Data processing system, data processing device, and data processing method

Abstract

This data processing system comprises an acquisition unit, a detection unit, and an extrapolation unit. The acquisition unit acquires time-series data of measurement values measured at a plurality of points including a first point and a second point. The detection unit detects missing time-series data acquired from the first point. When it is detected that the time-series data of the first point is missing, the extrapolation unit inserts data into the missing time-series data of the first point on the basis of a feature between the time-series data of the first point and the time-series data obtained by shifting the time of the second point.

Description

Data processing system, data processing device and data processing method

[0001] The present disclosure relates to a data processing system, a data processing device, and a data processing method.

[0002] There is known a technique for estimating a measurement value at an arbitrary point in a target space from measurement values ​​measured by sensors placed at multiple points. For example, in Patent Document 1, a measurement value at an arbitrary point is interpolated from the measurement values ​​at multiple points using the Kriging method.

[0003] Patent No. 6159594

[0004] By using an interpolation method such as that described in Patent Document 1, it is possible to estimate a measurement value at any point from measurement values ​​measured at multiple points. However, Patent Document 1 assumes that measurements used for interpolation are obtained from multiple sensors, and does not take into account cases where a measurement value at a certain point cannot be obtained. For example, if a measurement value at a certain point is missing, it is necessary to extrapolate the missing measurement value, but related techniques such as those described in Patent Document 1 are unable to extrapolate data with high accuracy.

[0005] In view of the above problems, one object of the present disclosure is to provide a data processing system, a data processing device, and a data processing method that are capable of extrapolating data with high accuracy.

[0006] A data processing system according to one aspect of the present disclosure includes an acquisition means for acquiring time series data of measurements taken at a plurality of locations including a first location and a second location, a detection means for detecting missing data in the time series data acquired from the first location, and an extrapolation means for, when missing data in the time series data of the first location is detected, inserting data into the missing time series data of the first location based on characteristics between the time series data of the first location and the time-shifted time series data of the second location.

[0007] A data processing device according to one aspect of the present disclosure includes an acquisition means for acquiring time series data of measurements taken at a plurality of locations including a first location and a second location, a detection means for detecting missing data in the time series data acquired from the first location, and an extrapolation means for, when missing data in the time series data of the first location is detected, inserting data into the missing time series data of the first location based on characteristics between the time series data of the first location and the time-shifted time series data of the second location.

[0008] A data processing method according to one aspect of the present disclosure acquires time series data of measurements taken at a plurality of locations including a first location and a second location, detects missing time series data acquired from the first location, and, if missing time series data of the first location is detected, inserts data into the missing time series data of the first location based on characteristics between the time series data of the first location and time-shifted time series data of the second location.

[0009] According to the present disclosure, data can be extrapolated with high accuracy.

[0010] 1 is a diagram illustrating a received power correction method. FIG. 2 is a diagram illustrating an example of a received power measurement value being missing. FIG. 3 is a configuration diagram illustrating an example of a data processing system according to some embodiments. FIG. 4 is a configuration diagram illustrating an example of a data processing device according to some embodiments. FIG. 5 is a flowchart illustrating an example of a data processing method according to some embodiments. FIG. 6 is a diagram illustrating a specific example of an extrapolation method according to some embodiments. FIG. 7 is a diagram illustrating a specific example of an extrapolation method according to some embodiments. FIG. 8 is a configuration diagram illustrating an example of an operation management system according to some embodiments. FIG. 9 is a configuration diagram illustrating an example of a wireless information transmission function according to some embodiments. FIG. 10 is a configuration diagram illustrating an example of an operation management server according to some embodiments. FIG. 11 is a configuration diagram illustrating an example of a received power map correction unit according to some embodiments. FIG. 12 is a flowchart illustrating an example of operation of the operation management server according to some embodiments. FIG. 13 is a diagram illustrating a received power map generation process according to some embodiments. FIG. 14 is a diagram illustrating a received power map generation process according to some embodiments. FIG. 15 is a diagram illustrating a received power map generation process according to some embodiments. FIG. 16 is a diagram illustrating a received power map generation process according to some embodiments. FIG. 1 is a diagram for explaining a received power map correction process according to some embodiments. FIG. 2 is a configuration diagram showing an example configuration of an operation management system according to some embodiments. FIG. 3 is a configuration diagram showing an example configuration of a received power map correction unit according to some embodiments. FIG. 4 is a flowchart showing an example operation of the received power map correction unit according to some embodiments. FIG. 5 is a flowchart showing an example operation of the received power map correction unit according to some embodiments. FIG. 6 is a diagram for explaining a received power map correction process according to some embodiments. FIG. 7 is a configuration diagram showing an overview of hardware of a computer according to some embodiments.

[0011] Hereinafter, embodiments will be described with reference to the drawings. In the drawings, the same elements are denoted by the same reference numerals, and redundant description will be omitted as necessary.

[0012] (Studies Leading to the Embodiments) The inventors have studied a method for correcting a reception power map showing the reception power distribution in a target space where a wireless communication service is provided, using measurements taken by a plurality of sensors.

[0013] Fig. 1 shows a received power correction method that the inventor has studied. In the example of Fig. 1, sensors S2 and S3 are placed at observation points P1 and P2, respectively. The sensors S2 and S3 receive radio waves emitted from base station B1 and measure the received power (received level) R ２ and R ３ At this time, the received power R ２ and R ３ In this case, the received power RSS is expressed by the following equation (1).

[0014] α and β in equation (1) are the received power R ２ and R ３ α and β are weighting coefficients of the distance d between the sensor S2 and the target point P0. ２ and the distance d between the sensor S3 and the target point P0 ３ This depends on the relationship of the following equation (2).

[0015] From the relationship of equation (2), when equation (1) is converted into an equation using only the coefficient α, the received power RSS at the target point P0 is given by the following equation (3).

[0016] In the reception power correction method of Fig. 1, first, a reception power map is generated in advance from the measured values ​​of the target space, and then a coefficient α at each point is calculated from the reception power at each point in the generated reception power map using equation (3). Then, the reception power R measured by sensors S2 and S3 at time t is calculated using the following equation (4). ２ and R ３The corrected received power RSS of the target point is calculated from the above. By correcting the received power of each point on the received power map, the received power map can be updated in real time.

[0017] Furthermore, the inventors have investigated the case where the number of sensors used in the received power correction method of FIG. 1 is increased to three or more. In this case, the received power at the target point can be corrected using the measured values ​​of three or more observation points. For example, when the received power R measured by three sensors S2 to S4 is ２ ~R ４ When using the above formula, the reception level RSS at the target point is expressed by the following formula (5): ４ is the distance between the sensor S4 and the target point P0.

[0018] As can be seen from equations (3) and (5), if any one of the measurements from the multiple sensors used is missing, the calculation of the equation does not hold. Therefore, with the received power correction method using equations (3) and (5), if a measurement value is missing, the received power at the target point cannot be determined, and a corrected received power map cannot be obtained. For example, due to reasons such as malfunction of the received power acquisition means or sensors in the multiple sensors, it is not always possible to acquire information from all sensors. In particular, as the number of sensors used increases, there is a high possibility that a specific value cannot be obtained from one of the sensors.

[0019] To solve this problem, the inventors have investigated a method of extrapolating the measurements of missing sensors. When extrapolating the measurements of missing sensors, unless appropriate candidates for extrapolation are selected, the difference between the extrapolated results and the true values ​​will become large, and the accuracy of the corrected received power map will deteriorate.

[0020] For example, one possible simple extrapolation method is to extrapolate using the value from one period prior to the missing value. Note that "period" is the unit of time for data contained in time series data. For example, the missing measurement value a(t) is corrected to the measurement value from one period prior, a(t-1). This method does not pose a problem as long as there is no fluctuation in the measurement value from one period prior. Another example of a simple extrapolation method is to extrapolate using the measurement value or fluctuation of the physically closest sensor. This method does not pose a problem as long as the measurement value of the missing sensor and the measurement value of the closest sensor are completely synchronized.

[0021] However, measurement values ​​may vary significantly depending on the environment, and the measurement values ​​of the nearest sensors may not be completely synchronized. In the example of Figure 2, sensors S2 and S3 are measuring radio waves from base station B1, and an obstacle X passes between sensors S2 and S3 and base station B1. The obstacle X moves from the side of sensor S2 to the side of sensor S3, parallel to the direction in which sensors S2 and S3 are lined up. For example, if the obstacle X is a large moving object such as a truck, when the obstacle X passes in front of each sensor, the sensor cannot receive radio waves from base station B1 and cannot notify the measurement values, resulting in missing measurement values ​​of the received power.

[0022] For example, when an obstruction X passes in front of sensor S3, the measurement value of the received power of sensor S3 is missing. In this case, if extrapolation is performed using the measurement value of sensor S3 from one period before the missing value, the correction will be made using the measurement value of a situation where the sensor is not obstructed. As a result, the obstructed situation disappears from the time series of measurement values, and accurate extrapolation is not possible. In addition, in this case, the measurement value of sensor S2, which is close to sensor S3, is not synchronized. In other words, when sensor S3 is obstructed by obstruction X, sensor S2 is not obstructed. Therefore, if the measurement value of sensor S3 is missing and the measurement value of sensor S2, which is close to sensor S3, is used for extrapolation, the obstructed situation also disappears from the time series of measurement values, and accurate extrapolation is not possible. In this way, simple extrapolation when a sensor measurement value is missing will not result in accurate extrapolation.

