Location estimation system and location estimation method

Abstract

To achieve high positioning accuracy while achieving aggregation in a short time.SOLUTION: A position estimation system comprises: a pre-learning point that emits a detection object parameter (PM) in a learning stage; an observation sensor that detects the PM emitted from the pre-learning point; a fusion center (FC) that aggregates information indicating the PM detected by the observation sensor in the learning stage, and records a result of aggregation in a database (DB); and a positioning object that emits the PM in an estimation stage. The observation sensor detects the PM emitted from the positioning object in the estimation stage. The FC aggregates, in the estimation stage, the information indicating the PM detected by the observation sensor, and estimates the position of the positioning object based on a result of the aggregation, the aggregation result recorded in the DB, and information on the position of the pre-learning point. A position fingerprint method is used as a position estimation method for the positioning object. The FC uses a PhyC-SN as a method for aggregating the information indicating the PM detected in the learning stage and the estimation stage.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a location estimation system and a location estimation method. [Background technology]

[0002] In recent years, the Internet of Things (IoT) has been attracting attention. Wireless sensor networks (WSNs) that can be used in small devices such as sensors have also been developed. Applications that use location information from multiple sensors are also being considered. Geolocation fingerprinting is known as a method that uses multiple sensors, and geolocation fingerprinting is described in, for example, Patent Document 1.

[0003] Known prior art location estimation techniques include those described in Non-Patent Documents 1 to 3, for example. Non-Patent Document 1 describes a mobile phone location estimation technique that uses a fingerprint method using multiple base stations. The technique described in Non-Patent Document 1 estimates the location using the received signal strength indicator (RSSI) of the terminal, etc. The technique described in Non-Patent Document 1 has the problem that it requires communication between multiple base stations and the mobile phone. Non-Patent Document 2 describes a location estimation technology based on a fingerprinting method using Bluetooth Low Energy (BLE) (registered trademark). In the technology described in Non-Patent Document 2, a BLE radio wave emitting device is installed indoors, and the target to be located measures the RSSI of the BLE, and location is determined using the ID and RSSI of each terminal as a data set. The technology described in Non-Patent Document 2 has the problem that it requires processing by the measuring terminal and an internet connection for the measuring terminal. Non-Patent Document 3 describes a location estimation method using radio wave sensors for estimating the location of a radio wave emission source. In the technology described in Non-Patent Document 3, the RSSI of radio waves from a radio wave source to be located is measured simultaneously using multiple radio wave sensors. The locations of the radio wave sensors are assumed to be known, and the location of the radio wave source to be located is estimated by a centroid addition calculation. The technology described in Non-Patent Document 3 does not require any processing for positioning, and does not require a connection to the Internet, etc., so the technology described in Non-Patent Document 3 can be used to identify illegal radio waves and estimate the location of existing cognitive radio systems. On the other hand, while collecting RSSI detected by multiple radio wave sensors is ideal, the technology described in Non-Patent Document 3 does not take this into consideration. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2020-159705 [Non-patent literature]

[0005] [Non-Patent Document 1] Kanazawa et al., Field Test Evaluation of a Location Estimation Method Using Machine Learning in LTE Networks, IEICE Technical Report SR2018-12, May 2018 [Non-patent document 2] Azuma et al., Indoor Location Estimation Using Relative Positions of Location Fingerprints Observed at Multiple Locations, IEICE Technical Report MoNaA2015-37, January 2016 [Non-patent document 3] Matsuno et al., A Study on Location Estimation of Multiple Wireless Stations for Dynamic Spectrum Sharing, IEICE General University, B-17-13, March 2020 [Non-patent document 4] T. Tsujino, T. Fujii, “High-precision location estimation using location fingerprinting method utilizing time-series received data,” IEICE Technical Report, vol. 120, no. 74, RCS2020-51, pp. 169-174, June 2020. [Non-Patent Document 5] Steven M. Kay, Fundamentals of Statistical Signal Processing, Volume 2: Detection Theory (Prentice-hall Signal Processing Series) 1998 / 1 / 27 Summary of the Invention [Problem to be solved by the invention]

[0006] In view of the above, an object of the present invention is to provide a position estimation system and a position estimation method that can achieve high positioning accuracy while achieving aggregation in a short time. [Means for solving the problem]

[0007] One aspect of the present invention is a position estimation system comprising: a pre-learning point that emits a detection target parameter, which is any one of radio waves, heat, and sound waves, in a learning phase; an observation sensor that detects the detection target parameter emitted from the pre-learning point in the learning phase; a fusion center that aggregates information indicating the detection results of the detection target parameter emitted from the pre-learning point by the observation sensor in the learning phase and records the aggregated information indicating the detection results of the detection target parameter emitted from the pre-learning point by the observation sensor in a database; and a positioning target that emits the detection target parameter in an estimation phase, wherein the observation sensor detects the detection target parameter emitted from the positioning target in the estimation phase, and the fusion center records the detection results of the detection target parameter emitted from the positioning target by the observation sensor in the estimation phase. aggregating information indicating the detection results by the observation sensor of the detection target parameters emitted from the target to be positioned, estimating the position of the target to be positioned based on the aggregated information indicating the detection results by the observation sensor of the detection target parameters emitted from the pre-learning points recorded in the database in the learning phase, and information indicating the positions of the pre-learning points; the fusion center uses a position fingerprinting method as a method for estimating the position of the target to be positioned; the fusion center uses a method for aggregating information indicating the detection results by the observation sensor of the detection target parameters emitted from the pre-learning points in the learning phase; and PhyC-SN (Physical Conversion Sensor Network) as a method for aggregating information indicating the detection results by the observation sensor of the detection target parameters emitted from the target to be positioned in the estimation phase.

[0008] In one aspect of the position estimation system of the present invention, the fusion center may estimate the position of the object to be positioned based on the similarity between the intensity distribution of the detection object parameters emitted from the pre-learning point detected by the observation sensor in the learning phase and the intensity distribution of the detection object parameters emitted from the object to be positioned detected by the observation sensor in the estimation phase.

[0009] In one embodiment of the position estimation system of the present invention, the position estimation system includes at least a first observation sensor and a second observation sensor as the observation sensors, and the fusion center may estimate the position of the object to be positioned based on the result of a vector distance calculation based on the difference between a series in which the intensities of the detection target parameters emitted from the pre-learning points detected by the first observation sensor in the learning phase and the intensities of the detection target parameters emitted from the pre-learning points detected by the second observation sensor are arranged in ascending order, and a series in which the intensities of the detection target parameters emitted from the object to be positioned detected by the first observation sensor in the estimation phase and the intensities of the detection target parameters emitted from the object to be positioned detected by the second observation sensor are arranged in ascending order.

[0010] In one embodiment of the position estimation system of the present invention, the position estimation system includes at least a first observation sensor and a second observation sensor as the observation sensors, and the fusion center may estimate the position of the object to be positioned based on a vector distance calculation result based on the difference between a series in which the strength of the detection target parameters emitted from the pre-learning point detected by the first observation sensor in the learning phase and the strength of the detection target parameters emitted from the pre-learning point detected by the second observation sensor are arranged in descending order, and a series in which the strength of the detection target parameters emitted from the object to be positioned detected by the first observation sensor in the estimation phase and the strength of the detection target parameters emitted from the object to be positioned detected by the second observation sensor are arranged in descending order.

[0011] In one embodiment of the position estimation system of the present invention, the position estimation system includes a plurality of observation sensors as the observation sensors, and the plurality of observation sensors include at least a first observation sensor arranged in a first area and a second observation sensor arranged in a second area, and the fusion center may, in the estimation stage, aggregate information indicating the detection results of the detection target parameters emitted from the positioning target by the first observation sensor into a first time zone, and aggregate information indicating the detection results of the detection target parameters emitted from the positioning target by the second observation sensor into a second time zone different from the first time zone.

[0012] In one embodiment of the position estimation system of the present invention, the fusion center may compare information indicating the detection results by the first observation sensor of the detection target parameters emitted from the positioning target, which are aggregated during the first time period of the estimation stage, with information indicating the detection results by the first observation sensor of the detection target parameters emitted from the pre-learning point, which are aggregated during the third time period of the learning stage, and may also compare information indicating the detection results by the second observation sensor of the detection target parameters emitted from the positioning target, which are aggregated during the second time period of the estimation stage, with information indicating the detection results by the second observation sensor of the detection target parameters emitted from the pre-learning point, which are aggregated during a fourth time period different from the third time period of the learning stage.

[0013] In one embodiment of the position estimation system of the present invention, the first area may include multiple observation sensors as the first observation sensors, the second area may include multiple observation sensors as the second observation sensors, the multiple observation sensors included in the first area may be arranged adjacent to each other, the first area may be strip-shaped, and the multiple observation sensors included in the second area may be arranged adjacent to each other, the second area may be strip-shaped.

[0014] In one embodiment of the position estimation system of the present invention, the first area may include multiple observation sensors as the first observation sensors, the second area may include multiple observation sensors as the second observation sensors, the multiple observation sensors included in the first area may be positioned at a distance from each other, the first area may be composed of multiple areas that are spaced apart from each other, the multiple observation sensors included in the second area may be positioned at a distance from each other, and the second area may be composed of multiple areas that are spaced apart from each other.

[0015] In one aspect of the position estimation system of the present invention, the position estimation system includes at least a first observation sensor and a second observation sensor as the observation sensors, and the fusion center generates a learning stage histogram corresponding to the spread of the detection target parameters emitted from the pre-learning points based on the intensities of the detection target parameters emitted from the pre-learning points detected by the first observation sensor in the learning stage and the intensities of the detection target parameters emitted from the pre-learning points detected by the second observation sensor, and generates a learning stage histogram corresponding to the spread of the detection target parameters emitted from the pre-learning points based on the intensities of the detection target parameters emitted from the positioning target detected by the first observation sensor in the estimation stage. a histogramming calculation unit that generates an estimated stage histogram corresponding to the spread of the detection object parameters emitted from the positioning object based on the intensities of the object parameters and the intensities of the detection object parameters emitted from the positioning object detected by the second observation sensor; and a Bhattacharyya distance calculation unit that calculates a Bhattacharyya distance that indicates a similarity between the learning stage histogram generated by the histogramming calculation unit and the estimated stage histogram, and the fusion center may estimate the position of the positioning object based on the Bhattacharyya distance calculated by the Bhattacharyya distance calculation unit and information indicating the positions of the pre-learning points.

[0016] One aspect of the present invention is a method for estimating the position of a target to be positioned in a position estimation system including a pre-learning point, an observation sensor, a fusion center, and a target to be positioned, the method comprising: a first step in which the pre-learning point emits a detection target parameter, which is one of radio waves, heat, and sound waves, during a learning phase; a second step in which the observation sensor detects the detection target parameter emitted from the pre-learning point during the first step during the learning phase; a third step in which the fusion center collects information indicating the detection results of the detection target parameter emitted from the pre-learning point detected by the observation sensor during the second step during the learning phase; a fourth step in which the fusion center records the collected information indicating the detection results of the detection target parameter emitted from the pre-learning point collected during the third step in a database during the learning phase; a fifth step in which the target to be positioned emits the detection target parameter during an estimation phase; and a sixth step in which the observation sensor detects the detection target parameter emitted from the target to be positioned during the fifth step during the estimation phase. The location estimation method includes, in the estimation stage, a seventh step in which the fusion center aggregates information indicating the detection results of the detection target parameters emitted from the location target detected by the observation sensor in the sixth step, and an eighth step in which the fusion center estimates the location of the location target based on the aggregated information indicating the detection results of the detection target parameters emitted from the location target aggregated in the seventh step, the aggregated information indicating the detection results of the detection target parameters emitted from the pre-learning points aggregated in the third step, and information indicating the location of the pre-learning points, wherein the location fingerprinting method is used as the method by which the fusion center aggregates information indicating the detection results of the detection target parameters emitted from the pre-learning points in the third step, and PhyC-SN is used as the method by which the fusion center aggregates information indicating the detection results of the detection target parameters emitted from the location target in the seventh step. [Effects of the Invention]

[0017] According to the present invention, it is possible to provide a position estimation system and a position estimation method that can achieve high positioning accuracy while achieving aggregation in a short time. [Brief explanation of the drawings]

