Wireless communication system, measurement device, wireless communication method, and wireless communication program

Abstract

To provide a wireless communication system which can determine a position of an object in a target area as a coordinate from channel state information (CSI) while preventing a local large detection error in the target area from occurring.SOLUTION: A wireless communication system of the disclosure includes a transmitter, a receiver, a wireless capture device, and a measurement device. The measurement device measures teacher CSI in arranging an object in one of a plurality of measured regions in a target area divided into the measured regions, for all of the plurality of measured regions. The measurement device identifies which of a plurality of determination regions each of the plurality of measured regions belongs to, the determination regions being defined so as to divide the target area. The measurement device produces a learning model from teacher data obtained by associating the teacher CSI with a determination region label. The measurement device produces a regression function from teacher data obtained by associating the teacher CSI with a position coordinate at which the object is arranged in the measurement. The measurement device determines the position coordinate of the object from newly input CSI using the learning model and the regression function.SELECTED DRAWING: Figure 8

Description

[Technical field]

[0001] The present disclosure relates to a technique for detecting the position of an object within a target area from propagation path information of a wireless signal communicated between wireless devices. [Background technology]

[0002] Patent document 1 discloses a technology that includes a propagation channel information (Channel State Information, hereinafter referred to as CSI) measuring means and a learning model generating means, inputs newly input CSI from the CSI measuring means to the learning model, and detects the presence or absence of an object within a target area.

[0003] On the other hand, Non-Patent Document 1 discloses a technology for detecting the position of an object in a target area from CSI. In this technology, machine learning is performed using the CSI when an object is placed at a certain position and the position coordinates of the object as training data. This makes it possible to create a regression function that determines the position coordinates of an object from the acquired CSI. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent Publication No. 2021-034878 [Non-patent literature]

[0005] [Non-Patent Document 1] “Rui Zhou, et al., “Device-Free Presence Detection and Localization With SVM and CSI Fingerprinting”, IEEE SENSORS JOURNAL, VOL. 17, NO. 23, DECEMBER 1, 2017. Summary of the Invention [Problem to be solved by the invention]

[0006] However, it is known that the above-mentioned methods for detecting the position of an object can cause large detection errors locally within a target area.

[0007] In order to solve the above-mentioned problems, a first objective of the present disclosure is to provide a wireless communication system that can determine the position of an object in a target area as coordinates from propagation channel information and suppress the occurrence of large localized detection errors within the target area.

[0008] A second object of the present disclosure is to provide a measurement device that can determine the position of an object within a target area as coordinates from propagation channel information and suppress the occurrence of locally large detection errors within the target area.

[0009] A third object of the present disclosure is to provide a wireless communication method that can determine the position of an object within a target area as coordinates from propagation channel information and suppress the occurrence of large localized detection errors within the target area.

[0010] A fourth object of the present disclosure is to provide a wireless communication program that can determine the position of an object within a target area as coordinates from propagation channel information and suppress the occurrence of large localized detection errors within the target area. [Means for solving the problem]

[0011] The first aspect of the present disclosure is A wireless communication system for detecting a position of an object located in a target area based on propagation channel information included in a wireless frame transmitted and received between a transmitter and a receiver, comprising: The transmitter; The receiver; a wireless capture device for capturing the wireless frames; a measurement device for measuring propagation channel information based on the radio frames captured by the radio capture device; Equipped with The measuring device is A teacher data measurement process for measuring teacher propagation channel information for all of a plurality of measurement regions in the target area divided into the plurality of measurement regions when an object is placed in one of the measurement regions; A process of identifying to which of a plurality of determination regions defined to divide the target area each of the plurality of measurement regions belongs; A process of creating first teacher data in which the teacher propagation channel information and the label of the judgment region are associated with each other; A process of creating second teacher data in which the teacher propagation channel information is associated with position coordinates at which an object is placed in the measurement; A process of creating a learning model using the first teacher data, which determines a position of an object as a label of the determination region from input propagation channel information; A process of generating a regression function for each of the judgment regions using the second teacher data, the regression function determining a position of an object as a coordinate from input propagation channel information; a label determination process for determining a position of an object as a label of the determination region from newly input propagation channel information using the learning model; a coordinate determination process for determining a position coordinate of an object from the newly input propagation channel information by using a regression function corresponding to a determination region indicated by the determined label; Preferably, the method is configured to execute:

[0012] In addition, the second aspect is A measurement device for detecting a position of an object disposed in a target area by measuring propagation channel information included in a wireless frame transmitted and received between a transmitter and a receiver, comprising: A teacher data measurement process for measuring teacher propagation channel information for all of a plurality of measurement regions in the target area divided into the plurality of measurement regions when an object is placed in one of the measurement regions; A process of identifying to which of a plurality of determination regions defined to divide the target area each of the plurality of measurement regions belongs; A process of creating first teacher data in which the teacher propagation channel information and the label of the judgment region are associated with each other; A process of creating second teacher data in which the teacher propagation channel information is associated with position coordinates at which an object is placed in the measurement; A process of creating a learning model using the first teacher data, which determines a position of an object as a label of the determination region from input propagation channel information; A process of generating a regression function for each of the judgment regions using the second teacher data, the regression function determining a position of an object as a coordinate from input propagation channel information; a label determination process for determining a position of an object as a label of the determination region from newly input propagation channel information using the learning model; a coordinate determination process for determining a position coordinate of an object from the newly input propagation channel information by using a regression function corresponding to a determination region indicated by the determined label; Preferably, the method is configured to execute:

