in time from the reference point that is the reference value immediately preceding the target candidate peak point, and the point at which the displacement amount changes to the downward direction for the first time after it exceeds the first threshold is identified as the maximum upward point. The first time-series data is searched in a time-backward direction from the point of maximum upward movement, and the point at which the displacement changes first in the upward direction after changing in the downward direction is identified as the non-sensor side boundary point. If the candidate waveform detection unit can identify all of the sensor-side boundary point, the maximum rise point, and the non-sensor-side boundary point for each of the one or more candidate peak points, it identifies the range between the non-sensor-side boundary point and the sensor-side boundary point in the first time series data as the vehicle candidate waveform. The information processing apparatus according to claim 6.

8. The sensor device detects the amount of displacement at the entrance end in the direction of travel of the vehicle. For each of the one or more candidate peak points, the candidate waveform detection unit: The first time-series data is searched in a direction backward in time from the target candidate peak point among the one or more candidate peak points, and the point where the displacement amount first returns to the reference value without changing in the downward direction is identified as the sensor-side boundary point. The first time-series data is searched in a direction that advances in time from the reference point which is the reference value immediately following the target candidate peak point, and the point at which the displacement amount changes to the downward direction for the first time after it exceeds the second threshold is identified as the maximum upward point. The first time-series data is searched in the direction of time advancement from the point of maximum upward movement, and the point at which the displacement changes first in the upward direction after changing in the downward direction is identified as the non-sensor side boundary point. If the candidate waveform detection unit can identify all of the sensor-side boundary point, the maximum rise point, and the non-sensor-side boundary point for each of the one or more candidate peak points, it identifies the range between the non-sensor-side boundary point and the sensor-side boundary point in the first time series data as the vehicle candidate waveform. The information processing apparatus according to claim 6.

9. The axle waveform detection unit is, For each of the one or more axle waveforms, the peak time and peak displacement amount at the point of change where the waveform changes from the downward direction to the upward direction are identified. For each of the one or more axle waveforms, the one on the downward direction of the displacement amount at the change point of two temporally adjacent valley waveforms is identified as the target valley waveform. From the one or more axle waveforms mentioned above, axle waveforms in which the absolute value of the difference between the displacement amount of the target valley waveform and the peak displacement amount is less than or equal to a preset reference wave height are removed as noise. Each of the two valley waveforms is a waveform in which the displacement changes in the upward direction before changing in the downward direction. The information processing apparatus according to claim 7 or 8.

10. The axle waveform detection unit is, The absolute value of the difference between the peak displacement amount in the maximum axle waveform, which is the one or more axle waveforms with the largest peak displacement amount, and the displacement amount at the maximum upward point is calculated as the maximum amplitude. From the one or more axle waveforms described above, axle waveforms in which the absolute value of the difference between the peak displacement and the displacement at the maximum rise point is less than or equal to a preset reference ratio with respect to the maximum amplitude, or the absolute value of the difference between the peak displacement and the maximum rise point is less than or equal to a preset reference amplitude, are removed as noise. The information processing apparatus according to claim 9.

11. The vehicle passage determination unit determines, if the first determination condition is met, that the candidate vehicle waveform is the waveform at the time of passage. The first determination condition is, The number of axle waveforms (one or more) after the noise has been removed is less than a preset threshold number. Among the one or more axle waveforms after the noise has been removed, the absolute value of the difference between the peak displacement amount in the largest axle waveform and the displacement amount at the maximum rise point is greater than or equal to a preset first threshold amplitude. The time difference between the time of the non-sensor side boundary point and the peak time of a predetermined axle waveform among the one or more axle waveforms after the noise has been removed is within a predetermined time range. The information processing apparatus according to claim 10.

12. The vehicle further comprises a passage time determination unit that identifies the peak time of the axle waveforms in a predetermined time order from among the one or more axle waveforms detected from the time range of the passage waveform as the time when the vehicle passed over the bridge. The information processing apparatus according to claim 9.

13. The vehicle speed estimation unit further comprises a speed estimation unit that estimates the vehicle speed based on the difference time between the peak time of the axle waveforms in a predetermined time order from among the one or more axle waveforms detected from the time range of the passing waveform, and the time of the non-sensor side boundary point. The information processing apparatus according to claim 9.

14. The system further includes a weight estimation unit that estimates the weight of the vehicle based on an accumulated value obtained by accumulating the difference values ​​in each of the one or more axle waveforms detected from the time range of the passing waveform, The difference value in each of the one or more axle waveforms is the absolute value of the difference between the peak displacement amount, which is the displacement amount at the point of change from the downward direction to the upward direction, and the displacement amount at the maximum upward point. The information processing apparatus according to claim 9.

15. A sensor device installed on the bridge to detect the amount of displacement of the bridge, Information processing device and Equipped with, The aforementioned information processing device is An acquisition unit that acquires time-series data of displacement amount representing the time change of the displacement amount of the bridge detected by the sensor device, A first preprocessor generates first time series data by low-pass filtering the displacement time series data with a first low-pass filter of a first cutoff frequency, A second preprocessor generates second time-series data by low-pass filtering the aforementioned displacement time-series data with a second low-pass filter having a second cutoff frequency higher than the first cutoff frequency. A candidate waveform detection unit detects candidate vehicle waveforms from the first time-series data that are estimated to occur during the period when the vehicle is passing over the bridge, An axle waveform detection unit detects one or more axle waveforms that are estimated to occur at the time the axle of the vehicle passes the position on the bridge where the sensor device is installed, based on the time range of the candidate vehicle waveform in the second time series data. A vehicle passage determination unit determines whether the candidate vehicle waveform is a passage waveform generated when the vehicle passes over the bridge, based on the one or more axle waveforms mentioned above. A sensor system having the following features.

16. The information processing device acquires time-series data of displacement, which represents the time change in the displacement of the bridge detected by a sensor device installed on the bridge. The information processing device generates first time-series data by low-pass filtering the displacement time-series data with a first low-pass filter having a first cutoff frequency. The information processing device generates a second time series data obtained by low-pass filtering the displacement time series data with a second low-pass filter having a second cutoff frequency higher than the first cutoff frequency. The information processing device detects, from the first time-series data, a candidate vehicle waveform that is estimated to occur during the period when the vehicle is passing over the bridge. The information processing device detects one or more axle waveforms from the time range of the candidate vehicle waveform in the second time series data that are estimated to occur at the timing when the axle of the vehicle passes the position on the bridge where the sensor device is installed. The information processing device determines, based on the one or more axle waveforms, whether the candidate vehicle waveform is a passing waveform generated when the vehicle passes over the bridge. Information processing methods.

17. Information processing equipment, An acquisition unit that acquires time-series data of displacement amount representing the time change of the displacement amount of the bridge detected by a sensor device installed on the bridge, A first preprocessor generates first time series data by low-pass filtering the displacement time series data with a first low-pass filter of a first cutoff frequency, A second preprocessor generates second time-series data by low-pass filtering the aforementioned displacement time-series data with a second low-pass filter having a second cutoff frequency higher than the first cutoff frequency. A candidate waveform detection unit detects candidate vehicle waveforms from the first time-series data that are estimated to occur during the period when the vehicle is passing over the bridge, An axle waveform detection unit detects one or more axle waveforms that are estimated to occur at the time the axle of the vehicle passes the position on the bridge where the sensor device is installed, based on the time range of the candidate vehicle waveform in the second time series data. A vehicle passage determination unit determines whether the candidate vehicle waveform is a passage waveform generated when the vehicle passes over the bridge, based on the one or more axle waveforms mentioned above. A program that makes something work.

Images