[0018] [Figure 1] FIG. 1 is a diagram illustrating a first example of a position estimation system 1 according to a first embodiment. [Figure 2] 10 is a diagram for explaining an example of the spread of radio waves emitted from a pre-learning point 1A1 in the learning stage of the position estimation system 1. FIG. [Figure 3] 10 is a diagram for explaining an example of the spread of radio waves emitted from a pre-learning point 1A2 in the learning stage of the position estimation system 1. FIG. [Figure 4] 10 is a diagram for explaining an example of the spread of radio waves emitted by a positioning object 1D in the estimation stage of the position estimation system 1. FIG. [Figure 5] 10 is a diagram for explaining a method for estimating the position of a positioning target 1D by a fusion center 1C. FIG. [Figure 6] 4 is a flowchart illustrating an example of processing executed in the position estimation system 1 of the first embodiment. [Figure 7] 10 is a diagram for explaining a method for estimating the position of a positioning target 1D by a fusion center 1C in a second example of the position estimation system 1 of the first embodiment. FIG. [Figure 8] This figure explains a specific example in which, in a second example of the position estimation system 1 of the first embodiment, the fusion center 1C compares the series related to the pre-learning point 1A1 shown in Figure 7 with the series related to the positioning target 1D, and compares the series related to the pre-learning point 1A2 with the series related to the positioning target 1D. [Figure 9] 10 is a diagram for explaining a method for estimating the position of a positioning target 1D by a fusion center 1C in a third example of the position estimation system 1 of the first embodiment. FIG. [Figure 10]This figure explains a specific example in which, in a third example of the position estimation system 1 of the first embodiment, the fusion center 1C compares the series related to the pre-learning point 1A1 shown in Figure 9 with the series related to the positioning target 1D, and compares the series related to the pre-learning point 1A2 with the series related to the positioning target 1D. [Figure 11] FIG. 10 is a diagram for explaining the aggregation by time division for each area of ​​information indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from a positioning target 1D in a fourth example of the position estimation system 1 of the first embodiment. [Figure 12] This figure is for explaining the aggregation by time division for each area of ​​information showing the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from pre-learning point 1A1 in the fourth example of the position estimation system 1 of the first embodiment, and the aggregation by time division for each area of ​​information showing the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from pre-learning point 1A2. [Figure 13] FIG. 10 is a diagram illustrating a fifth example of the position estimation system 1 according to the first embodiment. [Figure 14] FIG. 10 is a diagram for explaining area division (time division) and the like in a fifth example of the position estimation system 1 of the first embodiment. [Figure 15] FIG. 10 is a diagram illustrating a sixth example of the position estimation system 1 according to the first embodiment. [Figure 16] 10A and 10B are diagrams showing examples of a learning stage histogram, an estimation stage histogram, etc. generated by a histogram calculation unit 1C2. [Figure 17] FIG. 10 is a diagram showing simulation parameters. [Figure 18] FIG. 10 is a diagram showing simulation parameters. [Figure 19] FIG. 19 is a diagram showing the simulation results. [Figure 20] FIG. 17 is a diagram showing a simulation result when the histogram calculation shown in FIG. 16 is performed. [Figure 21] FIG. 1 is a diagram for explaining the concept of a real-machine test. [Figure 22] FIG. 1 is a diagram showing equipment used in a real-machine test. [Figure 23] FIG. 1 is a diagram showing the arrangement of sensors (observation sensors 1B1, 1B2, . . . ) in a real-machine test. [Figure 24] FIG. 10 is a diagram showing the arrangement of learning data measurement points (pre-learning points 1A1 to 1AN) and observation data measurement points (target 1D to be positioned) in a real-machine test. [Figure 25] FIG. 10 is a diagram showing a comparison of the results of an actual machine test. [Figure 26] FIG. 10 is a diagram showing other simulation results. [Figure 27] FIG. 1 is a diagram for explaining a positional fingerprinting method. [Figure 28] FIG. 10 is a diagram for explaining PhyC-SN. DETAILED DESCRIPTION OF THE INVENTION

[0019] Before describing the embodiments of the position estimation system and the position estimation method of the present invention, the technology on which the present invention is based will be described.

[0020] FIG. 27 is a diagram for explaining the position fingerprinting method. The system shown in FIG. 27 includes pre-learning points a to d, sensors SR1 to SR3, a fusion center FC, and an observation point (target of positioning) x. 27, wireless parameters (e.g., received signal strength indicator (RSSI) of radio waves) emitted from each of pre-learning points a to d are measured in advance at multiple points (sensors SR1 to SR3) within the area. The wireless parameters emitted from each of pre-learning points a to d measured in advance by sensors SR1 to SR3 are collected in a fusion center FC together with the location information of each of pre-learning points a to d and stored in a database. Furthermore, in the system shown in Figure 27, the position of observation point (target to be positioned) x is identified by pattern matching the radio parameters emitted from each of pre-learning points a to d and measured in advance by each of sensors SR1 to SR3 with the radio parameters emitted from observation point (target to be positioned) x and measured by each of sensors SR1 to SR3 (i.e., comparing the respective RSSI values ​​and selecting the closest point (performing vector distance calculation)). The technique shown in FIG. 27 is described in Non-Patent Document 4, for example.

[0021] The system shown in FIG. 27 has the advantage that the configuration of the sensors SR1 to SR3 can be simplified because the position information of the sensors SR1 to SR3 is not required to identify (estimate) the position of the observation point (target of positioning) x. Another advantage of the system shown in Figure 27 is that in order to identify (estimate) the position of observation point (target to be positioned) x, observation point (target to be positioned) x only needs to emit wireless parameters, and no processing is required for observation point (target to be positioned) x. On the other hand, a disadvantage of the type of system shown in Fig. 27 is that if the number of sensors SR1, SR2, SR3, ... is greater than the example shown in Fig. 4, there is a risk of information loss due to packet collisions caused by simultaneous access. Also, if a time-division access protocol is applied, the time required for aggregation will be longer. In other words, it can be said that an ingenious data collection method is necessary.

[0022] Therefore, as will be described later, in an embodiment of the location estimation system and location estimation method of the present invention, a collective aggregation method using wireless physical quantity conversion (PhyC-SN: Physical Conversion Sensor Network) is applied to the location fingerprint method shown in Figure 27. FIG. 28 is a diagram illustrating PhyC-SN. The system shown in FIG. 28 includes sensors #1 to #3 and a fusion center FC. In the system shown in Fig. 28, sensor #1 converts a sensing result (e.g., temperature 15°C) directly into a physical quantity of a carrier wave (carrier frequency f0 + 15 Hz) and transmits the corresponding subcarrier to the fusion center FC. Sensor #2 converts a sensing result (e.g., temperature -50°C) directly into a physical quantity of a carrier wave (carrier frequency f0 - 50 Hz) and transmits the corresponding subcarrier to the fusion center FC. Sensor #3 converts a sensing result (e.g., temperature 70°C) directly into a physical quantity of a carrier wave (carrier frequency f0 + 70 Hz) and transmits the corresponding subcarrier to the fusion center FC. 28 has the advantage that the fusion center FC can simultaneously collect (receive) subcarriers from sensors #1, #2, and #3, thereby reducing the time required for data collection. If the same physical quantity is selected in PhyC-SN, the handling will be as follows: When multiple sensors transmit the same information, they select the same physical quantity and receive multiple carriers simultaneously. In this case, the number of subcarriers can be identified by applying the multi-level detection described in Non-Patent Document 5, taking advantage of the characteristic that the amount of energy increases according to the number of subcarriers. As an alternative method, single-level detection with a single decision threshold, as described in Non-Patent Document 5, can be used to detect the presence of a notification even if the number of subcarriers is unknown. For example, if three sensors transmit the same information, single-level detection will determine whether the information is present or absent, and will recognize that only one sensor sent the information, resulting in the loss of two pieces of sensor information. In this way, single-level detection can be used by allowing for the loss of sensor information, thereby simplifying the detection method. Such loss of sensor information will result in a deterioration in the position estimation accuracy of the proposed method (embodiment of the position estimation system and position estimation method of the present invention), but if the deterioration is allowed, simple single-level detection can be used.

[0023] Hereinafter, embodiments of the position estimation system and the position estimation method of the present invention will be described.

[0024] First Embodiment FIG. 1 is a diagram showing a first example of a position estimation system 1 according to the first embodiment. 1, a position estimation system 1 is used to estimate the position of a positioning target (observation point) 1D. The position estimation system 1 includes pre-learning points 1A1, 1A2, ..., 1AN, observation sensors 1B1 to 1B9, a fusion center (FC) 1C, and the positioning target (observation point) 1D. In the example shown in Figure 1, the location estimation system 1 has N (N is any number greater than 2) pre-learning points 1A1, 1A2, ..., 1AN, but in other examples, the location estimation system 1 may have any number of pre-learning points other than N. In the example shown in FIG. 1, the position estimation system 1 includes nine observation sensors 1B1 to 1B9, but in other examples, the position estimation system 1 may include any number of observation sensors other than nine.

[0025] In the example shown in FIG. 1, pre-learning point 1A1 emits radio waves as a detection target parameter during the learning stage of position estimation system 1 (e.g., at a first timing during the learning stage of position estimation system 1). Pre-learning point 1A2 emits radio waves during the learning stage of position estimation system 1 (e.g., at a second timing different from the first timing during the learning stage of position estimation system 1). Pre-learning point 1AN emits radio waves during the learning stage of position estimation system 1 (e.g., at an Nth timing different from the first timing, second timing, etc. during the learning stage of position estimation system 1). The observation sensors 1B1 to 1B9 detect radio waves emitted from the respective pre-learning points 1A1, 1A2, . . . , 1AN during the learning stage of the position estimation system 1. The fusion center (FC) 1C includes a database 1C1. The fusion center 1C collects information (in the examples shown in FIGS. 2 and 3, "RSSI data") indicating the detection results by the observation sensors 1B1 to 1B9 of radio waves emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN during the learning stage of the position estimation system 1. In detail, the fusion center 1C uses PhyC-SN shown in FIG. 28 as a method for collecting information indicating the detection results by the observation sensors 1B1 to 1B9 of radio waves emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN during the learning stage of the position estimation system 1. In addition, the fusion center 1C records in the database 1C1 the aggregated information (in the examples shown in Figures 2 and 3, "preliminary information") indicating the detection results by the observation sensors 1B1 to 1B9 of the radio waves emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN during the learning phase of the position estimation system 1. Furthermore, the fusion center 1C records information indicating the positions of the pre-learning points 1A1, 1A2, . . . , 1AN (the position coordinates of the pre-learning points 1A1, 1A2, . . . , 1AN) in the database 1C1.

[0026] FIG. 2 is a diagram for explaining an example of the spread of radio waves emitted from the pre-learning point 1A1 during the learning stage of the position estimation system 1. In FIG. In the example shown in Figure 2, the RSSI value (approximately -40 [dBm]) of the radio waves emitted by pre-learning point 1A1 detected by observation sensor 1B4 is the strongest value. The RSSI value (approximately -50 [dBm]) of the radio waves emitted by pre-learning point 1A1 detected by observation sensors 1B1 and 1B5 is the second strongest value. The RSSI value (approximately -60 [dBm]) of the radio waves emitted by pre-learning point 1A1 detected by observation sensors 1B2 and 1B7 is the third strongest value. The RSSI value (approximately -70 [dBm]) of the radio waves emitted by pre-learning point 1A1 detected by observation sensor 1B6 is the fourth strongest value. The RSSI value (approximately -80 [dBm]) of the radio waves emitted by pre-learning point 1A1 detected by observation sensors 1B3 and 1B8 is the fifth strongest value. The RSSI value (approximately -90 [dBm]) of the radio wave emitted from the pre-learning point 1A1 detected by the observation sensor 1B9 is the weakest value.

[0027] In the example shown in Figure 2, the fusion center 1C uses PhyC-SN to aggregate information (RSSI data) indicating the RSSI value (approximately -40 [dBm]) of the radio waves detected by observation sensor 1B4, the RSSI value (approximately -50 [dBm]) of the radio waves detected by observation sensors 1B1 and 1B5, the RSSI value (approximately -60 [dBm]) of the radio waves detected by observation sensors 1B2 and 1B7, the RSSI value (approximately -70 [dBm]) of the radio waves detected by observation sensors 1B6, the RSSI value (approximately -80 [dBm]) of the radio waves detected by observation sensors 1B3 and 1B8, and the RSSI value (approximately -90 [dBm]) of the radio waves detected by observation sensor 1B9, which are radio waves emitted by pre-learning point 1A1 during the learning phase of the position estimation system 1. Therefore, the fusion center 1C can simultaneously receive information (RSSI data) indicating the detection results of the radio waves emitted from the pre-learning point 1A1 by the observation sensors 1B1 to 1B9 during the learning stage of the position estimation system 1 from each of the observation sensors 1B1 to 1B9.

[0028] In addition, in the example shown in Figure 2, the fusion center 1C records in the database 1C1 ``preliminary information of pre-learning point 1A1'' indicating that, as a result of aggregating information (RSSI data) indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from pre-learning point 1A1 during the learning stage of the position estimation system 1, the number of observation sensors whose radio wave RSSI value was approximately -40 [dBm] was ``1,'' the number of observation sensors whose radio wave RSSI value was approximately -50 [dBm] was ``2,'' the number of observation sensors whose radio wave RSSI value was approximately -60 [dBm] was ``2,'' the number of observation sensors whose radio wave RSSI value was approximately -70 [dBm] was ``1,'' the number of observation sensors whose radio wave RSSI value was approximately -80 [dBm] was ``2,'' and the number of observation sensors whose radio wave RSSI value was approximately -90 [dBm] was ``1.'' Furthermore, the fusion center 1C records information indicating the position of the pre-learning point 1A1 (position coordinates of the pre-learning point 1A1) in the database 1C1.

[0029] FIG. 3 is a diagram for explaining an example of the spread of radio waves emitted from the pre-learning point 1A2 during the learning stage of the position estimation system 1. In FIG. In the example shown in Figure 3, the RSSI value (approximately -40 [dBm]) of the radio waves emitted by pre-learning point 1A2 detected by observation sensor 1B7 is the strongest value. The RSSI value (approximately -50 [dBm]) of the radio waves emitted by pre-learning point 1A2 detected by observation sensor 1B5 is the second strongest value. The RSSI value (approximately -60 [dBm]) of the radio waves emitted by pre-learning point 1A2 detected by observation sensors 1B4 and 1B8 is the third strongest value. The RSSI value (approximately -70 [dBm]) of the radio waves emitted by pre-learning point 1A2 detected by observation sensors 1B2 and 1B6 is the fourth strongest value. The RSSI value (approximately -80 [dBm]) of the radio waves emitted by pre-learning point 1A2 detected by observation sensors 1B1 and 1B3 is the fifth strongest value. The RSSI value (approximately −90 [dBm]) of the radio wave emitted from the pre-learning point 1A2 detected by the observation sensor 1B9 is the weakest value.