[0023] (First Embodiment) Next, a first embodiment will be described. Fig. 3 shows an example of the configuration of a data processing system 10 according to some embodiments. The data processing system 10 is a system that generates and corrects a wireless communication quality map, such as a received power map, in a target space.

[0024] 3 , the data processing system 10 includes an acquisition unit 11, a detection unit 12, and an extrapolation unit 13. The acquisition unit 11 acquires time-series data of measurement values ​​measured at a plurality of points including a first point and a second point. For example, in order to correct a wireless communication quality map of a target space, sensors are placed at a plurality of points in the target space, and the acquisition unit 11 acquires time-series data of measurement values ​​measured by the sensors at each point.

[0025] For example, the measurement value is wireless communication quality information used to correct the wireless communication quality map. Wireless communication quality includes wireless quality related to radio waves or communication quality related to communication between communication devices. For example, wireless quality is an index indicating the received power or radio wave strength of radio waves, such as RSRP (Reference Signal Received Power), RSRQ (Reference Signal Received Quality), RSSI (Received Signal Strength Indicator), or SINR (Signal to Interference plus Noise power Ratio). For example, communication quality is an index indicating traffic quality, such as throughput or delay time.

[0026] The detection unit 12 detects a lack of time-series data of measurement values ​​acquired from a first location included in the multiple locations. A lack of time-series data means that data that should be arranged in time series at a predetermined interval cannot be obtained at the predetermined interval. For example, the detection unit 12 may detect a lack of time-series data of a specific value at the first location when measurement values ​​cannot be acquired from a sensor at the first location at the predetermined interval.

[0027] When the extrapolation unit 13 detects a missing piece of time series data of measurement values ​​at a first location, it inserts data into the missing piece of time series data at the first location. Inserting (or interpolating) data into the missing piece of time series data is called extrapolation. When the time series data at the first location is missing, the extrapolation unit 13 extrapolates the time series data at the first location based on features between the time series data at the first location and the time-shifted time series data at the second location. For example, the features between the time series data may be correlation, distance, similarity, etc. between the data.

[0028] The data processing system 10 may also include a calculation unit that calculates a characteristic indicating the relationship between a feature value indicating the characteristics between the time series data of a first location and the time-shifted time series data of a second location and the amount of time shift. For example, the calculation unit calculates this characteristic for each combination of multiple locations. The amount of time shift (shift) of the time series data is called lag. For example, the calculated characteristic may be a correlogram indicating the relationship between the lag and the cross-correlation of the time series data. The extrapolation unit 13 may identify data to be used for extrapolating the time series data of the first location from the time series data of the second location based on the calculated characteristic, such as the correlogram between the time series data of the first location and the time series data of the second location.

[0029] Furthermore, the data processing system 10 may include an identification unit that identifies a first time shift amount at which the feature value increases in the calculated characteristics. For example, the first time shift amount may be a peak lag at which the correlation coefficient is highest in a correlogram. The extrapolation unit 13 may identify data to be used for extrapolation from the time series data of the second location according to the identified first time shift amount. For example, the extrapolation unit 13 may extrapolate the time series data of the first location using measured values ​​of the second location at a time corresponding to the first time shift amount, or may extrapolate the time series data of the first location using fluctuations in measured values ​​of the second location at a time corresponding to the first time shift amount.

[0030] Furthermore, the extrapolation unit 13 may select data to be used for extrapolation depending on whether or not there is a fluctuation in the time series data of the second location corresponding to the first time shift amount. For example, if there is a fluctuation in the time series data of the second location corresponding to the first time shift amount, the extrapolation unit 13 may identify data to be used for extrapolating the time series data of the first location based on the time series data of the second location corresponding to the first time shift amount. If there is no fluctuation in the time series data of the second location corresponding to the first time shift amount, the extrapolation unit 13 may identify data to be used for extrapolating the time series data of the first location based on the time series data before the loss at the first location. For example, the extrapolation unit 13 may extrapolate the time series data of the first location using measured values ​​of the first location one period before, or may extrapolate the time series data of the first location using fluctuations in measured values ​​of the first location one period before.

[0031] The data processing system 10 may be configured by one device or by multiple devices. Fig. 4 shows an example configuration of a data processing device 20 according to some embodiments. In the example of Fig. 4, the data processing device 20 includes the acquisition unit 11, detection unit 12, and extrapolation unit 13 shown in Fig. 3. For example, the acquisition unit 11, detection unit 12, and extrapolation unit 13 may all be located in a management server or the like, or may be distributed across multiple devices.

[0032] 5 illustrates a data processing method according to some embodiments. For example, the data processing method according to some embodiments may be performed by the data processing system 10 of FIG. 3 or the data processing device 20 of FIG.

[0033] 5, the acquisition unit 11 acquires time-series data of measurement values ​​measured at a plurality of points (S11). For example, the acquisition unit 11 acquires time-series data of measurement values ​​such as received power from sensors at a plurality of points.

[0034] Next, the detection unit 12 detects a missing time series data of measurement values ​​acquired from a first location included in the multiple locations (S12). When a missing time series data of measurement values ​​from the first location is detected, the extrapolation unit 13 extrapolates the time series data of the first location (S13). The extrapolation unit 13 extrapolates the time series data of the first location based on characteristics between the time series data of the first location and time-shifted time series data of the second location. For example, in a correlogram obtained from the time series data of the first location and the time series data of the second location, the peak lag with the highest correlation coefficient is identified, and the time series data of the first location is extrapolated based on the time series data of the second location at the identified peak lag.

[0035] Specific examples of extrapolation methods according to some embodiments will be described using FIGS. 6 to 8. In this example, extrapolation is performed using a correlogram. FIG. 6 shows an example in which the extrapolation method according to some embodiments is applied to the situation shown in FIG. 2. It is assumed that the path and speed of a large object that has a particularly large impact on the received power are predetermined. For example, the path traveled by an AGV (Automated Guided Vehicle) in a factory or the path traveled by heavy machinery at a construction site are often predetermined. In particular, the larger the moving object, the more likely it is that the path it can travel will be limited.

[0036] In the example of Figure 6, similar to Figure 2, sensors S2 and S3 are measuring radio waves from base station B1, and an obstruction X passes between sensors S2 and S3 and base station B1 from the side of sensor S2 to the side of sensor S3.

[0037] When an obstacle X passes in front of sensor S3, the measurement value of the received power of sensor S3 is missing, so the received power of sensor S3 is extrapolated. In the example of Figure 6, for example, a correlogram of cross-correlation is used to find the peak lag k with the highest correlation coefficient. A correlogram of cross-correlation between the time series data of the received power of sensor S2 and the time series data of the received power of sensor S3 is generated, and the peak lag k with the highest correlation coefficient in the correlograms of sensors S2 and S3 is found.

[0038] 6, the fluctuations in the time-series data of the received power of sensor S2 k periods ago are similar to the fluctuations in the time-series data of the received power of sensor S3 at present (the correlation coefficient is high). Therefore, when the received power of sensor S3 is missing, the received power of sensor S3 is extrapolated using the fluctuations in the received power of sensor S2 k periods ago.

[0039] As shown in Fig. 6, the peak lag k with the highest correlation coefficient is found in the correlograms of sensors S2 and S3. Next, when the received power of sensor S3 is missing, the fluctuation of the received power of sensor S2 k periods before is used to extrapolate the missing received power, as shown in Fig. 7. For example, when the received power of sensor S3 is missing at t0, the received power of sensor S2 k periods before t0 is used to extrapolate the missing received power of sensor S3.

[0040] Fig. 8 shows an example different from Fig. 6, in which sensors S2 to S4 are arranged. The sensors are arranged in an L-shape in the order of S2, S3, S5, and S4. The sensors S3, S5, and S4 are arranged in a direction perpendicular to the direction in which the sensors S2 and S3 are arranged. In the example of Fig. 8, the obstruction X passes between the sensors S3, S5, and S4 and the base station B1 from the side of sensor S4 to the side of sensor S3.

[0041] 8, a correlogram of the cross-correlation of the time-series data of the received power is generated for all combinations of sensors S2 to S4, and the peak lag k with the highest correlation coefficient is found in the correlogram for each combination. For example, a correlogram of sensors S2 and S3, a correlogram of sensors S3 and S5, and a correlogram of sensors S3 and S4 are generated.

[0042] When an obstruction X passes through a path as shown in FIG. 8 , the correlation coefficient between sensors S2 and S3 is not high, as shown in the correlograms of sensors S2 and S3. Therefore, if the received power of sensor S3 is lost, the correlograms of sensors S2 and S3 are not used. For example, the correlogram of sensor S4 (correlograms of S3 and S4), which is affected by the obstruction in the same way as sensor S3, is used. The correlogram of sensor S5 (correlograms of S3 and S5), which is similarly affected, may also be used. For example, if the received power of sensor S3 is lost, the peak lag k a with the highest correlation coefficient in the correlograms of sensors S3 and S4 is identified, and the received power of sensor S4 identified ka periods before is used to extrapolate the missing received power of sensor S3. Furthermore, if the received power of sensor S3 is lost, the peak lag k b with the highest correlation coefficient in the correlograms of sensors S3 and S5 is identified, and the received power of sensor S5 identified kb periods before is used to extrapolate the missing received power of sensor S3.