[0013] Moreover, the third aspect is A wireless communication method for detecting a position of an object located in a target area based on propagation channel information included in a wireless frame transmitted and received between a transmitter and a receiver, comprising: a measurement device for measuring the propagation channel information based on the radio frame, In the target area divided into a plurality of measurement regions, measuring teacher propagation channel information for all of the plurality of measurement regions when an object is placed in one of the measurement regions; Identifying to which of a plurality of determination regions defined to divide the target area each of the plurality of measurement regions belongs; creating first teacher data in which the teacher propagation channel information and the label of the judgment region are associated with each other; creating second teacher data in which the teacher propagation channel information corresponds to position coordinates at which an object is placed in the measurement; creating a learning model using the first teacher data, which determines a position of an object as a label of the determination region from input propagation channel information; creating a regression function for each of the judgment regions using the second teacher data, the regression function determining a position of an object as a coordinate from input propagation channel information; determining a position of an object as a label of the determination region from newly input propagation channel information using the learning model; determining a position coordinate of an object from the newly input propagation channel information using a regression function corresponding to a determination region indicated by the determined label; It is preferred that the compound contains

[0014] Moreover, the fourth aspect is A wireless communication program for causing a measuring device to measure propagation channel information included in a wireless frame transmitted and received between a transmitter and a receiver and detect a position of an object placed in a target area, the program comprising: The measuring device, A teacher data measurement process for measuring teacher propagation channel information for all of a plurality of measurement regions in the target area divided into the plurality of measurement regions when an object is placed in one of the measurement regions; A process of identifying to which of a plurality of determination regions defined to divide the target area each of the plurality of measurement regions belongs; A process of creating first teacher data in which the teacher propagation channel information and the label of the judgment region are associated with each other; A process of creating second teacher data in which the teacher propagation channel information is associated with position coordinates at which an object is placed in the measurement; A process of creating a learning model using the first teacher data, which determines a position of an object as a label of the determination region from input propagation channel information; A process of generating a regression function for each of the judgment regions using the second teacher data, the regression function determining a position of an object as a coordinate from input propagation channel information; a label determination process for determining a position of an object as a label of the determination region from newly input propagation channel information using the learning model; a coordinate determination process for determining a position coordinate of an object from the newly input propagation channel information by using a regression function corresponding to a determination region indicated by the determined label; It is preferable that the program includes a program for executing the above. Effect of the Invention

[0015] According to the first to fourth aspects of the present disclosure, it is possible to provide a wireless communication system, a measurement device, a wireless communication method, and a wireless communication program that can determine the position of an object in a target area as coordinates from propagation channel information and suppress the occurrence of large localized detection errors within the target area. [Brief description of the drawings]

[0016] [Figure 1] 1 is a configuration example of a wireless communication system according to a first embodiment of the present disclosure. [Diagram 2] FIG. 1 is a diagram illustrating a method for detecting the position of an object in a conventional wireless communication system. [Diagram 3] FIG. 1 is a diagram illustrating a method for detecting the position of an object in a conventional wireless communication system. [Figure 4] 13 shows the result of position detection in a conventional wireless communication system when a regression function optimized only in the central portion is adopted. [Diagram 5] 13 shows the result of position detection in a conventional wireless communication system when a regression function optimized only in the central portion is adopted. [Figure 6] FIG. 2 is a diagram illustrating a method for creating a learning model and a regression function according to the first embodiment of the present disclosure. [Figure 7] FIG. 2 is a diagram illustrating a method for creating a learning model and a regression function according to the first embodiment of the present disclosure. [Figure 8] FIG. 2 is a diagram illustrating a method for creating a learning model and a regression function according to the first embodiment of the present disclosure. [Figure 9]FIG. 2 is a diagram illustrating a method for creating a learning model and a regression function according to the first embodiment of the present disclosure. [Figure 10] FIG. 2 is a diagram illustrating a method for determining an object position by a measuring device according to the first embodiment of the present disclosure. [Figure 11] FIG. 2 is a diagram illustrating a method for determining an object position by a measuring device according to the first embodiment of the present disclosure. [Figure 12] FIG. 4 is a diagram illustrating a position detection result according to the first embodiment of the present disclosure. [Figure 13] FIG. 4 is a diagram illustrating a position detection result according to the first embodiment of the present disclosure. [Figure 14] FIG. 4 is a diagram illustrating a position detection result according to the first embodiment of the present disclosure. [Figure 15] 1 is a block diagram showing a configuration example of a measurement device according to a first embodiment of the present disclosure. [Figure 16] FIG. 2 is a diagram illustrating an example of CSI according to the first embodiment of the present disclosure. [Figure 17] FIG. 2 is a block diagram illustrating a configuration example of a creation unit according to the first embodiment of the present disclosure. [Figure 18] 4 is an example of data stored in a storage unit according to the first embodiment of the present disclosure. [Figure 19] 13 illustrates a first modified example of the arrangement of determination regions according to the first embodiment of the present disclosure. [Figure 20] 13 illustrates a first modified example of the arrangement of determination regions according to the first embodiment of the present disclosure. [Figure 21] 13 illustrates a second modified example of the arrangement of the determination regions according to the first embodiment of the present disclosure. [Figure 22] 13 illustrates a second modified example of the arrangement of the determination regions according to the first embodiment of the present disclosure. [Figure 23] 11 is a flowchart of a process executed by a measurement device to determine the arrangement of a determination region according to a second embodiment of the present disclosure. [Figure 24] FIG. 11 is a block diagram showing a configuration example of a creation unit according to a second embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0017] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present disclosure will be described with reference to the drawings. The same or corresponding components are designated by the same reference numerals, and repeated description may be omitted.