[0030] In the example shown in Figure 3, the fusion center 1C uses PhyC-SN to aggregate information (RSSI data) indicating the RSSI value (approximately -40 [dBm]) of the radio waves detected by observation sensor 1B7, the RSSI value (approximately -50 [dBm]) of the radio waves detected by observation sensor 1B5, the RSSI value (approximately -60 [dBm]) of the radio waves detected by observation sensors 1B4 and 1B8, the RSSI value (approximately -70 [dBm]) of the radio waves detected by observation sensors 1B2 and 1B6, the RSSI value (approximately -80 [dBm]) of the radio waves detected by observation sensors 1B1 and 1B3, and the RSSI value (approximately -90 [dBm]) of the radio waves detected by observation sensor 1B9, which are radio waves emitted by pre-learning point 1A2 during the learning phase of the position estimation system 1. Therefore, the fusion center 1C can simultaneously receive information (RSSI data) indicating the detection results of the observation sensors 1B1 to 1B9 of the radio waves emitted from the pre-learning point 1A2 during the learning stage of the position estimation system 1 from each of the observation sensors 1B1 to 1B9.

[0031] In addition, in the example shown in Figure 3, the fusion center 1C records in the database 1C1 "preliminary information of pre-learning point 1A2" indicating that, as a result of aggregating information (RSSI data) indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from pre-learning point 1A2 during the learning stage of the position estimation system 1, the number of observation sensors whose radio wave RSSI value was approximately -40 [dBm] was "1", the number of observation sensors whose radio wave RSSI value was approximately -50 [dBm] was "1", the number of observation sensors whose radio wave RSSI value was approximately -60 [dBm] was "2", the number of observation sensors whose radio wave RSSI value was approximately -70 [dBm] was "2", the number of observation sensors whose radio wave RSSI value was approximately -80 [dBm] was "2", and the number of observation sensors whose radio wave RSSI value was approximately -90 [dBm] was "1". Furthermore, the fusion center 1C records information indicating the position of the pre-learning point 1A2 (position coordinates of the pre-learning point 1A2) in the database 1C1.

[0032] 1, the object 1D to be positioned emits radio waves as detection object parameters in the estimation stage (stage of estimating the position of the object 1D to be positioned) of the position estimation system 1. Observation sensors 1B1 to 1B9 detect the radio waves emitted from the object 1D to be positioned in the estimation stage of the position estimation system 1. The fusion center 1C aggregates information indicating the detection results (in the example shown in FIG. 4, "RSSI data") of the radio waves emitted from the object 1D to be positioned by the observation sensors 1B1 to 1B9 in the estimation stage of the position estimation system 1. In detail, the fusion center 1C uses PhyC-SN shown in FIG. 28 as a method for aggregating information indicating the detection results of the radio waves emitted from the object 1D to be positioned by the observation sensors 1B1 to 1B9 in the estimation stage of the position estimation system 1. In addition, the fusion center 1C estimates the position of the positioning target 1D based on the aggregated results of information indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from the positioning target 1D during the estimation phase of the position estimation system 1 (in the example shown in Figure 4, "observation information of the positioning target 1D"), the aggregated results of information (RSII data) indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN recorded in database 1C1 during the learning phase of the position estimation system 1 ("pre-information of pre-learning point 1A1", "pre-information of pre-learning point 1A2", ..., "pre-information of pre-learning point 1AN"), and information indicating the positions of each of the pre-learning points 1A1, 1A2, ..., 1AN recorded in database 1C1 during the learning phase of the position estimation system 1 ("position coordinates of pre-learning point 1A1", "position coordinates of pre-learning point 1A2", ..., "position coordinates of pre-learning point 1AN"). More specifically, the fusion center 1C uses the position fingerprinting method shown in FIG. 27 as a method for estimating the position of the positioning target 1D.

[0033] FIG. 4 is a diagram for explaining an example of the spread of radio waves emitted by the positioning object 1D in the estimation stage of the position estimation system 1. In FIG. In the example shown in FIG. 4, the RSSI value (approximately -40 [dBm]) of the radio waves emitted by the positioning target 1D detected by observation sensors 1B5 and 1B6 is the strongest value. The RSSI value (approximately -50 [dBm]) of the radio waves emitted by the positioning target 1D detected by observation sensors 1B9 is the second strongest value. The RSSI value (approximately -60 [dBm]) of the radio waves emitted by the positioning target 1D detected by observation sensors 1B2 and 1B4 is the third strongest value. The RSSI value (approximately -70 [dBm]) of the radio waves emitted by the positioning target 1D detected by observation sensors 1B3 and 1B8 is the fourth strongest value. The RSSI value (approximately -80 [dBm]) of the radio waves emitted by the positioning target 1D detected by observation sensor 1B7 is the fifth strongest value. The RSSI value (approximately -90 [dBm]) of the radio wave emitted by the positioning target 1D detected by the observation sensor 1B1 is the weakest value.

[0034] In the example shown in Figure 4, the fusion center 1C uses PhyC-SN to aggregate information (RSSI data) indicating the RSSI value (approximately -40 [dBm]) of the radio waves detected by observation sensors 1B5 and 1B6, the RSSI value (approximately -50 [dBm]) of the radio waves detected by observation sensor 1B9, the RSSI value (approximately -60 [dBm]) of the radio waves detected by observation sensors 1B2 and 1B4, the RSSI value (approximately -70 [dBm]) of the radio waves detected by observation sensor 1B3 and 1B8, the RSSI value (approximately -80 [dBm]) of the radio waves detected by observation sensor 1B7, and the RSSI value (approximately -90 [dBm]) of the radio waves detected by observation sensor 1B1, which are radio waves emitted by the positioning target 1D during the estimation stage of the position estimation system 1. Therefore, the fusion center 1C can simultaneously receive information (RSSI data) indicating the detection results of the observation sensors 1B1 to 1B9 of the radio waves emitted from the positioning target 1D in the estimation stage of the position estimation system 1 from each of the observation sensors 1B1 to 1B9.

[0035] In addition, in the example shown in Figure 4, the aggregated result of information (RSSI data) indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from the positioning target 1D during the estimation stage of the position estimation system 1 is "observation information of the positioning target 1D" indicating that the number of observation sensors whose radio wave RSSI value was approximately -40 [dBm] was "2," the number of observation sensors whose radio wave RSSI value was approximately -50 [dBm] was "1," the number of observation sensors whose radio wave RSSI value was approximately -60 [dBm] was "2," the number of observation sensors whose radio wave RSSI value was approximately -70 [dBm] was "2," the number of observation sensors whose radio wave RSSI value was approximately -80 [dBm] was "1," and the number of observation sensors whose radio wave RSSI value was approximately -90 [dBm] was "1."

[0036] FIG. 5 is a diagram for explaining a method for estimating the position of the object 1D to be positioned by the fusion center 1C. In the example shown in Figure 5, in order to estimate the position of the positioning target 1D, the fusion center 1C uses the ``observation information of the positioning target 1D'' obtained by the processing shown in Figure 4, the ``preliminary information of pre-learning point 1A1,'' ``preliminary information of pre-learning point 1A2,'' ..., ``preliminary information of pre-learning point 1AN'' obtained by the processing shown in Figures 2 and 3, and information indicating the respective positions of the pre-learning points 1A1, 1A2, ..., 1AN recorded in database 1C1 (``position coordinates of pre-learning point 1A1,'' ``position coordinates of pre-learning point 1A2,'' ..., ``position coordinates of pre-learning point 1AN'').

[0037] Fusion center 1C aggregates the results of information indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from target 1D to be positioned during the estimation stage of location estimation system 1 (i.e., "observation information of target 1D to be positioned") and the results of information (RSII data) indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from each of pre-learning points 1A1, 1A2, ..., 1AN recorded in database 1C1 during the learning stage of location estimation system 1 (i.e., "pre-learning data"). The position of the positioning target 1D is estimated by using the position fingerprint method shown in Figure 27 based on the information indicating the positions of the pre-learning points 1A1, 1A2, ..., 1AN recorded in the database 1C1 during the learning stage of the position estimation system 1 (i.e., "position coordinates of pre-learning point 1A1," "position coordinates of pre-learning point 1A2," ..., "position coordinates of pre-learning point 1AN"), and the information indicating the positions of the pre-learning points 1A1, 1A2, ..., 1AN recorded in the database 1C1 during the learning stage of the position estimation system 1 (i.e., "position coordinates of pre-learning point 1A1," "position coordinates of pre-learning point 1A2," ..., "position coordinates of pre-learning point 1AN"). For example, the fusion center 1C performs a comparative calculation between the "observation information of the positioning target 1D" and the "preliminary information of preliminary learning point 1A1," "preliminary information of preliminary learning point 1A2," ..., and "preliminary information of preliminary learning point 1AN," selects the "preliminary information of preliminary learning point" that is closest to the "observation information of the positioning target 1D," and estimates the position coordinates of the preliminary learning point of the selected "preliminary information of preliminary learning point" as the position coordinates of the positioning target 1D.

[0038] For example, the fusion center 1C estimates the position of the positioning target 1D based on the similarity between the intensity distribution of radio waves emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN detected by the observation sensors 1B1 to 1B9 during the learning phase of the position estimation system 1 (such as the intensity distribution of the graph of "Preliminary information of pre-learning point 1A1" shown in Figure 5 and the intensity distribution of the graph of "Preliminary information of pre-learning point 1A2" shown in Figure 5) and the intensity distribution of radio waves emitted from the positioning target 1D detected by the observation sensors 1B1 to 1B9 during the estimation phase of the position estimation system 1 (such as the intensity distribution of the graph of "Observation information of positioning target 1D" shown in Figure 5).

[0039] As described above, in the first example of the location estimation system 1 of the first embodiment, by applying the PhyC-SN shown in Figure 28 to the location fingerprint method shown in Figure 27, it is possible to achieve high positioning accuracy while achieving aggregation in a short period of time.

[0040] FIG. 6 is a flowchart illustrating an example of processing executed in the position estimation system 1 of the first embodiment. In the example shown in FIG. 6, in step S11A of the learning stage of the position estimation system 1, a pre-learning point 1A1 emits a detection target parameter (specifically, a radio wave). Next, in step S11B of the learning stage of position estimation system 1, observation sensors 1B1 to 1B9 detect the detection target parameters (radio waves) emitted from pre-learning point 1A1 in step S11A. Next, in step S11C of the learning stage of the position estimation system 1, the fusion center 1C aggregates information (specifically, RSSI data) indicating the detection results of the detection target parameters (radio waves) emitted from the pre-learning point 1A1 detected by the observation sensors 1B1 to 1B9 in step S11B. Next, in step S11D of the learning stage of the position estimation system 1, the fusion center 1C records the aggregated results of information (RSSI data) indicating the detection results of the detection target parameters (radio waves) emitted from the pre-learning point 1A1 aggregated in step S11C (specifically, the pre-information of the pre-learning point 1A1) in the database 1C1.

[0041] Furthermore, in step S12A of the learning stage of the position estimation system 1, the pre-learning point 1A2 emits a detection target parameter (radio wave). Next, in step S12B of the learning stage of position estimation system 1, observation sensors 1B1 to 1B9 detect the detection target parameters (radio waves) emitted from pre-learning point 1A2 in step S12A. Next, in step S12C of the learning stage of the position estimation system 1, the fusion center 1C collects information (RSSI data) indicating the detection results of the detection target parameters (radio waves) emitted from the pre-learning points 1A2 detected by the observation sensors 1B1 to 1B9 in step S12B. Next, in step S12D of the learning stage of the position estimation system 1, the fusion center 1C records in the database 1C1 the aggregated results (preliminary information of the preliminary learning point 1A2) of information (RSSI data) indicating the detection results of the detection target parameters (radio waves) emitted from the preliminary learning point 1A2 aggregated in step S12C.

[0042] In step S1NA of the learning stage of the position estimation system 1, the pre-learning point 1AN emits detection target parameters (radio waves). Next, in step S1NB of the learning stage of the position estimation system 1, the observation sensors 1B1 to 1B9 detect the detection target parameters (radio waves) emitted from the pre-learning point 1AN in step S1NA. Next, in step S1NC of the learning stage of the position estimation system 1, the fusion center 1C collects information (RSSI data) indicating the detection results of the detection target parameters (radio waves) emitted from the pre-learning points 1AN detected by the observation sensors 1B1 to 1B9 in step S1NB. Next, in step S1ND of the learning stage of the position estimation system 1, the fusion center 1C records the aggregated results (preliminary information of the preliminary learning point 1AN) of information (RSSI data) indicating the detection results of the detection target parameters (radio waves) emitted from the preliminary learning point 1AN aggregated in step S1NC in the database 1C1.

[0043] Next, in step S21 of the estimation stage of the position estimation system 1, the object 1D to be positioned emits detection object parameters (radio waves). Next, in step S22 of the estimation stage of the position estimation system 1, the observation sensors 1B1 to 1B9 detect the detection object parameters (radio waves) emitted from the positioning object 1D in step S21. Next, in step S23 of the estimation stage of the position estimation system 1, the fusion center 1C collects information (RSSI data) indicating the detection results of the detection object parameters (radio waves) emitted from the positioning object 1D detected by the observation sensors 1B1 to 1B9 in step S22. Next, in step S24 of the estimation stage of the position estimation system 1, the fusion center 1C estimates the position of the positioning target 1D based on the aggregated results of information (RSSI data) indicating the detection results of the detection target parameters (radio waves) emitted from the positioning target 1D aggregated in step S23 (specifically, observation information of the positioning target 1D), the aggregated results of information (RSII data) indicating the detection results of the detection target parameters (radio waves) emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN aggregated in steps S11C, S12C, and S1NC (specifically, prior information of pre-learning point 1A1, prior information of pre-learning point 1A2, ..., prior information of pre-learning point 1AN), and information indicating the positions of each of the pre-learning points 1A1, 1A2, ..., 1AN (specifically, the position coordinates of pre-learning point 1A1, the position coordinates of pre-learning point 1A2, ..., the position coordinates of pre-learning point 1AN).