[0043] As described above, in this embodiment, time series data of measurement values ​​measured by sensors at multiple locations is acquired. When time series data at a first location is missing, the time series data at the first location is extrapolated based on characteristics between the time series data at the first location and time-shifted time series data at a second location. For example, the time series data at the first location is extrapolated using time series data at a second location k periods before the peak lag, which has the highest correlation coefficient in the correlogram. This allows for accurate extrapolation even when data is missing due to a moving obstruction, as in FIGS. 6 and 8 . Therefore, when correcting a wireless communication quality map, such as a received power map, from sensor measurement values, the extrapolated measurement values ​​can be used to accurately correct the wireless communication quality map.

[0044] (Embodiment 2) Next, embodiment 2 will be described. In this embodiment, a specific example of embodiment 1 will be described. In this embodiment, when correcting the received power map, a correlogram of cross-correlation is used to extrapolate missing received power. Note that this embodiment can be implemented in combination with embodiment 1, and each configuration shown in embodiment 1 may be used as appropriate.

[0045] FIG. 9 illustrates an exemplary configuration of an operation management system 1 according to some embodiments. The operation management system 1 is a system that manages the operational status of an application terminal (not shown), such as an AGV, and manages the wireless quality and communication quality of a wireless communication environment (also referred to as a target wireless communication environment) in which the application terminal performs wireless communication. The application terminal is a mobile terminal, such as an AGV, an AMR (Autonomous Mobile Robot), an autonomous robot, a self-driving car, or a drone. The target wireless communication environment is a wireless communication environment enabled by radio waves transmitted from a predetermined transmission source. For example, the transmission source is a device that forms a communication area enabling wireless communication, and may be a base station for mobile communication or an access point for a wireless local area network (LAN). Furthermore, in a situation in which the terminal transmits radio waves, the terminal may also be the transmission source.

[0046] 9, the operation management system 1 includes an operation management server 100, an initial measurement device 200, a plurality of sensors 300, and a base station 400. The number of initial measurement devices 200 is not limited to one, and may be multiple. Furthermore, one device may include the functions of the initial measurement device 200, the sensors 300, the application terminal, and the like.

[0047] The initial measurement device 200 and the sensor 300 are connected to the base station 400 so as to be able to communicate wirelessly. The initial measurement device 200 and the sensor 300 may be connected to the base station 400 so as to be able to communicate wirelessly in both directions, or so as to be able to communicate wirelessly in at least one direction, either uplink or downlink. For example, they may be connected to be able to communicate via a wireless network such as 4G, local 5G / 5G, LTE (Long Term Evolution), or wireless LAN. Furthermore, the base station 400 and the operation management server 100 may be connected to be able to communicate wired or wirelessly. For example, they may be connected to be able to communicate via various networks such as a core network, the Internet, a WAN (Wide Area Network), or a LAN, or a combination thereof.

[0048] The initial measurement device 200 is an example of a device that measures the wireless quality of a target wireless communication environment. It measures the wireless quality of the target wireless communication environment before the application terminal is put into operation. For example, the operation management server 100 generates an initial reception power map (first reception power map) from the wireless quality measured before operation. Before the operation of the application terminal, for example, in the case of an AGV, this may be early morning before warehouse operations begin. However, the timing for measuring the wireless quality of the target wireless communication environment is not limited to before operation. It may be a time period when there are few mobile devices, such as a lunch break or an evening break, or any time when the wireless communication quality map needs to be reset. The initial measurement device 200 is a mobile terminal, such as a mobile phone, a smartphone, a tablet terminal, or an IoT (Internet of Things) terminal. The initial measurement device 200 measures the reception power of radio waves in the target wireless communication environment at each location to which it moves and transmits the measured reception power and location information to the operation management server 100 via the base station 400. Furthermore, the initial measurement device 200 is not limited to a mobile terminal; multiple immobile terminals may also be used as the initial measurement device 200. In this case, the multiple initial measurement devices 200 measure the received power of radio waves at each of the installed locations and transmit the measured received power and location information. For example, the initial measurement devices 200 transmit wireless information including the measured received power and location information using a wireless information transmission function described below. The initial measurement devices 200 may transmit the received power and location information, or may transmit only the received power. For example, the initial measurement devices 200 may perform measurements at preset locations and transmit the measured received power.

[0049] The sensor 300 is an example of a device that measures the wireless quality of a target wireless communication environment and measures the wireless quality of the target wireless communication environment while the application terminal is in operation. For example, the operation management server 100 generates a received power map for display from the wireless quality measured during this operation. The sensor 300 is a device that measures wireless quality at a predetermined observation point and is a fixed observation device whose position is fixed at the predetermined observation point. The sensor 300 is not limited to a fixed observation device, and may be a mobile observation device that can move to the predetermined observation point. The position of the observation point is not limited to a predetermined position and may be, for example, a position (measurement point) where the sensor 300 measures and reports location information. The sensor 300 is not limited to the sensor 300 and may be, for example, an IoT terminal, a base station, an access point, etc. The sensor 300 may be the same device as the initial measurement device 200. For example, the initial measurement device 200 may be the sensor 300 and measure the wireless quality at a predetermined observation point. The sensor 300 measures the received power of radio waves in a target wireless communication environment at a predetermined observation point and transmits the measured received power to the operation management server 100 via the base station 400. For example, the sensor 300 transmits wireless information including the measured received power using a wireless information transmission function described below. The sensor 300 may transmit wireless information including the measured received power and location information.

[0050] The base station 400 is a base station device in a wireless network between the initial measurement device 200 and the sensor 300. The base station 400 also serves as a relay device that relays communications between the initial measurement device 200, the sensor 300, and the operation management server 100. For example, the base station 400 is a local 5G base station, a 5G next generation Node B (gNB), an LTE eNB (evolved Node B), a wireless LAN access point, a Bluetooth (registered trademark) base station, or the like, but may also be another relay device. The base station 400 transfers the received power of the target wireless communication environment measured by the initial measurement device 200 and the sensor 300 to the operation management server 100. Therefore, for example, the base station 400 is a device separate from the source of the target wireless communication environment, but may also be the same device as the source of the target wireless communication environment.

[0051] The operation management server 100 is a server that manages the operation status of the application terminal and manages the wireless quality and communication quality of the target wireless communication environment. The operation management server 100 generates a received power map showing the received power distribution of the target wireless communication environment based on wireless information received from either the initial measurement device 200 or the sensor 300, and displays the generated received power map. In addition to the received power map, wireless quality maps showing the distribution of other wireless qualities may also be generated and displayed.

[0052] 10 shows an example of the configuration of the wireless information transmission function 210 according to some embodiments. The wireless information transmission function 210 is included in the initial measurement device 200 and the sensor 300. In the example of FIG. 10, the wireless information transmission function 210 includes a received power measurement unit 211, a position detection unit 212, and a wireless information transmission unit 213.

[0053] The received power measurement unit 211 measures the received power of radio waves in the target wireless communication environment. The received power measurement unit 211 receives radio waves transmitted from a transmission source in the target wireless communication environment and measures the received power (radio wave intensity) of the received radio waves. The received power is, for example, RSRP, RSRQ, RSSI, SINR, etc., or may be another index indicating wireless quality. The received power measurement unit 211 outputs received power information indicating the measured received power.

[0054] The position detection unit 212 detects the current position of the device (initial measurement device 200 and sensor 300) including the wireless information transmission function 210. For example, the position detection unit 212 may detect the position using a Global Navigation Satellite System (GNSS) such as the Global Positioning System (GPS), or may detect the position by analyzing video from a camera installed in a location overlooking a predetermined area. The position detection unit 212 may also detect the position of the device including the wireless information transmission function 210 based on information from a beacon, short-range wireless communication, or reflected light from a laser beam such as Lidar. The position detection unit 212 may also detect the position using a delay time in wireless communication. The wireless communication used for these purposes may be light such as visible light or infrared light, or radio waves with no frequency restrictions. The position may also be detected using other detection methods. The position detection unit 212 outputs position information indicating the detected position. Note that the sensor 300 does not necessarily have to be provided with the position detection unit 212 because measurements are performed at the position of a predetermined observation point. The same applies when the initial measurement device 200 performs measurement at a predetermined position.

[0055] The wireless information transmitter 213 transmits wireless information including the received power information measured by the received power measurement unit 211 and the location information detected by the location detection unit 212 to the operation management server 100 via the base station 400. In the case of the sensor 300, the wireless information transmitter 213 may transmit only the received power information as wireless information. The same applies when the initial measurement device 200 performs measurement at a predetermined location. The wireless information transmitter 213 is a communication interface capable of communicating with the base station 400, and is, for example, a wireless interface such as 4G, local 5G / 5G, LTE, or wireless LAN, but may also be a wireless interface of any other communication method.

[0056] FIG. 11 shows a configuration example of the operation management server 100 according to some embodiments. Note that this configuration is an example, and other configurations may be used as long as the operations of the operation management server 100 described later are possible. For example, part of the configuration of the operation management server 100 may be placed in another device, or the configuration of each part of the operation management server 100 may be distributed and placed in a plurality of devices.

[0057] In the example of FIG. 11 , the operation management server 100 includes a radio information receiving unit 110, a radio information storage unit 120, a received power map generating unit 130, a received power map storage unit 140, a received power map correcting unit 150, and a display unit 160.