[0018] First embodiment FIG. 1 illustrates an example of a configuration of a wireless communication system 100 according to a first embodiment of the present disclosure.

[0019] A wireless LAN transmitter (hereinafter referred to as a transmitter) 110 performs wireless LAN communication based on the wireless LAN standard IEEE802.11ac / ax. The transmitter 110 has one or more antennas 111(1), 111(2), ... 111(n). The transmitter 110 uses these antennas simultaneously to perform communication with one or more wireless LAN receivers (hereinafter referred to as a receiver) 120. Note that when the transmitter 110 has one antenna 111, the receiver has two or more antennas. However, normally, the transmitter 110 has two or more antennas 111.

[0020] The transmitter 110 transmits a radio signal including a known frame to the receiver 120. The radio signal transmitted from the transmitter 110 to the receiver 120 is affected by objects present in the target area 10.

[0021] The receiver 120 receives a radio signal from the transmitter 110 using one or more antennas. Based on the received radio signal, the receiver 120 calculates a difference from a known frame and derives compressed propagation channel information. The compressed propagation channel information (Channel State Information, CSI) is information on a radio propagation path. The information on the radio propagation path includes amplitude information and phase information for each OFDM subcarrier on the propagation path between the antenna 111 of the transmitter 110 and the antenna of the receiver 120, as well as information on relative values ​​between the antennas. The information may also include SNR (Signal-to-Noise Ratio) information.

[0022] Furthermore, the receiver 120 stores the compressed CSI in a radio frame such as a Compressed Beamforming Report (CBFR) and feeds it back to the antenna 111 of the transmitter 110.

[0023] The wireless capture device 130 captures wireless frames fed back from the receiver 120 to the transmitter 110. In addition, the wireless capture device 130 outputs the captured wireless frames to the measurement device 140.

[0024] The measuring device 140 measures the compressed CSI from the received wireless frame based on the IEEE 802.11ac / ax standard. Furthermore, based on the compressed CSI, the measuring device 140 determines the position coordinates of the object. Note that the coordinates here are not limited to two-dimensional coordinates, and may be three-dimensional coordinates.

[0025] In this manner, objects located within the target area 10 are measured using the compressed CSI. The wireless capture device 130 captures the wireless frames including the compressed CSI, and the measurement device 140 analyzes the wireless frames to detect the location of the objects.

[0026] The wireless capture device 130 may be built into the transmitter 110 or the receiver 120. Furthermore, the measurement device 140 and the wireless capture device 130 may be integrated together.

[0027] The wireless frame captured by the wireless capturing device 130 does not have to be limited to the wireless frame fed back from the receiver 120 to the transmitter 110. As long as the wireless frame is included in the wireless signal transmitted and received between the transmitter 110 and the receiver 120 and the compressed CSI can be extracted from the wireless frame, the type of the wireless frame is not limited.

[0028] Comparative Example Here, we will explain why large detection errors occur locally within the target area 10 in the conventional technology. Since large detection errors are particularly likely to occur on the wall side of the target area 10, we will explain this case as an example. Figures 2 and 3 are diagrams showing a method for detecting the position of an object in a conventional wireless communication system 200. Here, as shown in Figure 2, the wireless communication system 200 is assumed to have two receivers 120, which are denoted as receiver 120(1) and receiver 120(2).

[0029] In the conventional wireless communication system 200, an object placed in the target area 10 is measured using compressed CSI transmitted and received between the transmitter 110 and the receiver 120(1) or between the transmitter 110 and the receiver 120(2). In addition, machine learning is performed using the CSI when the object is placed at a certain position and the position coordinates of the object as training data. This makes it possible to create a regression function that determines the position coordinates of the object from the newly input CSI. That is, in the conventional wireless communication system 200, the position of the object can be determined as continuous two-dimensional coordinates. FIG. 3 shows an example of two-dimensional coordinates for indicating the position of an object in the target area.