[0044] 6, the fusion center 1C uses the position fingerprinting method as a method for estimating the position of the positioning target 1D in step S24. Also, the fusion center 1C uses PhyC-SN as a method for aggregating information (RSII data) indicating the detection results of the detection target parameters (radio waves) emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN in steps S11C, S12C, and S1NC, and as a method for aggregating information (RSII data) indicating the detection results of the detection target parameters (radio waves) emitted from the positioning target 1D in step S23.

[0045] In the second example of the position estimation system 1 of the first embodiment, the position estimation system 1 is configured in a manner generally similar to the position estimation system 1 shown in FIG. As described above, in the first example of the position estimation system 1 of the first embodiment shown in Figure 1, the position estimation system 1 has nine observation sensors 1B1 to 1B9, while in the second example of the position estimation system 1 of the first embodiment, the position estimation system 1 has five observation sensors 1B1 to 1B5.

[0046] In the second example of the position estimation system 1 of the first embodiment, the observation sensors 1B1 to 1B5 detect radio waves emitted from the respective pre-learning points 1A1, 1A2, . . . , 1AN during the learning stage of the position estimation system 1. The fusion center (FC) 1C aggregates information ("RSSI data" in the example shown in Figure 7) indicating the detection results by the observation sensors 1B1 to 1B5 of the radio waves emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN during the learning stage of the position estimation system 1. In addition, the fusion center 1C records in the database 1C1 the aggregated information indicating the detection results by the observation sensors 1B1 to 1B5 of the radio waves emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN during the learning phase of the position estimation system 1 (in the example shown in Figure 7, "pre-information for pre-learning point 1A1," "pre-information for pre-learning point 1A2," ...).

[0047] FIG. 7 is a diagram for explaining a method for estimating the position of the object 1D to be positioned by the fusion center 1C in the second example of the position estimation system 1 of the first embodiment. In the example shown in FIG. 7, during the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted by the preliminary learning point 1A1 among the observation sensors 1B1 to 1B5 with the strongest RSSI value is approximately −45 [dBm]. During the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted by the preliminary learning point 1A1 among the observation sensors 1B1 to 1B5 with the second strongest RSSI value is approximately −50 [dBm]. During the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted by the preliminary learning point 1A1 among the observation sensors 1B1 to 1B5 with the third strongest RSSI value is approximately −64 [dBm]. During the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted by the preliminary learning point 1A1 among the observation sensors 1B1 to 1B5 with the fourth strongest RSSI value is approximately −78 [dBm]. During the learning stage of position estimation system 1, the RSSI value of the observation sensor that detected the radio waves emitted by pre-learning point 1A1 with the weakest RSSI value among observation sensors 1B1 to 1B5 is approximately −80 [dBm].

[0048] In the example shown in FIG. 7, during the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted from the preliminary learning point 1A2 with the strongest RSSI value among the observation sensors 1B1 to 1B5 is approximately −44 [dBm]. During the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted from the preliminary learning point 1A2 with the second strongest RSSI value among the observation sensors 1B1 to 1B5 is approximately −55 [dBm]. During the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted from the preliminary learning point 1A2 with the third strongest RSSI value among the observation sensors 1B1 to 1B5 is approximately −60 [dBm]. During the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted from the preliminary learning point 1A2 with the fourth strongest RSSI value among the observation sensors 1B1 to 1B5 is approximately −83 [dBm]. During the learning stage of position estimation system 1, the RSSI value of the observation sensor that detected the radio waves emitted by pre-learning point 1A2 with the weakest RSSI value among observation sensors 1B1 to 1B5 is approximately −84 [dBm].

[0049] 7, during the estimation stage of the position estimation system 1, the RSSI value of the observation sensor among the observation sensors 1B1 to 1B5 that detects the radio waves emitted by the positioning target (observation point) 1D with the strongest RSSI value is approximately −44 [dBm]. During the estimation stage of the position estimation system 1, the RSSI value of the observation sensor among the observation sensors 1B1 to 1B5 that detects the radio waves emitted by the positioning target 1D with the second strongest RSSI value is approximately −55 [dBm]. During the estimation stage of the position estimation system 1, the RSSI value of the observation sensor among the observation sensors 1B1 to 1B5 that detects the radio waves emitted by the positioning target 1D with the third strongest RSSI value is approximately −60 [dBm]. During the estimation stage of the position estimation system 1, the RSSI value of the observation sensor among the observation sensors 1B1 to 1B5 that detects the radio waves emitted by the positioning target 1D with the fourth strongest RSSI value is approximately −83 [dBm]. During the estimation stage of position estimation system 1, the RSSI value of the observation sensor that detected the radio waves emitted by positioning target 1D with the weakest RSSI value among observation sensors 1B1 to 1B5 is approximately −84 [dBm].

[0050] In the example shown in FIG. 7, the fusion center 1C aggregates the information indicating the detection results by the observation sensors 1B1 to 1B5 of the radio waves emitted from the positioning target 1D in the estimation stage of the position estimation system 1 (i.e., “observation information of the positioning target 1D”) and the information indicating the detection results by the observation sensors 1B1 to 1B5 of the radio waves emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN recorded in the database 1C1 in the learning stage of the position estimation system 1 (i.e., The position of the positioning target 1D is estimated by using the position fingerprint method shown in Figure 27 based on the information indicating the positions of the pre-learning points 1A1, 1A2, ..., 1AN recorded in the database 1C1 during the learning stage of the position estimation system 1 (i.e., "position coordinates of pre-learning point 1A1", "position coordinates of pre-learning point 1A2", ..., "position coordinates of pre-learning point 1AN"), and the information indicating the positions of the pre-learning points 1A1, 1A2, ..., 1AN recorded in the database 1C1 during the learning stage of the position estimation system 1 (i.e., "position coordinates of pre-learning point 1A1", "position coordinates of pre-learning point 1A2", ..., "position coordinates of pre-learning point 1AN").

[0051] For example, the fusion center 1C estimates the position of the positioning target 1D based on the similarity between the intensity distribution of radio waves emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN detected by the observation sensors 1B1 to 1B5 during the learning phase of the position estimation system 1 (such as the intensity distribution of the graph of "Preliminary information for pre-learning point 1A1" shown in Figure 7 and the intensity distribution of the graph of "Preliminary information for pre-learning point 1A2" shown in Figure 7) and the intensity distribution of radio waves emitted from the positioning target 1D detected by the observation sensors 1B1 to 1B5 during the estimation phase of the position estimation system 1 (such as the intensity distribution of the graph of "Observation information for positioning target (observation point) 1D" shown in Figure 7).

[0052] Specifically, the fusion center 1C detects the strength of the radio wave emitted from the pre-learning point 1A1 detected by the observation sensor 1B1 in the learning stage of the position estimation system 1, the strength of the radio wave emitted from the pre-learning point 1A1 detected by the observation sensor 1B2, the strength of the radio wave emitted from the pre-learning point 1A1 detected by the observation sensor 1B3, the strength of the radio wave emitted from the pre-learning point 1A1 detected by the observation sensor 1B4, and the strength of the radio wave emitted from the pre-learning point 1A1 detected by the observation sensor 1B5. 1 in ascending order (i.e., from smallest to largest), a series of "-80 [dBm] → -78 [dBm] → -64 [dBm] → -50 [dBm] → -45 [dBm]"; the strength of the radio waves emitted from pre-learning point 1A2 detected by observation sensor 1B1 in the learning stage of position estimation system 1, the strength of the radio waves emitted from pre-learning point 1A2 detected by observation sensor 1B2, the strength of the radio waves emitted from pre-learning point 1A2 detected by observation sensor 1B3, and the strength of the radio waves emitted from observation sensor 1B4. The strength of the radio wave emitted from the pre-learning point 1A2 detected by the observation sensor 1B4 and the strength of the radio wave emitted from the pre-learning point 1A2 detected by the observation sensor 1B5 are arranged in ascending order (i.e., from smallest to largest), such as "-84 [dBm] → -83 [dBm] → -60 [dBm] → -55 [dBm] → -44 [dBm]", and the strength of the radio wave emitted from the positioning target 1D detected by the observation sensor 1B1 in the estimation stage of the position estimation system 1 and the strength of the radio wave emitted from the positioning target 1D detected by the observation sensor 1B2 are arranged in ascending order (i.e., from smallest to largest), such as "-84 [dBm] → -83 [dBm] → -60 [dBm] → -55 [dBm] → -44 [dBm]". The position of the object 1D to be positioned is estimated based on the result of a vector distance calculation based on the difference between the strength of the radio waves emitted from the object 1D to be positioned detected by observation sensor 1B3, the strength of the radio waves emitted from the object 1D to be positioned detected by observation sensor 1B4, and the strength of the radio waves emitted from the object 1D to be positioned detected by observation sensor 1B5, which are arranged in ascending order as follows: ``-84 [dBm] → -83 [dBm] → -60 [dBm] → -55 [dBm] → -44 [dBm]''.

[0053] Figure 8 is a diagram for explaining a specific example in which, in a second example of the position estimation system 1 of the first embodiment, the fusion center 1C compares the series related to the pre-learning point 1A1 shown in Figure 7 with the series related to the positioning target 1D, and compares the series related to the pre-learning point 1A2 with the series related to the positioning target 1D. In the example shown in Figures 7 and 8, the fusion center 1C compares the series "-80 [dBm] → -78 [dBm] → -64 [dBm] → -50 [dBm] → -45 [dBm]" for the pre-learning point 1A1 with the series "-84 [dBm] → -83 [dBm] → -60 [dBm] → -55 [dBm] → -44 [dBm]" for the positioning target 1D. Fusion center 1C also calculates the difference between the sequence "-80 [dBm] → -78 [dBm] → -64 [dBm] → -50 [dBm] → -45 [dBm]" for pre-learning point 1A1 and the sequence "-84 [dBm] → -83 [dBm] → -60 [dBm] → -55 [dBm] → -44 [dBm]" for positioning target 1D. Fusion center 1C then performs vector distance calculation based on the difference. In the examples shown in Figures 7 and 8, the difference between the series related to the pre-learning point 1A1 and the series related to the positioning target 1D is relatively large, so the fusion center 1C does not estimate that the position of the positioning target 1D approximately coincides with the position of the pre-learning point 1A1.

[0054] Also, in the examples shown in Figures 7 and 8, the fusion center 1C compares the series "-84 [dBm] → -83 [dBm] → -60 [dBm] → -55 [dBm] → -44 [dBm]" related to the pre-learning point 1A2 with the series "-84 [dBm] → -83 [dBm] → -60 [dBm] → -55 [dBm] → -44 [dBm]" related to the positioning target 1D. Fusion center 1C also calculates the difference between the sequence "-84 [dBm] → -83 [dBm] → -60 [dBm] → -55 [dBm] → -44 [dBm]" for pre-learning point 1A2 and the sequence "-84 [dBm] → -83 [dBm] → -60 [dBm] → -55 [dBm] → -44 [dBm]" for positioning target 1D. Fusion center 1C then performs vector distance calculation based on the difference. 7 and 8, the sequence "-84 [dBm] → -83 [dBm] → -60 [dBm] → -55 [dBm] → -44 [dBm]" for pre-learning point 1A2 matches the sequence "-84 [dBm] → -83 [dBm] → -60 [dBm] → -55 [dBm] → -44 [dBm]" for positioning object 1D, and the difference between the sequence for pre-learning point 1A2 and the sequence for positioning object 1D is zero. Therefore, the vector distance calculation result based on the difference (zero) between the sequence for pre-learning point 1A2 and the sequence for positioning object 1D is zero. As a result, the fusion center 1C estimates that the position of the target 1D to be positioned approximately coincides with the position of the pre-learning point 1A2. In the above description of Figures 7 and 8, no information loss occurs in the information aggregation (PhyC-SN) during the learning stage of the location estimation system 1. However, in some cases, information loss may occur in the information aggregation (PhyC-SN) during the learning stage of the location estimation system 1. For example, as described above, if three observation sensors transmit the same information, single-level detection will determine whether the information is present or absent, and will recognize that only one observation sensor sent the information, resulting in the loss of information from two observation sensors. When information loss occurs, the length of the sequence described above will change. For example, when information indicating the strength (-75 [dBm]) of radio waves emitted from pre-learning point 1A1 detected by observation sensor 1B1 during the learning phase of position estimation system 1 (information indicating the detection result), information indicating the strength (-85 [dBm]) of radio waves emitted from pre-learning point 1A1 detected by observation sensor 1B2 (information indicating the detection result), and information indicating the strength (-85 [dBm]) of radio waves emitted from pre-learning point 1A1 detected by observation sensor 1B3 (information indicating the detection result) are simultaneously transmitted from observation sensors 1B1 to 1B3 to fusion center 1C as data "-75 dBm-85 dBm-85 dBm," information loss occurs because the information indicating the detection result (-85 [dBm]) of observation sensor 1B2 and the information indicating the detection result (-85 [dBm]) of observation sensor 1B3 are identical. Assuming the above-mentioned single level detection, the detected data will be "-75dBm-85dBm", with one data "-85dBm" missing. This occurs not only in the learning stage of the position estimation system 1, but also in the estimation stage of the position estimation system 1. Therefore, when calculating the vector distance, a series is created in descending order, and then difference calculation is performed starting from the highest value. Any remaining data is then deleted (no difference calculation is performed). Specifically, if the data simultaneously transmitted to the fusion center 1C during the learning stage of the position estimation system 1 is "-75 dBm-85 dBm-85 dBm" and the data simultaneously transmitted to the fusion center 1C during the estimation stage of the position estimation system 1 is "-75 dBm-85 dBm", the data "85 dBm" during the learning stage of the position estimation system 1 is deleted, and in the vector distance calculation, ((-75 dBm)-(-75 dBm)) 2 +((-85dBm)-(-85dBm)) 2 is calculated.