[0058] The wireless information receiving unit 110 receives wireless information transmitted from the initial measurement device 200 and the sensor 300 via the base station 400. The wireless information receiving unit 110 is an acquiring unit that acquires wireless information including received power information and location information transmitted from the initial measurement device 200 and the sensor 300. The wireless information receiving unit 110 may acquire wireless information including only received power information from the sensor 300. For example, the wireless information receiving unit 110 corresponds to the acquiring unit 11 in FIG. 3. The wireless information receiving unit 110 is a communication interface capable of communicating with the Internet, a core network, etc., and is, for example, a wired interface for IP communication, but may also be a wired or wireless interface for any other communication method.

[0059] The wireless information storage unit 120 is a database that stores the wireless information received by the wireless information receiving unit 110. The wireless information storage unit 120 accumulates the wireless information acquired from the initial measurement device 200 and the sensor 300 as time-series data. The wireless information receiving unit 110 may output the received wireless information to the received power map generating unit 130 or the like.

[0060] The reception power map generating unit 130 generates a reception power map based on the wireless information received by the wireless information receiving unit 110 from the initial measurement device 200 and the sensor 300 and stored in the wireless information storage unit 120. The reception power map generating unit 130 generates a reception power map showing the reception power distribution in a predetermined area of ​​the target wireless communication environment based on the reception power information and location information included in the wireless information. For example, the predetermined area is a floor where the application terminal moves during operation.

[0061] The reception power map generating unit 130 may generate a reception power map based on radio information acquired from either the initial measurement device 200 or the sensor 300. For example, before operation, the reception power map generating unit 130 generates an initial reception power map based on radio information from each point measured by the initial measurement device 200 as it moves, and stores the generated reception power map in the reception power map storage unit 140. The reception power map generating unit 130 may output the generated initial reception power map to the display unit 160.

[0062] The received power map storage unit 140 stores the initial received power map generated by the received power map generation unit 130. The received power map storage unit 140 also stores the corrected received power map corrected by the received power map correction unit 150.

[0063] The reception power map correction unit 150 corrects, during operation, the initial reception power map that was generated by the reception power map generation unit 130 before operation and stored in the reception power map storage unit 140. During operation, the reception power map correction unit 150 corrects the initial reception power map based on radio information at each point measured by the sensor 300 at the observation point, and stores the corrected reception power map in the reception power map storage unit 140. The reception power map correction unit 150 may output the corrected reception power map to the display unit 160. Furthermore, during operation, the reception power map correction unit 150 may update the reception power map stored in the reception power map storage unit 140 (displayed on the display unit 160) as needed, based on radio information measured by the sensor 300 at the observation point.

[0064] The display unit 160 generates display information for displaying processing results, etc., of each unit of the operation management server 100. The display unit 160 may include a display device such as a liquid crystal display or an organic EL display, and may display the generated display information. The display unit 160 may also output the generated display information to a display device external to the operation management server 100 and display it on the external display device. For example, the display unit 160 displays the generated display information on an operation management screen using a GUI (Graphical User Interface). The display information includes information for displaying a received power map, etc. The display unit 160 displays, on the operation management screen, the received power map generated by the received power map generation unit 130 and the received power map corrected by the received power map correction unit 150. The display unit 160 may obtain information on a floor map indicating a specific floor from an internal storage unit of the operation management server 100 or an external storage device, and display the received power map on the floor map.

[0065] Fig. 12 shows an example configuration of the received power map correction unit 150 in the operation management server 100 according to some embodiments. In the example of Fig. 12, the received power map correction unit 150 includes a lag calculation unit 151, a missingness determination unit 152, a fluctuation detection unit 153, an extrapolation candidate determination unit 154, an extrapolation unit 155, a correction formula generation unit 156, and a correction calculation unit 157. Note that this configuration is just an example, and other configurations may be used as long as the operation of the received power map correction unit 150 described below is possible.

[0066] The lag calculation unit 151 calculates a correlogram of cross-correlation for all combinations of sensors 300 from the time series data of received power information included in the wireless information acquired from the sensors 300. The lag calculation unit 151 also functions as a correlogram calculation unit. The lag calculation unit 151 calculates a peak lag k at which the correlation coefficient is highest in the correlogram for each combination of sensors 300. The peak lag at which the correlation coefficient is highest for the time series data of sensor n (any of the sensors 300) and sensor m (any of the sensors 300 other than sensor n) is represented as k(n, m).

[0067] When generating a correlogram, the correlation coefficient may be calculated using the data as is, or may be calculated using values ​​transformed by Fourier transform or transformation using support vectors. As an example, a correlogram is created by calculating the correlation coefficient between data, but other indices relating to the similarity of data may also be calculated, not limited to the correlation coefficient. For example, the distance or similarity between data may be calculated. The distance between data may be the sum of squared errors, the Malanobis distance, or the distance using the center of gravity of the data.

[0068] The loss determination unit 152 determines whether or not there is a loss in the time-series data of the received power information included in the wireless information acquired from the sensor 300. The loss determination unit 152 references the time-series data of the received power information of the sensor 300 stored in the wireless information storage unit 120, and determines whether or not there is a loss in the time-series data of the received power information. For example, the loss determination unit 152 corresponds to the detection unit 12 in FIG. 3 .

[0069] When it is determined that there is a gap in the time-series data of the received power information of the sensor 300, the fluctuation detection unit 153 detects a fluctuation in the time-series data of a sensor 300 (also referred to as a correlated sensor) that has a high correlation with the missing sensor 300 from a correlogram obtained by combining the sensor 300 with the missing time-series data (also referred to as a missing sensor). The fluctuation detection unit 153 selects a correlated sensor 300 from a correlogram including a lag having a high correlation coefficient with the missing sensor 300, and identifies a peak lag k from the correlogram of the correlated sensor 300. The fluctuation detection unit 153 (or the lag calculation unit 151) also serves as an identification unit that identifies the peak lag k. The fluctuation detection unit 153 detects a fluctuation near the identified peak lag (near k periods before) in the time-series data of the received power information of the selected correlated sensor 300. For example, when time-series data of sensor m is missing, the fluctuation detection unit 153 detects a fluctuation k(n, m) periods before the time-series data of sensor n.

[0070] When a fluctuation is detected in the time-series data of the correlation sensor 300 k periods before the peak lag, the extrapolation candidate determination unit 154 selects the correlation sensor 300 that detected the fluctuation as an extrapolation candidate. When a fluctuation is detected k periods before the peak lag in multiple correlation sensors 300, the extrapolation candidate determination unit 154 selects one of the sensors 300. For example, the priority of sensors with similar received power from the same base station may be increased. That is, the extrapolation candidate determination unit 154 may select, as an extrapolation candidate, a sensor 300 with received power closest to that of the missing sensor 300 from among sensors with fluctuations k periods before. The sensor 300 with received power closest in physical distance to the missing sensor 300 may also be selected as an extrapolation candidate. Furthermore, the priority of sensors with high correlation coefficients may be increased. The extrapolation candidate determination unit 154 may select, as an extrapolation candidate, a sensor 300 with the highest correlation coefficient of the correlogram with the missing sensor 300 from among sensors with fluctuations k periods before.

[0071] The extrapolation unit 155 extrapolates the time series data of the missing sensor 300 using the time series data of the sensor 300 selected as an extrapolation candidate (also referred to as a candidate sensor). For example, the extrapolation unit 155 corresponds to the extrapolation unit 13 in FIG. 3 . The extrapolation unit 155 extrapolates the time series data of the missing sensor 300 using the received power or the fluctuation of the received power k periods before the peak lag of the candidate sensor 300. For example, if time series data is missing from sensor m and there is a fluctuation k(n, m) periods before the time series data of sensor n, the time series value data of the missing sensor m is extrapolated using the received power or the fluctuation of the received power k(n, m) periods before the time series data of sensor n.

[0072] Furthermore, if no fluctuation is detected in the time-series data of the correlated sensor 300 before the peak lag k periods, the extrapolation candidate determination unit 154 may select the missing sensor 300 or a sensor 300 close to the missing sensor 300 as an extrapolation candidate. In this case, the extrapolation unit 155 may extrapolate the time-series data of the missing sensor 300 using the received power or fluctuations in the received power of the missing sensor 300 one period before. The extrapolation unit 155 may extrapolate the time-series data of the missing sensor 300 using the received power or fluctuations in the received power of a sensor 300 whose received power is closest to that of the missing sensor 300. Alternatively, the received power or fluctuations in the received power of a sensor 300 whose physical distance from the missing sensor 300 is closest may be used. Alternatively, the received power or fluctuations in the received power of a sensor 300 whose correlogram has the highest correlation coefficient with the missing sensor 300 may be used.

[0073] The correction formula generation unit 156 generates a correction formula for correcting the received power at each point of the received power map. For example, the correction formula generation unit 156 refers to the initial received power map stored in the received power map storage unit 140 and calculates the coefficient α in the correction formula of the above formula (3) or formula (5).

[0074] The correction calculation unit 157 corrects the received power at each point on the received power map from the received power and the extrapolated received power acquired from the sensor 300, using the correction formula generated by the correction formula generation unit 156. For example, the correction calculation unit 157 inputs the received power acquired from the sensor 300 and the extrapolated received power into the correction formulas of the above formula (3) and formula (5) to obtain the corrected received power.

[0075] In this example, when the received power from the sensor 300 is missing, the missing received power is extrapolated and the extrapolated received power is input into the correction formula to calculate the corrected received power. However, the corrected received power may be calculated without extrapolating the missing received power. That is, a correction formula that can calculate the corrected received power even if an input value is missing may be used. For example, a correction formula may be prepared for each combination of measurements from one or more sensors, and the correction formula to be used may be changed depending on the combination at that time. For example, in the above formula (5), a distance term may be generated for each combination of sensors, and α may be calculated for each combination of sensors.