[0030] Usually, transmitter 110 and receiver 120 are almost always placed indoors. In conventional wireless communication system 200, transmitter 110 and two receivers 120 are also placed indoors and are surrounded on all sides by walls 50(1), 50(2), 50(3), and 50(4). Hereinafter, when there is no need to particularly distinguish between walls 50(1), 50(2), 50(3), and 50(4), they will be referred to as wall 50.

[0031] Here, the radio signal measuring an object in the center of the target area 10 is affected isotropically by reflections from the walls 50 and is not strongly affected by reflected waves from a specific wall 50. On the other hand, the radio signal measuring an object in the outer periphery of the target area 10 is closer to a specific wall 50 among the four walls 50(1), 50(2), 50(3), and 50(4), and is therefore strongly affected by reflections from that specific wall 50. In this way, the influence of reflected waves on the radio signal differs between the outer periphery and the center of the target area, which is the reason why large detection errors occur locally.

[0032] Specifically, a regression function created using a wireless signal in such an environment will be either a regression function optimized only in the central portion, a regression function optimized only in the peripheral portion, or a regression function optimized to some degree in both the central portion and the peripheral portion.

[0033] 4 and 5 show position detection results when a regression function optimized only in the central part is adopted in the conventional wireless communication system 200. FIG. 4 shows detection errors shown as a cumulative distribution function (CDF). The average error was 2.03 m in the outer periphery, 0.82 m in the central part, and 1.44 m for the entire target area 10. FIG. 5 is a three-dimensional diagram showing the distribution of detection errors in the target area 10. It is clear from FIG. 5 that the average error in the central part is smaller than that in the outer periphery.

[0034] As described above, in the conventional wireless communication system 200, the detection error varies greatly depending on the position of the object. If a regression function optimized to some extent is used for both the center and the outer periphery, it is expected that the distribution of errors in the center and the outer periphery will be eliminated. However, the average error in the center will be larger than the values ​​shown in Figs. 4 and 5.

[0035] 6 to 9 are diagrams for explaining a method for creating a learning model and a regression function according to the first embodiment of the present disclosure. First, a target area 10 is divided into a plurality of measurement regions 11. In the example of FIG. 6, the target area 10 is divided into 32 measurement regions 11. Note that the value of the number of divisions is an example and is not limited.

[0036] Furthermore, one or more arrangement points 12 are defined in each measurement region 11. In the example of Fig. 7, nine arrangement points 12 are defined in each measurement region 11. The measurement device 140 executes a process of measuring the teacher CSI for all of the arrangement points 12 for each of the multiple measurement regions 11 (hereinafter referred to as teacher data measurement process).

[0037] Here, each of the multiple measurement regions 11 belongs to one of the multiple judgment regions 13 defined to divide the target area 10. In the example of FIG. 8, the measurement regions 11 Nos. 1-9, 16, 17, and 24-32 located on the periphery belong to judgment regions 13 each having an individual label. On the other hand, the measurement regions 11 Nos. 10-15 and 18-23 located in the center belong to a judgment region 13 collectively having a single label. Note that the arrangement of the judgment regions 13 shown here is an example and is not limited to this.

[0038] 9 is a diagram showing the correspondence between the numbers of the measurement regions 11 and the labels of the judgment regions 13. The measurement regions 11 numbered 1-9, 16, 17, and 24-32 located on the periphery belong to the judgment regions 13 labeled "A"-"I", "K", "L", and "M"-"U", respectively. On the other hand, the measurement regions 11 numbered 10-15 and 18-23 located in the center collectively belong to the judgment region 13 labeled "J".

[0039] Furthermore, the measurement device 140 executes a process (hereinafter referred to as a first teacher data creation process) for creating first teacher data that associates the teacher CSI with the object's placement position in the measurement shown as the label of the determination region 13. Furthermore, the measurement device 140 executes a process (hereinafter referred to as a learning model creation process) for creating a learning model 60 using the first teacher data that determines the object's position as a label from the newly input CSI. Methods for creating the learning model 60 include known methods such as random forest, SVM (Support Vector Machine), K-nearest neighbor method, and neural network.

[0040] Furthermore, the measurement device 140 executes a process (hereinafter referred to as a second teacher data creation process) for creating second teacher data that associates the teacher CSI with the position coordinates where the object was placed in the measurement. Furthermore, the measurement device 140 executes a process (hereinafter referred to as a regression function creation process) for creating a regression function 70 for each judgment region 13 using the second teacher data, the regression function 70 determining the position of the object as coordinates from the newly input CSI. Examples of a method for creating the regression function 70 include the method described in Non-Patent Document 1.