[0055] In the third example of the position estimation system 1 of the first embodiment, the position estimation system 1 is configured in a manner generally similar to the position estimation system 1 shown in FIG. As described above, in the first example of the position estimation system 1 of the first embodiment shown in Figure 1, the position estimation system 1 has nine observation sensors 1B1 to 1B9, while in the third example of the position estimation system 1 of the first embodiment, the position estimation system 1 has five observation sensors 1B1 to 1B5.

[0056] In the third example of the position estimation system 1 of the first embodiment, the observation sensors 1B1 to 1B5 detect radio waves emitted from the respective pre-learning points 1A1, 1A2, . . . , 1AN during the learning stage of the position estimation system 1. The fusion center (FC) 1C collects information indicating the detection results (in the example shown in FIG. 9, "RSSI data") of radio waves emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN by the observation sensors 1B1 to 1B5 during the learning stage of the position estimation system 1. In addition, the fusion center 1C records in the database 1C1 the aggregated information showing the detection results by the observation sensors 1B1 to 1B5 of the radio waves emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN during the learning stage of the position estimation system 1 (in the example shown in Figure 9, "pre-information for pre-learning point 1A1," "pre-information for pre-learning point 1A2," ...).

[0057] FIG. 9 is a diagram for explaining a method for estimating the position of the object 1D to be positioned by the fusion center 1C in the third example of the position estimation system 1 of the first embodiment. In the example shown in FIG. 9, during the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted by the preliminary learning point 1A1 among the observation sensors 1B1 to 1B5 with the strongest RSSI value is approximately −45 [dBm]. During the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted by the preliminary learning point 1A1 among the observation sensors 1B1 to 1B5 with the second strongest RSSI value is approximately −50 [dBm]. During the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted by the preliminary learning point 1A1 among the observation sensors 1B1 to 1B5 with the third strongest RSSI value is approximately −64 [dBm]. During the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted by the preliminary learning point 1A1 among the observation sensors 1B1 to 1B5 with the fourth strongest RSSI value is approximately −78 [dBm]. During the learning stage of position estimation system 1, the RSSI value of the observation sensor that detected the radio waves emitted by pre-learning point 1A1 with the weakest RSSI value among observation sensors 1B1 to 1B5 is approximately −80 [dBm].

[0058] In the example shown in FIG. 9, during the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted from the preliminary learning point 1A2 with the strongest RSSI value among the observation sensors 1B1 to 1B5 is approximately −44 [dBm]. During the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted from the preliminary learning point 1A2 with the second strongest RSSI value among the observation sensors 1B1 to 1B5 is approximately −55 [dBm]. During the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted from the preliminary learning point 1A2 with the third strongest RSSI value among the observation sensors 1B1 to 1B5 is approximately −60 [dBm]. During the learning phase of the position estimation system 1, the RSSI value of the observation sensor that detects the radio waves emitted from the preliminary learning point 1A2 with the fourth strongest RSSI value among the observation sensors 1B1 to 1B5 is approximately −83 [dBm]. During the learning stage of position estimation system 1, the RSSI value of the observation sensor that detected the radio waves emitted by pre-learning point 1A2 with the weakest RSSI value among observation sensors 1B1 to 1B5 is approximately −84 [dBm].

[0059] 9, during the estimation stage of the position estimation system 1, the RSSI value of the observation sensor among the observation sensors 1B1 to 1B5 that detects the radio waves emitted by the positioning target (observation point) 1D with the strongest RSSI value is approximately −44 [dBm]. During the estimation stage of the position estimation system 1, the RSSI value of the observation sensor among the observation sensors 1B1 to 1B5 that detects the radio waves emitted by the positioning target 1D with the second strongest RSSI value is approximately −55 [dBm]. During the estimation stage of the position estimation system 1, the RSSI value of the observation sensor among the observation sensors 1B1 to 1B5 that detects the radio waves emitted by the positioning target 1D with the third strongest RSSI value is approximately −60 [dBm]. During the estimation stage of the position estimation system 1, the RSSI value of the observation sensor among the observation sensors 1B1 to 1B5 that detects the radio waves emitted by the positioning target 1D with the fourth strongest RSSI value is approximately −83 [dBm]. During the estimation stage of position estimation system 1, the RSSI value of the observation sensor that detected the radio waves emitted by positioning target 1D with the weakest RSSI value among observation sensors 1B1 to 1B5 is approximately −84 [dBm].

[0060] In the example shown in FIG. 9 , the fusion center 1C aggregates the information indicating the detection results by the observation sensors 1B1 to 1B5 of the radio waves emitted from the positioning target 1D during the estimation stage of the position estimation system 1 (i.e., “observation information of the positioning target 1D”) and the information indicating the detection results by the observation sensors 1B1 to 1B5 of the radio waves emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN recorded in the database 1C1 during the learning stage of the position estimation system 1 (i.e., “RSII data”). The position of the positioning target 1D is estimated by using the position fingerprint method shown in Figure 27 based on information indicating the respective positions of pre-learning points 1A1, 1A2, ..., 1AN recorded in database 1C1, for example, during the learning stage of the position estimation system 1 (i.e., "position coordinates of pre-learning point 1A1," "position coordinates of pre-learning point 1A2," ..., "position coordinates of pre-learning point 1AN"), and information indicating the respective positions of pre-learning points 1A1, 1A2, ..., 1AN recorded in database 1C1 (i.e., "position coordinates of pre-learning point 1A1," "position coordinates of pre-learning point 1A2," ..., "position coordinates of pre-learning point 1AN").

[0061] For example, the fusion center 1C estimates the position of the positioning target 1D based on the similarity between the intensity distribution of radio waves emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN detected by the observation sensors 1B1 to 1B5 during the learning phase of the position estimation system 1 (such as the intensity distribution of the graph of ``Preliminary information of pre-learning point 1A1'' shown in Figure 9 and the intensity distribution of the graph of ``Preliminary information of pre-learning point 1A2'' shown in Figure 9) and the intensity distribution of radio waves emitted from the positioning target 1D detected by the observation sensors 1B1 to 1B5 during the estimation phase of the position estimation system 1 (such as the intensity distribution of the graph of ``Observation information of positioning target 1D'' shown in Figure 9).

[0062] Specifically, the fusion center 1C detects the strength of the radio wave emitted from the pre-learning point 1A1 detected by the observation sensor 1B1 in the learning stage of the position estimation system 1, the strength of the radio wave emitted from the pre-learning point 1A1 detected by the observation sensor 1B2, the strength of the radio wave emitted from the pre-learning point 1A1 detected by the observation sensor 1B3, the strength of the radio wave emitted from the pre-learning point 1A1 detected by the observation sensor 1B4, and the strength of the radio wave emitted from the pre-learning point 1A1 detected by the observation sensor 1B5. 1 in descending order (i.e., from largest to smallest), the strength of the radio waves emitted from pre-learning point 1A2 detected by observation sensor 1B1 in the learning stage of location estimation system 1, the strength of the radio waves emitted from pre-learning point 1A2 detected by observation sensor 1B2, the strength of the radio waves emitted from pre-learning point 1A2 detected by observation sensor 1B3, and the strength of the radio waves emitted from observation sensor 1B4. The strength of the radio waves emitted from the pre-learning point 1A2 detected by the observation sensor 1B4 and the strength of the radio waves emitted from the pre-learning point 1A2 detected by the observation sensor 1B5 are arranged in descending order (i.e., from largest to smallest), such as "-44 [dBm] → -55 [dBm] → -60 [dBm] → -83 [dBm] → -84 [dBm]", and the strength of the radio waves emitted from the positioning target 1D detected by the observation sensor 1B1 in the estimation stage of the position estimation system 1 and the strength of the radio waves emitted from the positioning target 1D detected by the observation sensor 1B2 are arranged in descending order (i.e., from largest to smallest), such as "-44 [dBm] → -55 [dBm] → -60 [dBm] → -83 [dBm] → -84 [dBm]". The position of the object 1D to be positioned is estimated based on the result of a vector distance calculation based on the difference between the strength of the radio waves emitted from the object 1D to be positioned detected by observation sensor 1B3, the strength of the radio waves emitted from the object 1D to be positioned detected by observation sensor 1B4, and the strength of the radio waves emitted from the object 1D to be positioned detected by observation sensor 1B5, which are arranged in descending order as follows: ``-44 [dBm] → -55 [dBm] → -60 [dBm] → -83 [dBm] → -84 [dBm]''.

[0063] Figure 10 is a diagram for explaining a specific example in which, in a third example of the position estimation system 1 of the first embodiment, the fusion center 1C compares the series related to the pre-learning point 1A1 shown in Figure 9 with the series related to the positioning target 1D, and compares the series related to the pre-learning point 1A2 with the series related to the positioning target 1D. In the example shown in Figures 9 and 10, the fusion center 1C compares the series "-45 [dBm] → -50 [dBm] → -64 [dBm] → -78 [dBm] → -80 [dBm]" related to the pre-learning point 1A1 with the series "-44 [dBm] → -55 [dBm] → -60 [dBm] → -83 [dBm] → -84 [dBm]" related to the positioning target 1D. Fusion center 1C also calculates the difference between the sequence "-45 [dBm] → -50 [dBm] → -64 [dBm] → -78 [dBm] → -80 [dBm]" for pre-learning point 1A1 and the sequence "-44 [dBm] → -55 [dBm] → -60 [dBm] → -83 [dBm] → -84 [dBm]" for positioning target 1D. Fusion center 1C then performs vector distance calculation based on the difference. In the example shown in Figures 9 and 10, the difference between the series related to the pre-learning point 1A1 and the series related to the positioning target 1D is relatively large, so the fusion center 1C does not estimate that the position of the positioning target 1D approximately coincides with the position of the pre-learning point 1A1.

[0064] Also, in the examples shown in Figures 9 and 10, the fusion center 1C compares the series "-44 [dBm] → -55 [dBm] → -60 [dBm] → -83 [dBm] → -84 [dBm]" related to the pre-learning point 1A2 with the series "-44 [dBm] → -55 [dBm] → -60 [dBm] → -83 [dBm] → -84 [dBm]" related to the positioning target 1D. Fusion center 1C also calculates the difference between the sequence "-44 [dBm] → -55 [dBm] → -60 [dBm] → -83 [dBm] → -84 [dBm]" for pre-learning point 1A2 and the sequence "-44 [dBm] → -55 [dBm] → -60 [dBm] → -83 [dBm] → -84 [dBm]" for positioning target 1D. Fusion center 1C then performs vector distance calculation based on the difference. 9 and 10, the sequence "-44 [dBm] → -55 [dBm] → -60 [dBm] → -83 [dBm] → -84 [dBm]" for pre-learning point 1A2 matches the sequence "-44 [dBm] → -55 [dBm] → -60 [dBm] → -83 [dBm] → -84 [dBm]" for positioning object 1D, and the difference between the sequence for pre-learning point 1A2 and the sequence for positioning object 1D is zero. Therefore, the vector distance calculation result based on the difference (zero) between the sequence for pre-learning point 1A2 and the sequence for positioning object 1D is zero. As a result, the fusion center 1C estimates that the position of the target 1D to be positioned approximately coincides with the position of the pre-learning point 1A2.

[0065] 11A and 11B are diagrams for explaining the aggregation, by time division for each area, of information indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from a positioning target 1D in a fourth example of the position estimation system 1 according to the first embodiment. Specifically, Fig. 11A shows the aggregation of information indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from a positioning target 1D in a case where time division for each area is not performed. Fig. 11B shows the aggregation of information indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from a positioning target 1D in a fourth example of the position estimation system 1 according to the first embodiment in a case where time division for each area is performed.

[0066] In a fourth example of the position estimation system 1 of the first embodiment shown in Figure 11(B), similar to the example shown in Figure 1, the position estimation system 1 includes pre-learning points 1A1, 1A2, ..., 1AN, observation sensors 1B1 to 1B9, a fusion center 1C, and a positioning target (observation point) 1D.

[0067] 11(B), in a fourth example of the position estimation system 1 of the first embodiment, similar to the example shown in FIG. 1, a positioning target (observation point) 1D emits radio waves as a detection target parameter in the estimation stage of the position estimation system 1. Observation sensors 1B1 to 1B9 detect the radio waves emitted from the positioning target 1D in the estimation stage of the position estimation system 1. In the fourth example of the position estimation system 1 of the first embodiment shown in Figure 11(B), observation sensors 1B1, 1B2, and 1B3 are arranged in a first area AR1, observation sensors 1B4, 1B5, and 1B6 are arranged in a second area AR2, and observation sensors 1B7, 1B8, and 1B9 are arranged in a third area AR3. Observation sensors 1B1, 1B2, and 1B3 included in strip-shaped first area AR1 are arranged adjacent to one another. Similarly, observation sensors 1B4, 1B5, and 1B6 included in strip-shaped second area AR2 are arranged adjacent to one another. Furthermore, observation sensors 1B7, 1B8, and 1B9 included in strip-shaped third area AR3 are arranged adjacent to one another.