[0076] FIG. 13 illustrates an example of the operation of the operation management server 100 according to some embodiments. In the example of FIG. 13 , the operation management server 100 acquires radio information from the initial measurement device 200 before operation (S101). For example, before operation, a measurer moves around the floor carrying the initial measurement device 200. The initial measurement device 200 measures the received power at multiple locations and transmits the measured received power information and location information to the operation management server 100. Instead of a measurer, a robot such as an AGV may carry the initial measurement device 200 and move around the floor, measuring the received power at each location. Alternatively, the initial measurement device 200 may be installed at multiple locations and measure the received power at multiple fixed positions. The radio information receiving unit 110 of the operation management server 100 receives radio information including the received power information and location information at multiple locations from the initial measurement device 200 and stores the received radio information in the radio information storage unit 120.

[0077] For example, as shown in FIG. 14 , the initial measurement device 200 moves through the floor, measures the received power at multiple initial measurement points along the movement path, and stores radio information including the location information and received power information of the multiple initial measurement points in the radio information storage unit 120. The received power map generation unit 130 acquires the radio information of the multiple initial measurement points stored in the radio information storage unit 120. The intervals between the initial measurement points may be equal as shown in FIG. 14 or may be different. For example, measurements may be made at narrow intervals in areas where the received power changes significantly, and at wider intervals in areas where the received power changes less significantly. Measurements may be made at narrow intervals in areas close to the source of transmission and at wider intervals in areas far from the source of transmission. Measurements may be made at narrow intervals when the mobile terminal is moving slowly, and at wider intervals when the mobile terminal is moving quickly. Measurements may be made at narrow intervals in areas where application terminals use the device frequently, and at wider intervals in areas where application terminals use the device less frequently.

[0078] Next, the operation management server 100 generates an initial received power map (S102) based on the radio information acquired from the initial measurement device 200. The received power map generator 130 generates an initial received power map showing the received power distribution before operation based on the plurality of pieces of radio information including the position information and the received power information stored in the radio information storage unit 120.

[0079] For example, as shown in FIG. 15 , a reception power map showing the reception power distribution of the entire floor is generated from reception power information of a large number of initial measurement points. The reception power map generation unit 130 generates the reception power map of the entire floor by estimating reception power information of each point other than the initial measurement point (which may include the initial measurement point) based on the reception power information of the initial measurement point. The estimation method may be spatial interpolation such as linear interpolation or kriging, or another method. The reception power map is data in which reception power is mapped to each position in a target space such as a floor, and can be represented as a heat map or contour lines corresponding to the reception power at each position. The reception power map generation unit 130 stores the generated reception power map in the reception power map storage unit 140.

[0080] Next, the operation management server 100 acquires wireless information from the sensor 300 during operation (S103). After generating an initial received power map before operation, during operation the sensor 300 measures received power at a predetermined observation point and transmits wireless information including the measured received power information to the operation management server 100. The wireless information receiving unit 110 stores the received wireless information including the received power information in the wireless information storage unit 120.

[0081] 16 , sensors 300 are placed at two observation points M1 and M2, the received power at observation points M1 and M2 is measured, and the received power map correction unit 150 acquires received power information at observation points M1 and M2. The position information of the observation points may be acquired from the sensors 300, or may be set in advance. The number of observation points (sensors 300) that perform observation during operation is not limited to two, and may be any number equal to or greater than two.

[0082] Next, the operation management server 100 corrects the initial reception power map based on the wireless information acquired from the sensor 300 (S104). The reception power map correction unit 150 corrects the reception power information of the entire initial reception power map using the reception power information of the observation point measured by the sensor 300 and the reception power information of the entire initial reception power map that has been generated and stored.

[0083] For example, the received power information of the initial received power map in Fig. 15 is corrected using the received power information of observation points M1 and M2 in Fig. 16 to generate a corrected received power map as shown in Fig. 17. The received power map correction unit 150 stores the corrected received power map in the received power map storage unit 140. Furthermore, during operation, steps S103 to S104 may be repeated periodically to update the stored received power map as needed.

[0084] Next, the operation management server 100 displays an operation management screen (S105). The display unit 160 displays the generated and stored reception power map of the specified area on the operation management screen. Fig. 18 shows an example of the operation management screen displayed by the display unit 160. In the example of Fig. 18, the operation management screen 500 includes a reception power map screen 510.

[0085] The display unit 160 displays the received power map stored in the received power map storage unit 140 on a received power map screen 510 of the operation management screen 500. The display unit 160 may display the initial received power map generated by the received power map generation unit 130, and may also display a received power map corrected (updated) by the received power map correction unit 150. This makes it possible to display in real time the received power distribution of the floor estimated from the received power measured at the observation point.

[0086] For example, a floor map may be displayed on the received power map screen 510, and the generated received power map may be displayed on the floor map. The floor map may be a map showing the positions and shapes of objects placed on the floor, or may be a photographed image of the floor. The received power map may be displayed as a heat map according to the received power at each point, or may be displayed using contour lines indicating the received power. When displayed as a heat map, the color or hatching pattern of each point may be changed depending on the received power. Furthermore, in the received power map, areas where the received power is lower than a predetermined threshold may be highlighted using a color or hatching pattern.

[0087] 16, the observation points of the sensors 300 may be displayed on the received power map using icons or the like. For example, when time-series data of the sensor 300 is missing, the fact that the data is missing may be displayed in an identifiable manner. The color, size, shape, etc. of the icon of the sensor 300 from which time-series data is missing may be changed.

[0088] 19 shows an example of the operation of the received power map correction unit 150 according to some embodiments. The example of the operation in FIG. 19 includes the radio information acquisition (S103) and received power map correction (S104) in FIG.

[0089] In the example of Fig. 19, the reception power map correction unit 150 acquires reception power information from the plurality of sensors 300 (S201). As in S103 of Fig. 13, during operation, the wireless information receiving unit 110 receives wireless information including reception power information measured and transmitted by the sensors 300, and stores the received wireless information including the reception power information in the wireless information storage unit 120. The reception power map correction unit 150 acquires time-series data of the reception power information of the plurality of sensors 300 stored in the wireless information storage unit 120. The reception power map correction unit 150 may acquire position information of the sensors 300 included in the wireless information, as necessary.

[0090] Next, the received power map correction unit 150 calculates a correlogram of cross-correlation for all combinations of sensors 300 (S202). For example, as shown in FIG. 20 , the lag calculation unit 151 selects two sensors 300, e.g., sensor n and sensor m, and acquires time series data of received power information from the selected sensors n and m. The lag calculation unit 151 calculates a correlogram of cross-correlation for the acquired time series data from sensors n and m. The lag calculation unit 151 calculates a correlation coefficient while shifting a window W of a predetermined period of each time series data on the time axis. That is, it calculates the correlation coefficient between the data of window W for sensor n and the data of window W for sensor m when the time is shifted. The amount by which this data is shifted on the time axis is called a lag. The correlation coefficient calculated for each lag becomes a correlogram. As shown in FIG. 20 , the correlogram is expressed as a graph showing the relationship between each lag and the correlation coefficient. The lag calculation unit 151 calculates a correlogram for all combinations of sensors 300.

[0091] 21 shows all combinations of correlograms when the number of sensors is N. For example, the lag calculation unit 151 calculates correlograms C(1,2) to C(1,N-1) for a combination of sensor 1 and another sensor, calculates correlograms C(2,1) to C(2,N-1) for a combination of sensor 2 and another sensor, ..., calculates correlograms C(N-1,1) to C(N-1,N-2) for a combination of sensor N-1 and another sensor.

[0092] Next, the received power map correction unit 150 calculates the peak lag k(n, m) at which the correlation coefficient is highest in each correlogram (S203). As shown in FIG. 20, the lag calculation unit 151 obtains the maximum value of the correlation coefficient of the time-series data in the correlograms of sensors n and m, and determines the lag at which the correlation coefficient is maximum as the peak lag k. As shown in FIG. 21, the peak lag k is calculated for the correlogram of each combination. For example, the lag calculation unit 151 calculates peak lags k(1,2) to k(1,N-1) for correlograms C(1,2) to C(1,N-1), calculates peak lags k(2,1) to k(2,N-1) for correlograms C(2,1) to C(2,N-1), ..., calculates peak lags k(N-1,1) to k(N-1,N-2) for correlograms C(N-1,1) to C(N-1,N-2).

[0093] Thereafter, the received power map correction unit 150 acquires received power information from the plurality of sensors 300 at the timing of correcting the received power map (S201), and determines whether or not there is a gap in the time-series data of the received power information for any of the sensors m (S204). The gap determination unit 152 refers to the time-series data of the received power information for the plurality of sensors 300 stored in the wireless information storage unit 120, and determines whether or not there is a gap in the received power information for each time-series data.

[0094] For example, when the received power map correction unit 150 periodically corrects the received power map, if the received power information (wireless information) is not received from the sensor 300 within the correction period of the received power map, the missing determination unit 152 may determine that the time-series data of the sensor 300 is missing. Furthermore, the missing determination unit 152 may calculate the average and variance of the received power information (wireless information) and the arrival period (reception period) from the sensor 300, and may determine that the time-series data of the sensor 300 is missing if the arrival period of the received power information from the sensor 300 is delayed so much that it is not within a predetermined x% (if the received power information is not received).