[0041] 10 and 11 are diagrams for explaining a method for determining an object position by a measurement device 140 according to the first embodiment of the present disclosure. Here, as shown in FIG. 10, a case is shown in which a target area 10 is divided into five determination regions 13 with labels "A" to "E". The measurement device 140 executes a process (hereinafter referred to as a label determination process) of determining the object position as the label of the determination region 13 from the newly input CSI using a learning model 60. This allows the position of the object in the target area 10 to be roughly grasped at first.

[0042] Furthermore, the measurement device 140 executes a process (hereinafter referred to as coordinate determination process) of determining the position coordinates of the object from the newly input CSI using a regression function 70 corresponding to the determination region 13 indicated by the determined label. In the example of FIG. 11, regression functions 70 from fA to fE corresponding to the determination regions 13 of the labels "A" to "E" are created respectively. In the coordinate determination process, the position coordinates are determined using the corresponding regression function based on the label determined in the label determination process. By using the regression function 70 corresponding to each determination region 13, it is possible to prevent the occurrence of large detection errors locally.

[0043] 12 to 14 are diagrams for explaining position detection results according to the first embodiment of the present disclosure. The wall side region of the determination region 13, which is located on the wall side, is divided into 4, 6, 10, and 18 parts as shown in FIG. 12. On the other hand, the inner region, which is located inside the wall side region, is unified into 2 parts as shown in FIG. 13. FIG. 14 shows the detection error in the entire target area 10 shown as a CDF. The average error was 1.140 m when the wall side region was divided into 4 parts, 0.879 m when it was divided into 6 parts, 0.809 m when it was divided into 10 parts, and 0.627 m when it was divided into 18 parts. As described above, in the conventional technology, the average error in the entire target area 10 was 1.44 m. It is clear that the detection error has been improved in the present disclosure. In the present disclosure, the average error in the entire target area 10 can also be improved by preventing the occurrence of large detection errors locally.

[0044] 15 is a block diagram showing an example of a configuration of a measuring device 140 according to the first embodiment of the present disclosure. An extracting unit 41 extracts compressed CSI from a wireless frame output from the wireless capturing device 130. A preprocessing unit 42 performs preprocessing, which will be described later, on the compressed CSI. Furthermore, if the CSI is a teacher CSI, the preprocessing unit 42 outputs the preprocessed CSI to a creating unit 43. On the other hand, if the CSI is CSI measured for object position detection, the preprocessing unit 42 outputs the preprocessed CSI to a label determining unit 44.

[0045] The creation unit 43 executes the above-mentioned first teacher data creation process, second teacher data creation process, learning model creation process, and regression function creation process based on the teacher CSI received from the preprocessing unit 42. Furthermore, it outputs the created learning model 60 and regression function 70 to the label determination unit 44 and the position determination unit 45, respectively.

[0046] The label determination unit 44 executes the above-mentioned label determination process based on the learning model 60 received from the creation unit 43. In addition, the label determination unit 44 outputs the determined label to the position determination unit 45. The position determination unit 45 executes the above-mentioned coordinate determination process based on the regression function 70 received from the creation unit 43.

[0047] 16 is a diagram illustrating an example of CSI according to the first embodiment of the present disclosure. The diagram illustrates combinations of the number of antennas (denoted as i) of the transmitter 110 and the number of antennas (denoted as j) of the receiver 120, and the number and type of compressed CSI obtained according to the combination. Here, φij is the phase difference of the radio signal between the antenna 111 of the transmitter 110 and the antenna of the receiver 120. Also, Ψij is a value expressed as the arctangent of the absolute value of the amplitude ratio of the radio signal between the antenna 111 of the transmitter 110 and the antenna of the receiver 120.

[0048] The preprocessing performed by the preprocessing unit 42 may include, for example, calculating cosine and sine from φij and Ψij to create vectors. Alternatively, a CSI matrix may be created from φij and Ψij, and a correlation matrix may be created from the CSI matrix. In addition, a subcarrier average may be calculated for the correlation matrix. Also, an SNR may be extracted from φij and Ψij. Furthermore, a moving average may be calculated for these parameters. Alternatively, φij and Ψij may be output as they are without being preprocessed.

[0049] 17 is a block diagram showing a configuration example of the creation unit 43 according to the first embodiment of the present disclosure. The teacher data creation unit 431 identifies whether the coordinates belong to the multiple measurement regions 11 or the multiple judgment regions 13, based on the teacher CSI received from the preprocessing unit 42 and the coordinates (x, y) of the object in the measurement of the CSI. Furthermore, the teacher data creation unit 431 executes a first teacher data creation process and a second teacher data creation process. Furthermore, the created first teacher data, second teacher data, the number of the measurement region 11, and information on the label of the judgment region 13 are output to the storage unit 432.

[0050] The storage unit 432 stores the data received from the teacher data creation unit 431. The storage unit 432 outputs the teacher CSI and information on the label of the determination region 13 to the learning model creation unit 434. In addition, the storage unit 432 outputs the teacher CSI and the coordinates (x, y) of the object to the regression function creation unit 435.