[0068] Fusion center 1C aggregates information (RSSI data) indicating the detection results of radio waves emitted from object 1D by observation sensors 1B1, 1B2, and 1B3 in the estimation stage of position estimation system 1, for example, in time period TF1 in the estimation stage of position estimation system 1. Specifically, the aggregated results for time period TF1 of information (RSSI data) indicating the detection results by observation sensors 1B1 to 1B3 of radio waves emitted from the positioning target 1D during the estimation stage of the position estimation system 1 are that the number of observation sensors (more specifically, observation sensor 1B3) whose radio wave RSSI value was approximately -70 [dBm] is "1", the number of observation sensors (more specifically, observation sensor 1B2) whose radio wave RSSI value was approximately -80 [dBm] is "1", and the number of observation sensors (more specifically, observation sensor 1B1) whose radio wave RSSI value was approximately -90 [dBm] is "1".

[0069] The fusion center 1C aggregates information (RSSI data) indicating the detection results of the radio waves emitted from the positioning target 1D by the observation sensors 1B4, 1B5, and 1B6 during the estimation stage of the position estimation system 1, for example, into a time period TF2 different from the time period TF1 during the estimation stage of the position estimation system 1. Specifically, the aggregated results for time period TF2 of information (RSSI data) indicating the detection results by observation sensors 1B4 to 1B6 of radio waves emitted from the positioning target 1D during the estimation stage of the position estimation system 1 are that the number of observation sensors (more specifically, observation sensor 1B6) whose radio wave RSSI value was approximately -50 [dBm] is "1", the number of observation sensors (more specifically, observation sensor 1B5) whose radio wave RSSI value was approximately -60 [dBm] is "1", and the number of observation sensors (more specifically, observation sensor 1B4) whose radio wave RSSI value was approximately -80 [dBm] is "1".

[0070] The fusion center 1C aggregates information (RSSI data) indicating the detection results of the radio waves emitted from the positioning target 1D by the observation sensors 1B7, 1B8, and 1B9 during the estimation stage of the position estimation system 1, for example, into a time period TF3 different from the time periods TF1 and TF2 during the estimation stage of the position estimation system 1. Specifically, the aggregated results for time period TF3 of information (RSSI data) indicating the detection results by observation sensors 1B7 to 1B9 of radio waves emitted from the positioning target 1D during the estimation stage of the position estimation system 1 are that the number of observation sensors (more specifically, observation sensor 1B9) whose radio wave RSSI value was approximately -40 [dBm] is "1", the number of observation sensors (more specifically, observation sensor 1B8) whose radio wave RSSI value was approximately -50 [dBm] is "1", and the number of observation sensors (more specifically, observation sensor 1B7) whose radio wave RSSI value was approximately -60 [dBm] is "1".

[0071] 12A and 12B are diagrams illustrating the aggregation, by area, of information indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from pre-learning point 1A1 in a fourth example of position estimation system 1 according to the first embodiment, divided by time for each area, and the aggregation, by area, of information indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from pre-learning point 1A2. In detail, Fig. 12A illustrates the aggregation, without division by area, of information indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from target 1D, the aggregation of information indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from pre-learning point 1A1, and the aggregation of information indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from pre-learning point 1A2. Figure 12(B) shows an aggregation of information showing the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from the positioning target 1D in a fourth example of the first embodiment of the position estimation system 1, in which aggregation is performed by time division for each area, an aggregation of information showing the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from pre-learning point 1A1, and an aggregation of information showing the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from pre-learning point 1A2.

[0072] 12(B), similar to the example shown in FIG. 1, pre-learning point 1A1 emits radio waves as a parameter to be detected in the learning stage of position estimation system 1. Observation sensors 1B1 to 1B9 detect the radio waves emitted from pre-learning point 1A1 in the learning stage of position estimation system 1.

[0073] Fusion center 1C aggregates information (RSSI data) indicating the detection results of observation sensors 1B1, 1B2, and 1B3 of radio waves emitted from pre-learning point 1A1 during the learning stage of position estimation system 1, for example, into time period TFA during the learning stage of position estimation system 1. Specifically, during the learning stage of the position estimation system 1, the aggregated results for the time period TFA of information (RSSI data) indicating the detection results by observation sensors 1B1 to 1B3 of radio waves emitted from pre-learning point 1A1 are that the number of observation sensors (more specifically, observation sensor 1B1) whose radio wave RSSI value was approximately -40 [dBm] is "1", the number of observation sensors (more specifically, observation sensor 1B2) whose radio wave RSSI value was approximately -50 [dBm] is "1", and the number of observation sensors (more specifically, observation sensor 1B3) whose radio wave RSSI value was approximately -60 [dBm] is "1".

[0074] The fusion center 1C collects information (RSSI data) indicating the detection results of the radio waves emitted from the pre-learning point 1A1 by the observation sensors 1B4, 1B5, and 1B6 during the learning stage of the position estimation system 1, for example, into a time period TFB during the learning stage of the position estimation system 1. Specifically, during the learning stage of the position estimation system 1, the aggregated results for the time period TFB of information (RSSI data) indicating the detection results by observation sensors 1B4 to 1B6 of radio waves emitted from pre-learning point 1A1 are that the number of observation sensors (more specifically, observation sensor 1B4) whose radio wave RSSI value was approximately -50 [dBm] is "1", the number of observation sensors (more specifically, observation sensor 1B5) whose radio wave RSSI value was approximately -60 [dBm] is "1", and the number of observation sensors (more specifically, observation sensor 1B6) whose radio wave RSSI value was approximately -80 [dBm] is "1".

[0075] The fusion center 1C collects information (RSSI data) indicating the detection results of the radio waves emitted from the pre-learning point 1A1 by the observation sensors 1B7, 1B8, and 1B9 during the learning stage of the position estimation system 1, for example, into a time period TFC during the learning stage of the position estimation system 1. Specifically, during the learning stage of the position estimation system 1, the aggregated results for the time period TFC of information (RSSI data) indicating the detection results by observation sensors 1B7 to 1B9 of radio waves emitted from pre-learning point 1A1 are that the number of observation sensors (more specifically, observation sensor 1B7) whose radio wave RSSI value was approximately -70 [dBm] is "1", the number of observation sensors (more specifically, observation sensor 1B8) whose radio wave RSSI value was approximately -80 [dBm] is "1", and the number of observation sensors (more specifically, observation sensor 1B9) whose radio wave RSSI value was approximately -90 [dBm] is "1".

[0076] The fusion center 1C aggregates information (RSSI data) indicating the detection results of the radio waves emitted from the pre-learning point 1A2 by the observation sensors 1B1, 1B2, and 1B3 during the learning stage of the position estimation system 1 into, for example, a time period TFD during the learning stage of the position estimation system 1. Specifically, during the learning stage of the position estimation system 1, the aggregated results for the time period TFD of information (RSSI data) indicating the detection results by observation sensors 1B1 to 1B3 of radio waves emitted from pre-learning point 1A2 are that the number of observation sensors (more specifically, observation sensor 1B3) whose radio wave RSSI value was approximately -70 [dBm] is "1", the number of observation sensors (more specifically, observation sensor 1B2) whose radio wave RSSI value was approximately -80 [dBm] is "1", and the number of observation sensors (more specifically, observation sensor 1B1) whose radio wave RSSI value was approximately -90 [dBm] is "1".

[0077] Fusion center 1C collects information (RSSI data) indicating the detection results of observation sensors 1B4, 1B5, and 1B6 of radio waves emitted from pre-learning point 1A2 during the learning stage of position estimation system 1, for example, into time period TFE during the learning stage of position estimation system 1. Specifically, during the learning stage of the position estimation system 1, the aggregated results for time period TFE of information (RSSI data) indicating the detection results by observation sensors 1B4 to 1B6 of radio waves emitted from pre-learning point 1A2 are that the number of observation sensors (more specifically, observation sensor 1B6) whose radio wave RSSI value was approximately -50 [dBm] is "1", the number of observation sensors (more specifically, observation sensor 1B5) whose radio wave RSSI value was approximately -60 [dBm] is "1", and the number of observation sensors (more specifically, observation sensor 1B4) whose radio wave RSSI value was approximately -80 [dBm] is "1".

[0078] Fusion center 1C aggregates information (RSSI data) indicating the detection results of observation sensors 1B7, 1B8, and 1B9 of radio waves emitted from pre-learning point 1A1 during the learning stage of position estimation system 1 into time period TFF during the learning stage of position estimation system 1, for example. Specifically, during the learning stage of the position estimation system 1, the aggregated results for the time period TFF of information (RSSI data) indicating the detection results by observation sensors 1B7 to 1B9 of radio waves emitted from pre-learning point 1A2 are that the number of observation sensors (more specifically, observation sensor 1B9) whose radio wave RSSI value was approximately -40 [dBm] is "1", the number of observation sensors (more specifically, observation sensor 1B8) whose radio wave RSSI value was approximately -50 [dBm] is "1", and the number of observation sensors (more specifically, observation sensor 1B7) whose radio wave RSSI value was approximately -60 [dBm] is "1".

[0079] In a fourth example of the position estimation system 1 of the first embodiment (the example shown in Figure 12(B)), in order to estimate the position of the positioning target 1D, the fusion center 1C uses the observation information (aggregation results) of each of the first area AR1, second area AR2 and third area AR3 of the positioning target 1D shown in Figure 12, the prior information (aggregation results) of each of the first area AR1, second area AR2 and third area AR3 of the pre-learning point 1A1, the prior information (aggregation results) of each of the first area AR1, second area AR2 and third area AR3 of the pre-learning point 1A2, and information indicating the positions of each of the pre-learning points 1A1, 1A2, ..., 1AN recorded in the database 1C1.

[0080] In a fourth example of the position estimation system 1 of the first embodiment (the example shown in Figure 12(B)), the fusion center 1C estimates the position of the positioning target 1D by using the position fingerprint method shown in Figure 27 based on the aggregated results of information indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from the positioning target 1D during the estimation phase of the position estimation system 1, the aggregated results of information (RSII data) indicating the detection results by observation sensors 1B1 to 1B9 of radio waves emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN recorded in database 1C1 during the learning phase of the position estimation system 1, and information indicating the positions of each of the pre-learning points 1A1, 1A2, ..., 1AN recorded in database 1C1 during, for example, the learning phase of the position estimation system 1. In the example shown in Figure 12(B), the fusion center 1C performs comparative calculations between the observation information (aggregated results) of each of the first area AR1, second area AR2, and third area AR3 of the positioning target 1D and the prior information (aggregated results) of each of the first area AR1, second area AR2, and third area AR3 of the pre-learning point 1A1, and the prior information (aggregated results) of each of the first area AR1, second area AR2, and third area AR3 of the pre-learning point 1A2, and selects the prior information (aggregated results) of each of the first area AR1, second area AR2, and third area AR3 of the pre-learning point 1A2 that is closest to the observation information (aggregated results) of each of the first area AR1, second area AR2, and third area AR3 of the positioning target 1D, and estimates the position coordinates of the selected pre-learning point 1A2 as the position coordinates of the positioning target 1D.

[0081] In the fourth example of the position estimation system 1 of the first embodiment (the example shown in FIG. 11(B) and FIG. 12(B)), the RSSI distribution of each area can be summarized by aggregating information by time division for each area. Furthermore, by dividing the area, it is possible to distinguish between regions, thereby improving the positioning accuracy.

[0082] FIG. 13 is a diagram showing a fifth example of the position estimation system 1 according to the first embodiment. In the example shown in FIG. 13, the position estimation system 1 includes pre-learning points 1A1, 1A2, . . . , 1AN, observation sensors 1B1 to 1B20, a fusion center 1C, and a positioning target (observation point) 1D. Observation sensors 1B1 to 1B20 detect radio waves emitted from each of pre-learning points 1A1, 1A2, ..., 1AN during the learning stage of position estimation system 1. Fusion center 1C collects information indicating the detection results by observation sensors 1B1 to 1B20 of radio waves emitted from each of pre-learning points 1A1, 1A2, ..., 1AN during the learning stage of position estimation system 1. Furthermore, the observation sensors 1B1 to 1B20 detect radio waves emitted from the object 1D to be positioned during the estimation stage of the position estimation system 1. Fusion center 1C aggregates information indicating the detection results by observation sensors 1B1-1B20 of radio waves emitted from positioning target 1D during the estimation stage of position estimation system 1. Fusion center 1C also estimates the position of positioning target 1D based on the aggregated result of information indicating the detection results by observation sensors 1B1-1B20 of radio waves emitted from positioning target 1D during the estimation stage of position estimation system 1, the aggregated result of information indicating the detection results by observation sensors 1B1-1B20 of radio waves emitted from each of pre-learning points 1A1, 1A2, ..., 1AN recorded in database 1C1 during the learning stage of position estimation system 1, and information indicating the positions of each of pre-learning points 1A1, 1A2, ..., 1AN recorded in database 1C1 during the learning stage of position estimation system 1, for example.

[0083] Fig. 14 is a diagram for explaining area division (time division) etc. in a fifth example of the position estimation system 1 of the first embodiment. In detail, Fig. 14(A) shows the relationship between the observation sensors 1B1-1B20 and the areas AR1-AR4 in an example in which each of the areas AR1-AR4 is configured in a strip shape, and Fig. 14(B) shows the relationship between the observation sensors 1B1-1B20 and the areas AR1-AR4 in the fifth example of the position estimation system 1 of the first embodiment. In the fifth example of the position estimation system 1 of the first embodiment shown in Figure 14(B), observation sensors 1B1, 1B5, 1B9, 1B13, and 1B17 are arranged in a first area AR1, observation sensors 1B2, 1B6, 1B10, 1B14, and 1B18 are arranged in a second area AR2, observation sensors 1B3, 1B7, 1B11, 1B15, and 1B19 are arranged in a third area AR3, and observation sensors 1B4, 1B8, 1B12, 1B16, and 1B20 are arranged in a fourth area AR4.