[0095] If the received power information is not missing from the multiple sensors 300, there is no need to extrapolate the time-series data, so the received power map correction unit 150 proceeds to S208 and corrects the received power map. If the received power information is missing from any of the sensors m, the time-series data is extrapolated from S205 onwards.

[0096] If received power information is missing for sensor m, the received power map correction unit 150 determines whether there is a fluctuation k(n, m) periods ago in the time-series data for sensor n (S205). When time-series data is missing for sensor m, the fluctuation detection unit 153 references correlograms c(m, 1) to c(m, N) obtained by combining the missing sensor m with N-1 sensors (excluding sensor m) from the N sensors. For example, as shown in FIG. 20 , the fluctuation detection unit 153 extracts a correlogram whose maximum correlation coefficient is greater than a threshold th, and selects sensor n (correlated sensor) combined with sensor m in the extracted correlogram. This eliminates sensors with low correlation with sensor m. The fluctuation detection unit 153 identifies the peak lag k(n, m) of the extracted correlogram.

[0097] The fluctuation detection unit 153 detects fluctuations in the time-series data of the selected correlation sensor n. The fluctuation detection unit 153 determines whether or not there is a fluctuation in the time-series data of the correlation sensor n k(n, m) periods before the peak lag. It is assumed that received power information k(n, m) periods or more ago is stored in the wireless information storage unit 120.

[0098] For example, as shown in FIG. 22 , the fluctuation detection unit 153 acquires the missing time series data k(n, m) periods before t0 using a window W of the same size as the window used to calculate the correlation coefficient of the correlogram. The fluctuation detection unit 153 detects whether or not there is a change point in the acquired time series data of window W k(n, m) periods before. If there is a change point in the data of window W at correlation sensor n, the fluctuation detection unit 153 determines that there is a fluctuation. For example, as shown in Example 1 of FIG. 22 , it may be determined that there is a fluctuation when the received power increases in the data of window W. As shown in Example 2 of FIG. 22 , it may be determined that there is a fluctuation when the received power decreases in the data of window W. For example, even if the point of fluctuation is slightly before or after window W, it may still be determined that there is a fluctuation as long as the change is a decrease in the received power.

[0099] If there is a fluctuation in the time series data of sensor n k(n,m) periods ago, the received power map correction unit 150 extrapolates the time series data of missing sensor m based on the time series data of sensor n (S206). If there is a fluctuation in the time series data of correlated sensor n k(n,m) periods ago, the extrapolation unit 155 extrapolates the time series data of missing sensor m using the value or fluctuation of the time series data of correlated sensor n k(n,m) periods ago. For example, the extrapolation unit 155 may extrapolate the time series data of missing sensor m by setting the received power a at the time of missing sensor m to equal the received power b of correlated sensor n k(n,m) periods ago. Alternatively, the extrapolation unit 155 may set the fluctuation a at the time of missing sensor m to equal the fluctuation b of correlated sensor n k(n,m) periods ago, and extrapolate the time series data of missing sensor m based on this fluctuation a. In addition, the extrapolation unit 155 may extrapolate the time series data of the missing sensor m based on the fluctuation a, where a = fluctuation b of the correlated sensor n k(n, m) periods ago / deviation b × deviation a of the missing sensor m.

[0100] Furthermore, if there is no fluctuation in the time series data of sensor n k(n,m) periods prior, the received power map correction unit 150 extrapolates the time series data of the missing sensor m using the time series data of the missing sensor m one period prior (S207). If there is no fluctuation in the time series data of the correlated sensor n k(n,m) periods prior, the extrapolation unit 155 extrapolates the time series data of the missing sensor m using the value or fluctuation of the time series data of the missing sensor m one period prior. Furthermore, the extrapolation unit 155 may use the value or fluctuation of the time series data of the sensor 300 whose received power is closest to that of the missing sensor m, the sensor 300 whose physical distance is closest to the missing sensor m, or the sensor 300 whose correlogram has the highest correlation coefficient with the missing sensor 300. Note that even if there is no fluctuation in the time series data of sensor n k(n,m) periods prior, the value or fluctuation of the time series data of sensor n k(n,m) periods prior may be used for extrapolation, as in S206.

[0101] For example, when using the value of the missing sensor m from one period before, the extrapolation unit 155 may extrapolate the time series data of the missing sensor m by setting the received power a(t) of the missing sensor m when it is missing = the received power b(t-1) of the missing sensor m one period before. Furthermore, when using the value of a sensor whose received power or distance is closest to the missing sensor m, the extrapolation unit 155 may extrapolate the time series data of the missing sensor m by setting the received power a(t) of the missing sensor m when it is missing = the received power b(t-1) of the nearest sensor one period before. Furthermore, when using the fluctuation of the sensor whose received power or distance is closest to the missing sensor m, the extrapolation unit 155 may extrapolate the time series data of the missing sensor m based on the fluctuation a. Furthermore, when using the deviation of fluctuations between the missing sensor m and a sensor having the closest received power or distance, the extrapolation unit 155 may extrapolate the time series data of the missing sensor m based on the fluctuation a, assuming that the fluctuation a of the missing sensor m = the fluctuation b of the closest sensor one period before / the deviation b × the deviation a of the sensor m. Furthermore, when using the fluctuations of the sensor with the highest correlation coefficient with the sensor m, the extrapolation unit 155 may extrapolate the time series data of the missing sensor m based on the fluctuation a, assuming that the fluctuation a of the missing sensor m = the fluctuation b of the sensor with the highest correlation coefficient one period before / the deviation b × the deviation a of the sensor m.

[0102] Next, the received power map correction unit 150 corrects the received power map stored in the received power map storage unit 140 (S208). If there are no missing data in the time-series data of any of the sensors, the correction calculation unit 157 inputs the received power acquired from the sensor 300 into the above formula (3) or formula (5) to calculate the received power at each point on the received power map. If there are missing data in the time-series data of sensor m, the correction calculation unit 157 inputs the received power acquired from the sensor 300 and the received power extrapolated in S206 or S207 into the above formula (3) or formula (5) to calculate the received power at each point on the received power map. Note that the extrapolated received power may be input as the received power at multiple points in formula (3) or formula (5).

[0103] As described above, in this embodiment, a correlogram of cross-correlation between multiple sensors is created, and the lag k(n, m) at which the correlation coefficient is highest in each correlogram is calculated. When time-series data is missing at sensor m, if there is a fluctuation in the time-series data of sensor n k(n, m) periods prior, the fluctuation is used to extrapolate the time-series data at sensor m. This makes it possible to appropriately extrapolate the missing data, thereby suppressing deterioration in the correction accuracy of the received power map.

[0104] (Embodiment 3) Next, embodiment 3 will be described. In this embodiment, in addition to embodiment 2, an example in which an obstructing object is detected and tracked by a camera will be described. Note that this embodiment can be implemented in combination with embodiment 1 or 2, and each configuration shown in embodiment 1 or 2 may be used as appropriate.

[0105] Fig. 23 shows an example configuration of an operation management system 1 according to some embodiments. The example of Fig. 23 includes a camera 600 in addition to the configuration of Fig. 9. The camera 600 captures an image of the target wireless communication environment measured by the sensor 300. The camera 600 captures an image between the sensor 300 and the base station 400, and captures, for example, an image of an obstructing object passing between the sensor 300 and the base station 400. The camera 600 transmits the captured image to the operation management server 100.

[0106] Fig. 24 shows an example of the configuration of the received power map correction unit 150 of the operation management server 100 according to some embodiments. In the example of Fig. 24, in addition to the configuration of Fig. 12, an object recognition unit 158 ​​and an object tracking unit 159 are provided.

[0107] The object recognition unit 158 ​​acquires an image captured by the camera 600 and recognizes objects around the sensor 300 from the acquired image. The object recognition unit 158 ​​recognizes objects in the image using a machine-learned object recognition model or the like, and detects obstructions between the sensor 300 and the base station 400. Obstructions are objects that may obstruct radio waves between the sensor 300 and the base station 400. For example, obstructions are moving objects such as trucks and AGVs that are large enough to affect radio waves.

[0108] The object tracking unit 159 tracks the obstructing object detected by the object recognition unit 158. The object tracking unit 159 tracks the trajectory (movement path) of the obstructing object passing between the sensor 300 and the base station 400. The object tracking unit 159 determines whether the sensor 300 is hidden behind the tracked obstructing object. The sensor 300 being hidden behind the obstructing object means that the obstructing object is located between the sensor 300 and the base station 400 and blocks radio waves from the base station 400 to the sensor 300. The object tracking unit 159 records, for each sensor 300, the time when the obstructing object is hidden behind the sensor 300, i.e., the time when the obstructing object passes in front of the sensor 300. For example, the obstructing object lag k, which corresponds to the peak lag k of the correlogram, may be calculated from the time when each sensor is hidden behind the same obstructing object. In this example, the time series data of the sensor 300 used for extrapolation may be identified based on images captured between the sensor 300 and the base station 400. The lag k may be identified from the image, and the time-series data at the identified lag k may be used for extrapolation. The lag k may be identified by the object tracking unit 159 or by another block. For example, the lag k may be identified by the fluctuation detection unit 153, the lag calculation unit 151, or the extrapolation unit 155.

[0109] Fig. 25 shows an example of the operation of the received power map correction unit 150 according to some embodiments. Fig. 25 shows an example of extrapolation using a lag obtained from a camera image without using a correlogram.