[0051] The learning model creation unit 434 executes a learning model creation process to create the learning model 60. Furthermore, the regression function creation unit 435 executes a regression function creation process to create the regression function 70.

[0052] 18 is an example of data stored in the storage unit 432 according to the first embodiment of the present disclosure. For the measurement of the teacher CSI performed for each arrangement point 12, the CSI, the object coordinates (x1, y1), the number of the measurement area 11, and the label of the judgment area 13 are stored. The CSI is a vector such as cosφij, sinφij, cosΨij, and sinΨij.

[0053] The processing performed by the measuring device 140 may be executed by a communication program using a computer having a CPU and memory and storing a program in the memory. Alternatively, the processing may be executed by a communication program using an integrated circuit such as an FPGA (Field Programmable Gate Array). The communication program may be provided by being recorded on a storage medium, or may be provided through a network. This point is common to all the following embodiments.

[0054] As described above, according to this embodiment, it is possible to provide a wireless communication system, a measurement device, a wireless communication method, and a wireless communication program that can determine the position of an object in a target area as coordinates from propagation channel information and suppress the occurrence of large localized detection errors within the target area.

[0055] As described above, the arrangement of the determination region 13 is not limited to the example in Fig. 8. In the following, several modified examples of the arrangement of the determination region 13 will be shown.

[0056] <Variation 2> 19 and 20 show a first modified example of the arrangement of the determination regions 13 according to the first embodiment of the present disclosure. Here, the arrangement of the determination regions 13 is determined according to the distance from the wall 50. As shown in FIG. 19, the measurement regions 11 numbered 1-9, 16, 17, and 24-32 belong to the determination regions 13 each having an individual label. This is because these measurement regions 11 are wall side regions particularly close to any one of the walls 50(1), 50(2), 50(3), and 50(4), and the distances from the particularly close wall 50 can be said to be approximately equal to each other.

[0057] On the other hand, the measurement areas 11 at the center, numbered 10-12 and numbered 21-23, belong to the judgment area 13, which has a single label, because the distances of these measurement areas 11 to the walls 50(1), 50(2), and 50(4) can be said to be approximately equal to each other.

[0058] Similarly, the measurement areas 11 at the center, numbered 13-15 and numbered 18-20, belong collectively to a judgment area 13 having a single label. This is because the distances of these measurement areas 11 to the walls 50(2), 50(3), and 50(4) can be said to be approximately equal to each other.

[0059] Fig. 20 is a diagram showing the correspondence between the labels of the determination area 13 determined by the method explained in Fig. 19 and the numbers of the measurement area 11. Since the diagram can be read in the same way as Fig. 9, the explanation will be omitted.

[0060] <Variation 3> 21 and 22 show a second modified example of the arrangement of the determination region 13 according to the first embodiment of the present disclosure. Here, as shown in Fig. 21, each of the antennas 111 of the transmitter 110 is connected to an antenna of the receiver 120 by lines 81, 82, 83, and 84. In addition, the determination region 13 is set so that the determination region 13 through which the lines 81, 82, 83, and 84 do not pass (hereinafter referred to as a non-passing region) is smaller than the determination region 13 through which the lines 81, 82, 83, and 84 pass (hereinafter referred to as a passing region).

[0061] For example, measurement regions 11 Nos. 25-27 are collectively classified as one judgment region 13. Measurement regions 11 Nos. 17-21 are also collectively classified as one judgment region 13. These are all measurement regions 11 through which line 81 passes. The same applies to measurement regions 11 through which lines 81, 82, and 83 pass.

[0062] On the other hand, the measurement regions 11 at numbers 1-5 and 28-32 are each classified as individual judgment regions 13. These measurement regions 11 are regions through which none of the lines 81, 82, 83, 84 passes.

[0063] Fig. 22 is a diagram showing the correspondence between the labels of the determination area 13 determined by the method explained in Fig. 21 and the numbers of the measurement area 11. Since the diagram can be read in the same way as Fig. 9, the explanation will be omitted.

[0064] Embodiment 2 In this embodiment, the arrangement of the determination regions 13 is determined based on the results of detection errors. In the following, changes from the first embodiment will be described.

[0065] 23 is a flowchart of a process executed by the measurement device 140 to determine the arrangement of the determination regions 13 according to the second embodiment of the present disclosure. First, the process is started. Next, the arrangement of the determination regions 13 is determined (step S01). Here, it is assumed that the multiple determination regions 13 are labeled with "A", "B", "C", ....

[0066] Next, the CSI when an object is placed at the position coordinates where the object was placed in the teacher data measurement process is measured for all of the measurement areas 11, and is set as a confirmation CSI (step S02). Furthermore, using the learning model 60 and the regression function 70, a label determination process and a coordinate determination process are respectively performed on the confirmation CSI (step S03).

[0067] Furthermore, the determined position coordinates of the object are compared with the position coordinates where the object was placed in the corresponding measurement, and the detection error is calculated for all of the multiple measurement regions 11 (step S04). Furthermore, the calculated detection errors are averaged for all of the multiple measurement regions 11 to obtain an average error in the entire target area 10 (step S05).