[0084] The plurality of observation sensors 1B1, 1B5, 1B9, 1B13, and 1B17 included in the first area AR1 are arranged spaced apart from one another. The first area AR1 is made up of a plurality of areas that are spaced apart from one another. The plurality of observation sensors 1B2, 1B6, 1B10, 1B14, and 1B18 included in the second area AR2 are arranged spaced apart from one another. The second area AR2 is also made up of a plurality of areas that are spaced apart from one another. The plurality of observation sensors 1B3, 1B7, 1B11, 1B15, and 1B19 included in the third area AR3 are arranged spaced apart from one another. The third area AR3 is also made up of a plurality of areas that are spaced apart from one another. The plurality of observation sensors 1B4, 1B8, 1B12, 1B16, and 1B20 included in the fourth area AR4 are arranged spaced apart from one another. The fourth area AR4 is also made up of a plurality of areas that are spaced apart from one another.

[0085] Fusion center 1C aggregates information indicating the detection results of observation sensors 1B1, 1B5, 1B9, 1B13, and 1B17 of radio waves emitted from object 1D during the estimation stage of position estimation system 1, for example, into time period TF1 during the estimation stage of position estimation system 1. The fusion center 1C aggregates information indicating the detection results of the radio waves emitted from the positioning target 1D by the observation sensors 1B2, 1B6, 1B10, 1B14, and 1B18 during the estimation stage of the position estimation system 1, for example, into a time period TF2 different from the time period TF1 during the estimation stage of the position estimation system 1. The fusion center 1C aggregates information indicating the detection results of the radio waves emitted from the positioning target 1D by the observation sensors 1B3, 1B7, 1B11, 1B15, and 1B19 during the estimation stage of the position estimation system 1, for example, into a time period TF3 different from the time periods TF1 and TF2 during the estimation stage of the position estimation system 1. The fusion center 1C aggregates information indicating the detection results of the radio waves emitted from the positioning target 1D by the observation sensors 1B4, 1B8, 1B12, 1B16, and 1B20 during the estimation stage of the position estimation system 1, for example, into a time period TF4 that is different from the time periods TF1, TF2, and TF3 during the estimation stage of the position estimation system 1.

[0086] FIG. 15 is a diagram showing a sixth example of the position estimation system 1 according to the first embodiment. In the example shown in FIG. 15, the position estimation system 1 includes pre-learning points 1A1, 1A2, . . . , 1AN, observation sensors 1B1 to 1B9, a fusion center 1C, and a positioning target (observation point) 1D. Fusion center 1C includes database 1C1, histogram calculation unit 1C2, and Bhattacharyya distance calculation unit 1C3. The histogram calculation unit 1C2 generates a learning stage histogram (see FIG. 16(A)) in the learning stage of the position estimation system 1, and generates an estimation stage histogram (see FIG. 16(C)) in the estimation stage of the position estimation system 1. Specifically, when a histogram is generated in the learning stage or estimation stage of the position estimation system 1, data such as "-83 dBm-79 dBm-78 dBm-70 dBm" is acquired by information aggregation (PhyC-SN). When a histogram is calculated, the bin interval is set to, for example, "3 dB intervals." For example, a frequency value of "1" is set for "-68.5 dBm to -71.4 Bm," a frequency value of "0" is set for "-71.5 dBm to -74.4 Bm," a frequency value of "0" is set for "-74.5 dBm to -77.4 Bm," a frequency value of "2" is set for "-77.5 dBm to -80.4 Bm," and a frequency value of "1" is set for "-80.5 dBm to -83.4 Bm." Even if the above-described lack of information occurs, in the example where the Bhattacharyya distance is used, only the frequency value changes, and the calculation method remains unchanged.

[0087] FIG. 16 is a diagram showing an example of a learning stage histogram, an estimation stage histogram, etc. generated by the histogram calculation unit 1C2. In detail, Fig. 16(B) shows the intensity distribution of the detection target parameter (radio wave) emitted from pre-learning point 1A1 (not shown in Fig. 16(B)) during the learning stage of position estimation system 1. Fig. 16(A) shows a learning stage histogram (more specifically, a histogram of RSSI data) which is information indicating the detection results (RSSI values) of the detection target parameter (radio wave) emitted from pre-learning point 1A1 during the learning stage of position estimation system 1 by each of observation sensors 1B1 to 1B9. Fig. 16(D) shows the intensity distribution of detection target parameters (radio waves) emitted from positioning target (observation point) 1D (not shown in Fig. 16(D)) in the estimation stage of position estimation system 1. Fig. 16(C) shows an estimation stage histogram (more specifically, a histogram of RSSI data) which is information indicating the detection results (RSSI values) of detection target parameters (radio waves) emitted from positioning target 1D by each of observation sensors 1B1 to 1B9 in the estimation stage of position estimation system 1.

[0088] 16, histogram calculation unit 1C2 generates the learning-stage histogram shown in Fig. 16(A) based on the detection results (RSSI values) by each of observation sensors 1B1 to 1B9 of radio waves emitted from preliminary learning point 1A1 during the learning stage of position estimation system 1. In other words, the learning-stage histogram generated by histogram calculation unit 1C2 corresponds to the spread of radio waves (radio wave intensity distribution) emitted from preliminary learning point 1A1 during the learning stage of position estimation system 1. 16(C) based on the detection results (RSSI values) by each of the observation sensors 1B1 to 1B9 of the radio waves emitted from the object 1D to be positioned in the estimation stage of the position estimation system 1. That is, the estimation stage histogram generated by the histogram calculation unit 1C2 corresponds to the spread of the radio waves (radio wave intensity distribution) emitted from the object 1D to be positioned in the estimation stage of the position estimation system 1.

[0089] In the example shown in FIG. 15, the Bhattacharyya distance calculation unit 1C3 calculates the Bhattacharyya distance BD(H1, H2) indicating the similarity between the learning stage histogram (see FIG. 16(A)) generated by the histogram calculation unit 1C2 and the estimated stage histogram (see FIG. 16(C)). The Bhattacharyya distance BD(H1, H2) calculated by the Bhattacharyya distance calculation unit 1C3 is expressed, for example, by the following equation (1). In equation (1), H1 indicates the first histogram, H2 indicates the second histogram, H1(I) indicates the histogram value H1 when I, and H2(I) indicates the histogram value H2 when I. A normalized histogram is used.

[0090]

number

[0091] In the example shown in Figure 15, the fusion center 1C estimates the position of the positioning target 1D based on the Bhattacharyya distance BD(H1, H2) calculated by the Bhattacharyya distance calculation unit 1C3 and information indicating the positions of each of the pre-learning points 1A1, 1A2, ..., 1AN. In the sixth example of the location estimation system 1 of the first embodiment, the magnitude of the aggregated RSSI value can be taken into consideration. Also, the difference in the contour lines showing the spread of radio waves can be expressed by a histogram, and the difference in the surface spread can be expressed one-dimensionally.

[0092] [Example] The inventors performed an evaluation of the position estimation system 1 of the first embodiment through a simulation using ray tracing.

[0093] 17 and 18 are diagrams showing simulation parameters. An urban space measuring 800m x 800m was created as the environment. 136 observation sensors (see Figure 17) and 37 pre-training points (see Figure 18) were placed in this environment. The observation sensors were placed at intervals of approximately 50m, and the pre-training points were placed at intervals of approximately 100m. As shown in Figure 18, the positioning targets (observation points) were selected from the pre-training points that were close to each other, and placed at 20m intervals between those two points. Five such locations were created. The radio waves sent from the pre-training points and the observation points were assumed to be LPWA LoRa (registered trademark). Two things were evaluated in this simulation. The first was the number of divisions, and the second was the division method. The first simply determined how many divisions were needed to achieve good accuracy, while the second compared and evaluated two division methods. The evaluation method was based on the cumulative distribution (CDF) (vertical axis of Figure 19) of the distance (horizontal axis of Figure 19) between the actual position of the observation point (target of positioning) and the estimated position.

[0094] In the simulation results described below, "Ideal" indicates the result when the nearest point is selected as the result of position estimation. "Conventional Method" shows the results when the location estimation system and location estimation method of the present invention are not applied (i.e., the conventional location fingerprint method). The number of time slots required for aggregation is 136. "Proposed method without area division" shows the results of the method corresponding to the first to third examples of the location estimation system 1 of the first embodiment (i.e., the method applying PhyC-SN to the location fingerprinting method). The number of time slots required for aggregation is one. "Proposed method, area divided into 4, randomly" shows the results of a method corresponding to the fifth example of the location estimation system 1 of the first embodiment (i.e., a method in which PhyC-SN is applied to the location fingerprinting method and the division of Figure 14(B) is further applied). The number of time slots required for aggregation is 4. "Proposed method, area divided into 4 strips" shows the results of a method corresponding to the fourth example of the location estimation system 1 of the first embodiment (i.e., a method in which the PhyC-SN is applied to the location fingerprinting method and the division of Figure 14(A) is further applied). The number of time slots required for aggregation is 4.

[0095] Fig. 19 is a diagram showing the simulation results. In Fig. 19, the horizontal axis represents the distance between the observation point (target 1D to be positioned) and the estimation point (estimated position of target 1D to be positioned), and the vertical axis represents the CDF (cumulative distribution function). It was found that area division can achieve accuracy close to that of conventional methods. Regarding area division, it was found that dividing into bands (in other words, aggregating into bands) as in the example shown in Figure 14(A) can achieve higher accuracy than dividing into separate areas as in the example shown in Figure 14(B). It was found that aggregating into bands makes it possible to detect regional characteristics. It was found that when dividing into separate areas, the screen resolution of the individual aggregated results decreases, making them difficult to distinguish.

[0096] Fig. 20 is a diagram showing the simulation results when the histogram calculation shown in Fig. 16 is performed. In Fig. 20, the horizontal axis indicates the distance between the observation point (target 1D to be positioned) and the estimation point (estimated position of target 1D to be positioned), and the vertical axis indicates the CDF (cumulative distribution function). "Sorted by size + 4 division" shows the results of a method equivalent to a combination of the second or third example of the position estimation system 1 of the first embodiment and the fourth or fifth example of the position estimation system 1 of the first embodiment. "Histogram+4 division" shows the results of a method equivalent to a combination of the fourth or fifth example of the position estimation system 1 of the first embodiment and the sixth example of the position estimation system 1 of the first embodiment. The area was divided into four areas as shown in Fig. 14(A) or 14(B). The histogram was evaluated by dividing the data into eight areas. By dividing the area, accuracy could be improved, and we were able to achieve almost the same accuracy as the conventional method. By performing histogram calculations, we were able to achieve accuracy that exceeded that of conventional methods in some areas.

[0097] The inventors also conducted an actual test of the position estimation system 1 of the first embodiment. Figure 21 is a diagram to explain the concept of the actual device test. In the actual device test, the RSSI value from LPWA communication was used as sensor information. Taking into account the symmetry of radio waves, the RSSI obtained from each sensor was taken as the RSSI received by the sensor. In other words, it was assumed that the RSSI value from "observation sensor → aggregation station (fusion center)" was the same as the RSSI value from "aggregation station → observation sensor." It was assumed that the aggregation of sensor information was ideal. Figure 22 shows the equipment used in the actual test. In the actual test, a LoRaWAN gateway (LPS8-JP) manufactured by Dragino was used as the aggregation station, and a transmitter (LHT65) manufactured by Dragino was used as the transmitting terminal. FIG. 23 is a diagram showing the arrangement of sensors (observation sensors 1B1, 1B2, . . . ) in the actual equipment test. FIG. 24 is a diagram showing the arrangement of learning data measurement points (pre-learning points 1A1 to 1AN) and observation data measurement points (target 1D to be positioned) in the actual machine test.

[0098] Figure 25 shows a comparison of the results of the actual test. In Figure 25, the horizontal axis shows the distance between the observation point (target 1D to be positioned) and the estimation point (estimated position of target 1D to be positioned), and the vertical axis shows the CDF (cumulative distribution function). The definitions of "ideal," "conventional method," "proposed method without area division," and "proposed method with 4-area division and strip-like area" are the same as those described above. The "proposed method without area division" and the "proposed method with 4-area division and strip-like shape" achieved results equivalent to or better than the "conventional method." With a 10m interval between pre-points (pre-training points 1A1 to 1AN), an accuracy of 50% was achieved at intervals of 10m or less. A difference from the "ideal" was confirmed for the "proposed method without area division" and the "proposed method with 4-area division in a strip." It is believed that this difference can be reduced by increasing the number of sensors (observation sensors 1B1 to 1B20).

[0099] Fig. 26 is a diagram showing other simulation results. In Fig. 26, the horizontal axis represents the distance between the observation point (target 1D to be positioned) and the estimation point (estimated position of target 1D to be positioned), and the vertical axis represents the CDF (cumulative distribution function). The measurement specifications are the same as those in Figures 17 and 18. In FIG. 26, "Ideal" indicates the result when the closest point is selected as a result of position estimation. "Size order + 4 division" shows the results of a method corresponding to the fourth example of the location estimation system 1 of the first embodiment (i.e., a method in which PhyC-SN is applied to the location fingerprinting method and the division of Figure 14(A) is further applied). The number of time slots required for aggregation is 4. (This is the same as "Proposed method, area divided into 4, strip-like" in Figure 19.) Of these, this is the case where multi-level detection is applied to PhyC-SN signal detection and information is received without loss. "Barcode Strip Four-Division" shows the results of a method corresponding to the fourth example of the position estimation system 1 of the first embodiment (i.e., a method in which PhyC-SN is applied to the position fingerprinting method and the division of Figure 14(A) is further applied). The number of time slots required for aggregation is four. Of these, this is the case where single-level detection is applied to PhyC-SN signal detection. In addition, quantization is applied to divide the RSSI into 512 levels for PhyC-SN transmission. "Barcode fragmented into four" shows the results of a method corresponding to the fifth example of the position estimation system 1 of the first embodiment (i.e., a method in which PhyC-SN is applied to the position fingerprinting method and the fragmentation of Figure 14(B) is further applied). The number of time slots required for aggregation is four. Of these, this is the case where single-level detection is applied to PhyC-SN signal detection. In addition, quantization is applied to divide the RSSI into 512 levels for PhyC-SN transmission. As shown in Figure 26, it is observed that the positioning accuracy of "barcode strip 4 division" is degraded by about 5% compared to "size order + 4 division", but if a degradation of about 5% is allowed, the same accuracy can be achieved and single-level detection can be applied. That is, FIG. 26 shows a comparison of the results when using two detectors in conjunction with the multi-level and single-level detection described with reference to FIG.