[0110] 25 , similarly to S201 in FIG. 19 , the received power map correction unit 150 acquires received power information from the multiple sensors 300 (S301). Next, the received power map correction unit 150 determines whether any of the sensors m is hidden behind an obstruction (S302). For example, based on the result of tracking the obstruction from the image of the camera 600, the object tracking unit 159 determines whether the obstruction has passed in front of any of the sensors m and whether any of the sensors m is hidden behind the obstruction.

[0111] If it is determined that sensor m is hidden by an obstruction, the received power map correction unit 150 determines whether sensor n was hidden by the same obstruction as sensor m k periods ago (S303). For example, the object tracking unit 159 determines whether the obstruction that obstructed sensor m also obstructed (passed by) another sensor n before sensor m, based on the results of tracking the obstruction from the image of the camera 600. The object tracking unit 159 determines the time when the obstruction obstructed (passed by) sensor n before sensor m as lag k.

[0112] If it is determined that the sensor n was hidden by the same obstruction as the sensor m k periods ago, the received power map correction unit 150 extrapolates the time series data of the sensor m based on the time series data of the sensor n (S304). The extrapolation unit 155 extrapolates the time series data of the sensor m using the value or fluctuation of the time series data of the sensor n k periods ago, similar to S206 in FIG. 19 .

[0113] If sensor n was not behind the same obstruction as sensor m k periods ago, i.e., if there is no sensor that was behind the same obstruction, the received power map correction unit 150 extrapolates the time series data of sensor m using the time series data of sensor m from one period ago (S305). Similar to S207 in FIG. 19 , the extrapolation unit 155 extrapolates the time series data of sensor m using the value or fluctuation of the time series data of sensor m from one period ago. Then, similar to S208 in FIG. 19 , the received power map correction unit 150 corrects the received power map stored in the received power map storage unit 140 (S306).

[0114] Figure 26 shows an example of the operation of the received power map correction unit 150 according to some embodiments. Figure 26 is an example that combines the example of using the correlogram in Figure 19 and the example of using the camera image in Figure 25.

[0115] In the example of FIG. 26, the received power map correction unit 150 calculates the correlogram and the peak lag k(n, m) that maximizes the correlation coefficient from the received powers of the multiple sensors 300, as in FIG. 19 (S201 to S203).

[0116] Thereafter, the received power map correction unit 150 determines whether or not the sensor m is hidden by an obstruction (S302), as in Fig. 25. If it is determined that the sensor m is hidden by an obstruction, the received power map correction unit 150 determines whether or not there is a gap in the time-series data of the received power information for the sensor m (S204), as in Fig. 19. That is, it determines whether or not the time-series data of the sensor m hidden by an obstruction is actually missing from the camera image.

[0117] If the received power information for sensor m is missing, the received power map correction unit 150 determines whether sensor n was hidden by the same obstruction as sensor m k periods ago, as in FIG. 25 (S303). The lag k is a lag identified from the camera image. If it is determined that sensor n was hidden by the same obstruction as sensor m k periods ago, the received power map correction unit 150 determines whether there is a fluctuation in the time-series data of sensor n k(n, m) periods ago, as in FIG. 19 (S205). The lag k(n, m) is a lag identified from the correlogram. In other words, it determines whether there is actually a fluctuation in the time-series data of sensor n that was hidden by the same obstruction as sensor m in the camera image.

[0118] 19 , if there is a fluctuation in the time series data of sensor n k(n, m) periods ago, the received power map correction unit 150 extrapolates the time series data of the missing sensor m based on the time series data of sensor n (S206), and if there is no fluctuation in the time series data of sensor n k(n, m) periods ago, the received power map correction unit 150 extrapolates the time series data of the missing sensor m using the time series data of the missing sensor m one period ago (S207). Thereafter, the received power map correction unit 150 corrects the received power map stored in the received power map storage unit 140 (S208), as in FIG. 19 .

[0119] 26 , when calculating the peak lag k(n, m) in each correlogram, the peak lag k(n, m) may be associated with the object recognition result. For example, as shown in FIG. 27 , when peak lag k1(n, m) and peak lag k2(n, m) are calculated in each correlogram, information identifying an obstruction recognized from the camera image is associated with each peak lag. For example, if peak lag k1(n, m) is the lag resulting from the influence of obstruction A, peak lag k1(n, m) is associated with obstruction A, and if peak lag k2(n, m) is the lag resulting from the influence of obstruction B, peak lag k2(n, m) is associated with obstruction B. Next, when data from sensor m is missing, peak lag k1(n, m) or peak lag k2(n, m) corresponding to the obstruction recognized based on the camera image may be selected from the correlogram, and the selected peak lag k(n, m) may be used to extrapolate the data from sensor m.

[0120] As described above, an obstruction may be tracked from a camera image, and the lag used in extrapolation may be calculated based on the tracking results. Furthermore, for a sensor that detected obstruction from a camera image, data loss or fluctuation may be checked, and extrapolation may be performed in the same manner as in embodiment 1. This allows for extrapolation using, for example, multimodal or multisensor methods, by using the detection results of the camera image, rather than just the sensor measurement values, and further improves the extrapolation accuracy.

[0121] The present disclosure is not limited to the above-described embodiment, and can be modified as appropriate within the scope of the present disclosure.

[0122] Each component in the above-described embodiments may be configured with hardware or software, or both, and may be configured with a single piece of hardware or software, or may be configured with multiple pieces of hardware or software. Each device and each function (processing) of the operation management server, etc., may be realized by a computer 30 having a processor 31 such as a CPU (Central Processing Unit) and a memory 32 serving as a storage device, as shown in FIG. 28. For example, a program for performing the method (data processing method) in the embodiment may be stored in the memory 32, and each function may be realized by the processor 31 executing the program stored in the memory 32.

[0123] These programs include instructions (or software code) that, when loaded into a computer, cause the computer to perform one or more functions described in the embodiments. The programs may be stored on a non-transitory computer-readable medium or a tangible storage medium. By way of example and not limitation, computer-readable media or tangible storage media include random-access memory (RAM), read-only memory (ROM), flash memory, solid-state drive (SSD) or other memory technology, CD-ROM, digital versatile disc (DVD), Blu-ray disc or other optical disk storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage device. The programs may also be transmitted on a transitory computer-readable medium or communication medium. By way of example and not limitation, transitory computer-readable media or communication media include electrical, optical, acoustic, or other forms of propagated signals.

[0124] Although the present disclosure has been described above with reference to the embodiments, the present disclosure is not limited to the above-described embodiments. Various modifications that can be understood by those skilled in the art can be made to the configuration and details of the present disclosure within the scope of the present disclosure. Furthermore, each embodiment can be combined with other embodiments as appropriate.

[0125] Each drawing is merely an example for describing one or more embodiments. Each drawing may not relate to only one particular embodiment, but may also relate to one or more other embodiments. As will be understood by those skilled in the art, various features or steps described with reference to any one drawing can be combined with features or steps shown in one or more other drawings to create, for example, an embodiment not explicitly shown or described. Not all features or steps shown in any one drawing are necessary to describe an exemplary embodiment, and some features or steps may be omitted. The order of steps described in any drawing may be changed as appropriate.

[0126] Some or all of the above embodiments may also be described as, but not limited to, the following supplementary notes: Some or all of the elements described in any supplementary note may be applied to various hardware, software, recording means for recording software, systems, and methods.