[0068] Furthermore, the average error in the judgment region 13 of the label "A" is calculated (step S06). Here, the detection error is averaged for the measurement region 11 included in the judgment region 13 of the label "A". Next, the difference between the average error in the judgment region 13 of the label "A" and the average error in the entire target area 10 is calculated (step S07).

[0069] Next, it is calculated whether the obtained difference is equal to or less than a threshold value having a positive sign (step S08). If the difference is equal to or less than the threshold value, the decision region 13 with the label "A" determined in step S01 and the corresponding regression function 70 are adopted (step S09). Furthermore, steps S06 to S09 are executed for all decision regions 13. On the other hand, if the difference is greater than the threshold value, steps S06 to S09 are also executed for all decision regions 13.

[0070] When steps S06 to S09 have been executed for all the judgment regions 13, it is judged whether or not all the judgment regions 13 and the corresponding regression functions 70 have been determined (step S10). If they have been determined, the process ends. On the other hand, if there are undetermined judgment regions 13 and their corresponding regression functions, steps S01 and after are recursively executed only for those judgment regions 13 (step S11).

[0071] In step S07, the difference between the average error in the determination region 13 of the label "A" and the average error in the entire target area 10 is calculated. However, the amount is not limited to the difference as long as it allows a comparison of the average error in the determination region 13 of the label "A" and the average error in the entire target area 10. For example, the ratio of the average error in the determination region 13 of the label "A" to the average error in the entire target area 10 may be used. In this case, the threshold in step S08 may be 1 or more.

[0072] 24 is a block diagram showing a configuration example of the creation unit 43 according to the second embodiment of the present disclosure. The creation unit 43 of this embodiment further includes a label determination unit 436, a position determination unit 437, an error calculation unit 438, and an average error calculation unit 439 in addition to the configuration example of the first embodiment. The label determination unit 436 performs a label determination process on the measured confirmation CSI. The position determination unit 437 performs a coordinate determination process on the measured confirmation CSI. The error calculation unit 438 calculates the detection error in each measurement region 11. The average error calculation unit 439 calculates the average error in the entire target area 10 and the average error for each determination region 13.

[0073] As described above, in this embodiment, the layout of the determination region 13 is determined based on the actual detection error. This makes it possible to divide the determination region 13 and determine the layout until the detection error becomes small, thereby making it possible to improve the detection accuracy more than in the first embodiment.

[0074] The present disclosure is not limited to the above-described embodiment, and various modifications can be made in the implementation stage without departing from the spirit of the present disclosure. In addition, the embodiments may be implemented in appropriate combination, and in that case, the combined effects can be obtained. [Explanation of symbols]

[0075] 10 target area, 11 measurement area, 12 placement point, 13 judgment area, 41 extraction unit, 42 preprocessing unit, 43 creation unit, 44 label determination unit, 45 position determination unit, 50 wall, 60 learning model, 70 regression function, 81, 82, 83, 84 straight line, 100 wireless communication system, 110 transmitter, 111 antenna, 120 receiver, 130 wireless capture device, 140 measuring device, 200 conventional wireless communication system, 431 teacher data creation unit, 432 memory unit, 434 learning model creation unit, 435 regression function creation unit, 436 label determination unit, 437 position determination unit, 438 error calculation unit, 439 average error calculation unit

Claims

1. A wireless communication system for detecting a position of an object located in a target area based on propagation channel information included in a wireless frame transmitted and received between a transmitter and a receiver, comprising: The transmitter; The receiver; a wireless capture device for capturing the wireless frames; a measurement device for measuring propagation channel information based on the radio frames captured by the radio capture device; Equipped with The measuring device is A teacher data measurement process for measuring teacher propagation channel information for all of a plurality of measurement regions in the target area divided into the plurality of measurement regions when an object is placed in one of the measurement regions; A process of identifying to which of a plurality of determination regions defined to divide the target area each of the plurality of measurement regions belongs; A process of creating first teacher data in which the teacher propagation channel information and the label of the judgment region are associated with each other; A process of creating second teacher data in which the teacher propagation channel information corresponds to position coordinates at which an object is placed in the measurement; A process of creating a learning model using the first teacher data, which determines a position of an object as a label of the determination region from input propagation channel information; A process of generating a regression function for each of the judgment regions using the second teacher data, the regression function determining a position of an object as a coordinate from input propagation channel information; a label determination process for determining a position of an object as a label of the determination region from newly input propagation channel information using the learning model; a coordinate determination process for determining a position coordinate of an object from the newly input propagation channel information by using a regression function corresponding to a determination region indicated by the determined label; A wireless communication system configured to:

2. Each of the plurality of measurement regions includes a plurality of arrangement points; The wireless communication system according to claim 1 , wherein the teacher data measurement process includes a process of measuring the teacher propagation channel information for all of the arrangement points for each of the plurality of measurement areas.