[0100] Areas where the location estimation system 1 of the first embodiment is expected to be applied include radio wave environment monitoring (identifying illegal radio waves), monitoring the usage status of existing systems for frequency sharing (see the study by KDDI Research Institute), environmental monitoring (floods, fires, building structures, etc.), and indoor positioning (positioning in places where satellites cannot be used), i.e., use in production sites and commercial facilities, for example.

[0101] Second Embodiment A second embodiment of the position estimation system and position estimation method of the present invention will be described below. The position estimation system 1 of the second embodiment is configured similarly to the position estimation system 1 of the first embodiment described above, except for the points described below. Therefore, the position estimation system 1 of the second embodiment can achieve the same effects as the position estimation system 1 of the first embodiment described above, except for the points described below.

[0102] As described above, in the position estimation system 1 of the first embodiment, each of the pre-learning points 1A1, 1A2, ..., 1AN emits radio waves as a detection target parameter during the learning stage of the position estimation system 1. The observation sensors 1B1-1B9 detect the radio waves as a detection target parameter emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN during the learning stage of the position estimation system 1. Furthermore, during the estimation stage of the position estimation system 1, the positioning target 1D emits radio waves as a detection target parameter. The observation sensors 1B1-1B9 detect the radio waves as a detection target parameter emitted from the positioning target 1D. The fusion center 1C estimates the position of the positioning target 1D based on the aggregated information indicating the detection results by the observation sensors 1B1 to 1B9 of radio waves as detection target parameters emitted from the positioning target 1D during the estimation stage of the position estimation system 1, the aggregated information indicating the detection results by the observation sensors 1B1 to 1B9 of radio waves as detection target parameters emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN during the learning stage of the position estimation system 1, and information indicating the positions of each of the pre-learning points 1A1, 1A2, ..., 1AN.

[0103] On the other hand, in the position estimation system 1 of the second embodiment, each of the pre-learning points 1A1, 1A2, ..., 1AN emits heat as a parameter to be detected during the learning stage of the position estimation system 1. The observation sensors 1B1 to 1B9 detect the heat as a parameter to be detected emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN during the learning stage of the position estimation system 1. Furthermore, in the position estimation system 1 of the second embodiment, the object 1D to be positioned emits heat as a detection object parameter in the estimation stage of the position estimation system 1. The observation sensors 1B1 to 1B9 detect the heat as a detection object parameter emitted from the object 1D to be positioned. Furthermore, in the position estimation system 1 of the second embodiment, the fusion center 1C estimates the position of the positioning target 1D based on the aggregated results of information indicating the detection results by the observation sensors 1B1 to 1B9 of heat as a detection target parameter emitted from the positioning target 1D during the estimation stage of the position estimation system 1, the aggregated results of information indicating the detection results by the observation sensors 1B1 to 1B9 of heat as a detection target parameter emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN during the learning stage of the position estimation system 1, and information indicating the positions of each of the pre-learning points 1A1, 1A2, ..., 1AN.

[0104] Third Embodiment A third embodiment of the position estimation system and position estimation method of the present invention will be described below. The position estimation system 1 of the third embodiment is configured similarly to the position estimation system 1 of the first embodiment described above, except for the points described below. Therefore, the position estimation system 1 of the third embodiment can achieve the same effects as the position estimation system 1 of the first embodiment described above, except for the points described below.

[0105] In the position estimation system 1 of the third embodiment, each of the pre-learning points 1A1, 1A2, ..., 1AN emits a sound wave as a parameter to be detected during the learning stage of the position estimation system 1. The observation sensors 1B1 to 1B9 detect the sound waves as a parameter to be detected emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN during the learning stage of the position estimation system 1. Furthermore, in the position estimation system 1 of the third embodiment, the object 1D to be positioned emits sound waves as detection object parameters in the estimation stage of the position estimation system 1. The observation sensors 1B1 to 1B9 detect the sound waves as detection object parameters emitted from the object 1D to be positioned. Furthermore, in the position estimation system 1 of the third embodiment, the fusion center 1C estimates the position of the positioning target 1D based on the aggregated results of information indicating the detection results by the observation sensors 1B1 to 1B9 of sound waves as detection target parameters emitted from the positioning target 1D during the estimation stage of the position estimation system 1, the aggregated results of information indicating the detection results by the observation sensors 1B1 to 1B9 of sound waves as detection target parameters emitted from each of the pre-learning points 1A1, 1A2, ..., 1AN during the learning stage of the position estimation system 1, and information indicating the positions of each of the pre-learning points 1A1, 1A2, ..., 1AN.

[0106] Although the present invention has been described above using the embodiments, the present invention is not limited to these embodiments, and various modifications and substitutions can be made without departing from the spirit of the present invention. The configurations described in the above-described embodiments and examples may be combined.

[0107] Note that all or part of the functions of each unit of the position estimation system 1 in the above-described embodiment may be realized by recording a program for realizing these functions on a computer-readable recording medium, and reading and executing the program recorded on the recording medium into a computer system. Note that the term "computer system" here includes hardware such as an OS and peripheral devices. Furthermore, "computer-readable recording media" refers to portable media such as flexible disks, optical magnetic disks, ROMs, and CD-ROMs, as well as storage units such as hard disks built into computer systems. Furthermore, "computer-readable recording media" may also include devices that dynamically store programs for a short period of time, such as communication lines when transmitting programs over networks like the Internet or communication lines like telephone lines, or devices that store programs for a fixed period of time, such as volatile memory within computer systems that serve as servers or clients in such cases. Furthermore, the above-mentioned programs may be programs that realize some of the aforementioned functions, or may be programs that can realize the aforementioned functions in combination with programs already stored in the computer system. [Explanation of symbols]

[0108] 1...Location estimation system, 1A1 to 1AN...Pre-learning points, 1B1 to 1B20...Observation sensors, 1C...Fusion center (FC), 1C1...Database, 1C2...Histogram calculation unit, 1C3...Bhattacharyya distance calculation unit, 1D...Positioning target (observation point), AR1, AR2, AR3, AR4...Area

Claims

1. a pre-learning point that emits a detection target parameter that is one of radio waves, heat, and sound waves in a learning stage; a first observation sensor and a second observation sensor as observation sensors that detect the detection target parameters emitted from the pre-learning points in the learning stage; a fusion center that aggregates information indicating the detection results of the detection target parameters by the observation sensors issued from the pre-learning points during the learning stage, and records the aggregated information indicating the detection results of the detection target parameters by the observation sensors issued from the pre-learning points in a database; a positioning object that emits the detection object parameters in an estimation stage, the observation sensor detects the detection object parameters emitted from the positioning object in the estimation step; The fusion center, in the estimation step, aggregating information indicating the detection results of the detection object parameters emitted from the positioning object by the observation sensors; estimating a position of the object to be positioned based on an aggregated result of information indicating the detection results, by the observation sensor, of the detection object parameters emitted from the object to be positioned, an aggregated result of information indicating the detection results, by the observation sensor, of the detection object parameters emitted from the pre-learning points recorded in the database in the learning stage, and information indicating the positions of the pre-learning points; The fusion center uses a location fingerprinting method as a method for estimating the location of the target to be positioned, a fusion center uses a Physical Conversion Sensor Network (PhyC-SN) as a method for aggregating information indicating the detection results by the observation sensors of the detection target parameters emitted from the pre-learning points in the learning stage, and a method for aggregating information indicating the detection results by the observation sensors of the detection target parameters emitted from the target to be positioned in the estimation stage; The fusion center includes: generating a learning stage histogram corresponding to the spread of the detection target parameters emitted from the pre-learning points based on the intensities of the detection target parameters emitted from the pre-learning points detected by the first observation sensor in the learning stage and the intensities of the detection target parameters emitted from the pre-learning points detected by the second observation sensor; a histogramming calculation unit that generates an estimation-stage histogram corresponding to a spread of the detection object parameters emitted from the positioning object based on the intensities of the detection object parameters emitted from the positioning object detected by the first observation sensor in the estimation stage and the intensities of the detection object parameters emitted from the positioning object detected by the second observation sensor; a Bhattacharyya distance calculation unit that calculates a Bhattacharyya distance that indicates a similarity between the learning stage histogram generated by the histogram calculation unit and the estimated stage histogram, the fusion center estimates the position of the target to be positioned based on the Bhattacharyya distance calculated by the Bhattacharyya distance calculation unit and information indicating the positions of the pre-learning points. Location estimation system.

2. the position estimation system includes a plurality of observation sensors as the observation sensors, the plurality of observation sensors include at least a first observation sensor arranged in a first area and a second observation sensor arranged in a second area; The fusion center, in the estimation step, aggregating information indicating a detection result of the detection object parameter emitted from the positioning object by the first observation sensor into a first time period; aggregating information indicating a detection result of the detection object parameter emitted from the positioning object by the second observation sensor into a second time period different from the first time period; The location estimation system of claim 1 .

3. The fusion center includes: information indicating the detection results, by the first observation sensor, of the detection object parameters emitted from the positioning object, which are collected during the first time period in the estimation step; comparing the information indicating the detection results by the first observation sensor of the detection target parameters emitted from the pre-learning points, which are collected during a third time period of the learning stage; information indicating the detection results, by the second observation sensor, of the detection object parameters emitted from the positioning object, which are collected during the second time period in the estimation step; comparing the information indicating the detection results by the second observation sensor of the detection target parameters emitted from the pre-learning point, which are collected in a fourth time period different from the third time period of the learning stage; The location estimation system of claim 2 .

4. the first area includes a plurality of observation sensors as the first observation sensors; the second area includes a plurality of observation sensors as the second observation sensors; The plurality of observation sensors included in the first area are arranged adjacent to each other, and the first area is strip-shaped; the plurality of observation sensors included in the second area are arranged adjacent to each other, and the second area is strip-shaped; The location estimation system of claim 2 .

5. the first area includes a plurality of observation sensors as the first observation sensors; the second area includes a plurality of observation sensors as the second observation sensors; The plurality of observation sensors included in the first area are arranged spaced apart from each other, and the first area is configured by a plurality of areas spaced apart from each other; The plurality of observation sensors included in the second area are arranged spaced apart from each other, and the second area is configured by a plurality of areas spaced apart from each other. The location estimation system of claim 2 .

6. A method for estimating a position of an object to be positioned in a position estimation system including a pre-learning point, a first observation sensor and a second observation sensor as observation sensors, a fusion center, and the object to be positioned, the method comprising: a first step in which the pre-learning point emits a detection target parameter, which is one of radio waves, heat, and sound waves, during a learning stage; a second step in which the observation sensor detects the detection target parameter emitted from the pre-learning point in the first step during the learning stage; a third step in which the fusion center generates a learning stage histogram corresponding to the spread of the detection target parameters emitted from the pre-learning points based on the intensities of the detection target parameters emitted from the pre-learning points detected by the first observation sensor and the intensities of the detection target parameters emitted from the pre-learning points detected by the second observation sensor during the learning stage, thereby aggregating information indicating the detection results of the detection target parameters emitted from the pre-learning points detected by the observation sensor in the second step; a fourth step in which the fusion center records, in the learning stage, a summary of information indicating the detection results of the detection target parameters emitted from the pre-learning points summarized in the third step in a database; a fifth step in which the object to be positioned emits the object parameters in an estimation stage; a sixth step in which the observation sensor detects the detection object parameters emitted from the positioning object in the fifth step during the estimation stage; a seventh step in which the fusion center generates an estimation stage histogram corresponding to the spread of the detection object parameters emitted from the positioning object based on the intensities of the detection object parameters emitted from the positioning object detected by the first observation sensor and the intensities of the detection object parameters emitted from the positioning object detected by the second observation sensor in the estimation stage, thereby aggregating information indicating detection results of the detection object parameters emitted from the positioning object detected by the observation sensors in the sixth step; the fusion center includes an eighth step in the estimation step of calculating a Bhattacharyya distance indicating a similarity between the learning stage histogram and the estimation stage histogram based on the estimation stage histogram, which is an aggregation result of information indicating detection results of the detection object parameters emitted from the positioning object aggregated in the seventh step, the learning stage histogram, which is an aggregation result of information indicating detection results of the detection object parameters emitted from the pre-learning points aggregated in the third step, and information indicating positions of the pre-learning points, and estimating a position of the positioning object based on the calculated Bhattacharyya distance and the information indicating the positions of the pre-learning points; a location fingerprinting method is used as a method for the fusion center to estimate the location of the target; PhyC-SN is used as a method by which the fusion center collects information indicating the detection results of the detection target parameters transmitted from the pre-learning points in the third step, and as a method by which the fusion center collects information indicating the detection results of the detection target parameters transmitted from the positioning target in the seventh step. Location estimation method.

Images