[0127] (Supplementary Note 1) A data processing system comprising: an acquisition means for acquiring time series data of measurement values ​​measured at a plurality of locations including a first location and a second location; a detection means for detecting a loss of the time series data acquired from the first location; and an extrapolation means for, when a loss of the time series data of the first location is detected, inserting data into the missing time series data of the first location based on characteristics between the time series data of the first location and the time-shifted time series data of the second location. (Supplementary Note 2) The data processing system according to Supplementary Note 1 further comprises a calculation means for calculating a characteristic indicating a relationship between a feature amount indicating a characteristic between the time series data of the first location and the time-shifted time series data of the second location and the amount of time shift, wherein the extrapolation means identifies data to be used for the insertion from the time series data of the second location based on the calculated characteristic. (Supplementary Note 3) The data processing system according to Supplementary Note 2, further comprising: a specifying means for specifying a first time shift amount at which the feature value becomes high in the calculated characteristic; and the extrapolation means for specifying data to be used for the insertion from the time series data of the second location according to the specified first time shift amount. (Supplementary Note 4) The data processing system according to Supplementary Note 3, wherein the extrapolation means for specifying data to be used for the insertion from the time series data of the second location when there is a fluctuation in data included in the time series data of the second location corresponding to the first time shift amount. (Supplementary Note 5) The data processing system according to Supplementary Note 3 or 4, wherein the extrapolation means for specifying data to be used for the insertion from data before the loss in the time series data of the first location when there is no fluctuation in data included in the time series data of the second location corresponding to the first time shift amount. (Supplementary Note 6) The data processing system according to any one of Supplements 2 to 5, wherein the characteristic is a correlogram showing the relationship between a lag and the cross-correlation of the time series data. (Supplementary Note 7) A data processing system according to any one of Supplementary Notes 1 to 6, further comprising a recognition unit that recognizes an object from images taken of the first location and the second location, and the identification means identifies the first time shift amount based on the object recognition results from the images of the first location and the second location.(Supplementary Note 8) The data processing system according to Supplementary Note 7, wherein the calculated characteristics associate the first time shift amount with each object, and the identification means identifies the first time shift amount corresponding to the recognized object. (Supplementary Note 9) A data processing device comprising: acquisition means for acquiring time series data of measurement values ​​measured at a plurality of points including a first point and a second point; detection means for detecting a loss of the time series data acquired from the first point; and extrapolation means for, when a loss of the time series data of the first point is detected, inserting data into the missing time series data of the first point based on a feature between the time series data of the first point and the time-shifted time series data of the second point. (Supplementary Note 10) The data processing device according to Supplementary Note 9, further comprising: calculation means for calculating a characteristic indicating a relationship between the time shift amount and a feature amount indicating a feature between the time series data of the first point and the time-shifted time series data of the second point, and the extrapolation means for identifying data to be used for the insertion from the time series data of the second point based on the calculated characteristic. (Supplementary Note 11) The data processing device according to Supplementary Note 10, further comprising: a specification means for specifying a first time shift amount at which the feature value becomes high in the calculated characteristic; and the extrapolation means specifies data to be used for the insertion from the time series data of the second location according to the specified first time shift amount. (Supplementary Note 12) The data processing device according to Supplementary Note 11, wherein the extrapolation means specifies data to be used for the insertion from the time series data of the second location when there is a fluctuation in data included in the time series data of the second location corresponding to the first time shift amount. (Supplementary Note 13) The data processing device according to Supplementary Note 11 or 12, wherein the extrapolation means specifies data to be used for the insertion from data before the loss in the time series data of the first location when there is no fluctuation in data included in the time series data of the second location corresponding to the first time shift amount. (Supplementary Note 14) The data processing device according to any one of Supplements 10 to 13, wherein the characteristic is a correlogram showing the relationship between lag and cross-correlation of the time series data.(Supplementary Note 15) A data processing method comprising: acquiring time series data of measurement values ​​measured at a plurality of locations including a first location and a second location; detecting a loss of the time series data acquired from the first location; and, if a loss of the time series data of the first location is detected, inserting data into the missing time series data of the first location based on characteristics between the time series data of the first location and the time-shifted time series data of the second location. (Supplementary Note 16) The data processing method according to Supplementary Note 15, calculating a characteristic indicating a relationship between a feature amount indicating a characteristic between the time series data of the first location and the time-shifted time series data of the second location and the time shift amount, and identifying data to be used for the insertion from the time series data of the second location based on the calculated characteristic. (Supplementary Note 17) The data processing method according to Supplementary Note 16, further comprising: identifying a first time shift amount at which the feature amount becomes high in the calculated characteristic; and identifying data to be used for the insertion from the time series data of the second location according to the identified first time shift amount. (Supplementary Note 18) The data processing method according to Supplementary Note 17, wherein, when there is a fluctuation in data included in the time series data of the second location corresponding to the first time shift amount, data to be used for the insertion is identified from the time series data of the second location. (Supplementary Note 19) The data processing method according to Supplementary Note 17 or 18, wherein, when there is no fluctuation in data included in the time series data of the second location corresponding to the first time shift amount, data to be used for the insertion is identified from data before the loss in the time series data of the first location. (Supplementary Note 20) The data processing method according to any one of Supplements 16 to 19, wherein the characteristic is a correlogram indicating the relationship between a lag and the cross-correlation of the time series data. (Supplementary Note 21) The data processing device according to any one of Supplements 9 to 14, further comprising a recognition unit that recognizes an object from images taken of the first location and the second location, and wherein the identification means identifies the first time shift amount based on a recognition result of the object from the images of the first location and the second location.(Supplementary Note 22) The data processing device according to Supplementary Note 21, wherein the calculated characteristics associate the first time shift amount with each object, and the identification means identifies the first time shift amount corresponding to the recognized object. (Supplementary Note 23) The data processing method according to any one of Supplements 15 to 20, wherein the object is recognized from images taken of the first location and the second location, and the first time shift amount is identified based on the object recognition result from the images of the first location and the second location. (Supplementary Note 24) The data processing method according to Supplementary Note 23, wherein the calculated characteristics associate the first time shift amount with each object, and the first time shift amount corresponding to the recognized object is identified.

[0128] 1 Operation management system 10 Data processing system 11 Acquisition unit 12 Detection unit 13 Extrapolation unit 20 Data processing device 30 Computer 31 Processor 32 Memory 100 Operation management server 110 Wireless information receiving unit 120 Wireless information storage unit 130 Received power map generation unit 140 Received power map storage unit 150 Received power map correction unit 151 Lag calculation unit 152 Missingness determination unit 153 Fluctuation detection unit 154 Extrapolation candidate determination unit 155 Extrapolation unit 156 Correction formula generation unit 157 Correction calculation unit 158 ​​Object recognition unit 159 Object tracking unit 160 Display unit 200 Initial measurement device 210 Wireless information transmission function 211 Received power measurement unit 212 Position detection unit 213 Wireless information transmission unit 300 Sensor 400 Base station 500 Operation management screen 510 Reception power map screen 600 Camera

Claims

1. An acquisition means for acquiring time-series data of measurement values measured at a plurality of points including a first point and a second point; a detection means for detecting a missing of the time-series data acquired from the first point; and an extrapolation means for inserting data into the missing time-series data of the first point based on a feature between the time-series data of the first point and the time-series data with the time of the second point shifted, a data processing system.

2. A calculation means for calculating a characteristic indicating a relationship between a feature amount indicating a feature between the time-series data of the first point and the time-series data with the time of the second point shifted and the amount of time shift; the extrapolation means specifying data to be used for the insertion from the time-series data of the second point based on the calculated characteristic, the data processing system according to claim 1.

3. A specifying means for specifying a first time shift amount at which the feature amount becomes high in the calculated characteristic; the extrapolation means specifying data to be used for the insertion from the time-series data of the second point according to the specified first time shift amount, the data processing system according to claim 2.

4. The extrapolation means specifying data to be used for the insertion from the time-series data of the second point when there is a variation in the data included in the time-series data of the second point corresponding to the first time shift amount, the data processing system according to claim 3.

5. The extrapolation means specifying data to be used for the insertion from the data before the missing in the time-series data of the first point when there is no variation in the data included in the time-series data of the second point corresponding to the first time shift amount, the data processing system according to claim 3 or 4.

6. The characteristic is a correlogram showing a relationship between a lag and a cross-correlation of the time-series data, the data processing system according to any one of claims 2 to 5.

7. A recognition unit for recognizing an object from images of the first point and the second point; the specifying means specifying the first time shift amount based on recognition results of the object from the images of the first point and the second point, the data processing system according to any one of claims 3 to 5.

8. In the calculated characteristics, the shift amount of the first time is associated with each object, and the specifying means specifies the shift amount of the first time corresponding to the recognized object. The data processing system according to claim 7.

9. An acquisition means for acquiring time series data of measurement values measured at a plurality of points including a first point and a second point, a detection means for detecting a loss of the time series data acquired from the first point, and when the loss of the time series data of the first point is detected, an extrapolation means for inserting data into the lost time series data of the first point based on a feature between the time series data of the first point and the time series data with the time shifted at the second point. A data processing apparatus comprising:

10. A calculating means for calculating a characteristic indicating a relationship between a feature amount indicating a feature between the time series data of the first point and the time series data with the time shifted at the second point and the shift amount of the time, and the extrapolation means specifies the data to be used for the insertion from the time series data of the second point based on the calculated characteristic. The data processing apparatus according to claim 9.

11. The calculated characteristic includes a specifying means for specifying a shift amount of a first time when the feature amount becomes high, and the extrapolation means specifies the data to be used for the insertion from the time series data of the second point according to the specified shift amount of the first time. The data processing apparatus according to claim 10.

12. When there is a variation in the data included in the time series data of the second point corresponding to the shift amount of the first time, the extrapolation means specifies the data to be used for the insertion from the time series data of the second point. The data processing apparatus according to claim 11.

13. When there is no variation in the data included in the time series data of the second point corresponding to the shift amount of the first time, the extrapolation means specifies the data to be used for the insertion from the data before the loss in the time series data of the first point. The data processing apparatus according to claim 11 or 12.

14. The characteristic is a correlogram showing a relationship between a lag and a cross-correlation of the time series data. The data processing apparatus according to any one of claims 10 to 13.

15. Obtain time-series data of measurement values measured at a plurality of points including a first point and a second point, detect a missing value in the time-series data obtained from the first point, and when a missing value in the time-series data of the first point is detected, insert data into the missing time-series data of the first point based on a feature between the time-series data of the first point and the time-series data with the time of the second point shifted. A data processing method.

16. Calculate a characteristic indicating the relationship between a feature amount indicating a feature between the time-series data of the first point and the time-series data with the time of the second point shifted and the amount of time shift, and based on the calculated characteristic, identify the data to be used for the insertion from the time-series data of the second point. The data processing method according to claim 15.

17. In the calculated characteristic, identify a first amount of time shift at which the feature amount becomes high, and based on the identified first amount of time shift, identify the data to be used for the insertion from the time-series data of the second point. The data processing method according to claim 16.

18. When there is a variation in the data included in the time-series data of the second point corresponding to the first amount of time shift, identify the data to be used for the insertion from the time-series data of the second point. The data processing method according to claim 17.

19. When there is no variation in the data included in the time-series data of the second point corresponding to the first amount of time shift, identify the data to be used for the insertion from the data before the missing value in the time-series data of the first point. The data processing method according to claim 17 or 18.

20. The characteristic is a correlogram showing the relationship between the lag and the cross-correlation of the time-series data. The data processing method according to any one of claims 16 to 19.

Images