3. 3. The wireless communication system according to claim 1, wherein the plurality of determination regions include a wall side region arranged on a wall side surrounding the target area and an inner region arranged inside the wall side region, and the wall side region is set to be smaller than the inner region.

4. 3. The wireless communication system according to claim 1, wherein the plurality of determination areas include a passing area through which a straight line connecting the antenna of the transmitter and the antenna of the receiver passes, and a non-passing area through which the straight line does not pass, and the non-passing area is set to be smaller than the passing area.

5. The measuring device is a process of measuring propagation channel information for all of the plurality of measurement areas when an object is placed at the position coordinates where the object was placed in the teacher data measurement process, and obtaining the propagation channel information for confirmation; determining a position of an object as a label of the determination region from the confirmation propagation channel information using the learning model; determining a position coordinate of an object from the confirmation propagation channel information using a regression function corresponding to a determination region indicated by the determined label; comparing the determined object position coordinates with the position coordinates where the object was located in the corresponding measurement and calculating a detection error for all of the plurality of measurement areas; A process of averaging the calculated detection errors for all of the plurality of measurement regions to obtain an average error for the entire target area; A process of averaging the calculated detection errors for each of the determination regions to obtain an average error for each of the determination regions; comparing the average error for each decision region with the average error for the entire target area; If the average error for each of the judgment regions is equal to or less than a specified value based on the comparison, a process of employing the plurality of determination regions dividing the target area and the regression functions corresponding to the determination regions in the label determination process and the coordinate determination process; The wireless communication system of claim 1 or 2, further configured to perform the following:

6. A measurement device for detecting a position of an object disposed in a target area by measuring propagation channel information included in a wireless frame transmitted and received between a transmitter and a receiver, comprising: A teacher data measurement process for measuring teacher propagation channel information for all of a plurality of measurement regions in the target area divided into the plurality of measurement regions when an object is placed in one of the measurement regions; A process of identifying to which of a plurality of determination regions defined to divide the target area each of the plurality of measurement regions belongs; A process of creating first teacher data in which the teacher propagation channel information and the label of the judgment region are associated with each other; A process of creating second teacher data in which the teacher propagation channel information corresponds to position coordinates at which an object is placed in the measurement; A process of creating a learning model using the first teacher data, which determines a position of an object as a label of the determination region from input propagation channel information; A process of generating a regression function for each of the judgment regions using the second teacher data, the regression function determining a position of an object as a coordinate from input propagation channel information; a label determination process for determining a position of an object as a label of the determination region from newly input propagation channel information using the learning model; a coordinate determination process for determining a position coordinate of an object from the newly input propagation channel information by using a regression function corresponding to a determination region indicated by the determined label; The measurement device is configured to perform the following steps:

7. A wireless communication method for detecting a position of an object located in a target area based on propagation channel information included in a wireless frame transmitted and received between a transmitter and a receiver, comprising: a measurement device for measuring the propagation channel information based on the radio frame, measuring teacher propagation channel information for all of a plurality of measurement regions in a target area divided into the plurality of measurement regions when an object is placed in one of the measurement regions; Identifying to which of a plurality of determination regions defined to divide the target area each of the plurality of measurement regions belongs; creating first teacher data in which the teacher propagation channel information and the label of the judgment region are associated with each other; creating second teacher data in which the teacher propagation channel information corresponds to position coordinates at which an object is placed in the measurement; creating a learning model using the first teacher data, which determines a position of an object as a label of the determination region from input propagation channel information; creating a regression function for each of the judgment regions using the second teacher data, the regression function determining a position of an object as a coordinate from input propagation channel information; determining a position of an object as a label of the determination region from newly input propagation channel information using the learning model; determining a position coordinate of an object from the newly input propagation channel information using a regression function corresponding to a determination region indicated by the determined label; A wireless communication method comprising:

8. A wireless communication program for causing a measuring device to measure propagation channel information included in a wireless frame transmitted and received between a transmitter and a receiver and detect a position of an object placed in a target area, the program comprising: The measuring device, A teacher data measurement process for measuring teacher propagation channel information for all of a plurality of measurement regions in the target area divided into the plurality of measurement regions when an object is placed in one of the measurement regions; A process of identifying to which of a plurality of determination regions defined to divide the target area each of the plurality of measurement regions belongs; A process of creating first teacher data in which the teacher propagation channel information and the label of the judgment region are associated with each other; A process of creating second teacher data in which the teacher propagation channel information corresponds to position coordinates at which an object is placed in the measurement; A process of creating a learning model using the first teacher data, which determines a position of an object as a label of the determination region from input propagation channel information; A process of generating a regression function for each of the judgment regions using the second teacher data, the regression function determining a position of an object as a coordinate from input propagation channel information; a label determination process for determining a position of an object as a label of the determination region from newly input propagation channel information using the learning model; a coordinate determination process for determining a position coordinate of an object from the newly input propagation channel information by using a regression function corresponding to a determination region indicated by the determined label; A program for wireless communication, comprising a program for executing the above.

Images