Terminal identification device, terminal identification method, and program

Abstract

A terminal identification device 1 comprises: an acquisition unit 11 that acquires communication information indicating the communication quality of a communication terminal 4 in a predetermined period from a device 2 that performs wireless communication with the communication terminal 4, and acquires sensor information in the predetermined period from a sensor 3; a quality analysis unit 12 that analyzes the communication information and specifies the point in time at which the communication quality of the communication terminal 4 sharply delcines; a sensor analysis unit 13 that, when object regions of a plurality of objects detected from the sensor information at the point in time overlap, determines whether shielding is present on the basis of the size of the overlapping region; and an identification unit 14 that, when the sensor analysis unit 12 determines that there is shielding, associates the communication terminal 4 with an object on the shielded side of the overlapping region.

Description

Terminal identification device, terminal identification method, and program

[0001] The present disclosure relates to a terminal identification device, a terminal identification method, and a program.

[0002] There is a high demand for estimating the location of communication terminals, and it is used in various situations, such as understanding people flow or behavior, predicting communication quality, etc. For this reason, terminal location estimation technologies have been proposed that use communication quality information such as the reception strength of communication terminals and spatial information from cameras, etc.

[0003] Non-Patent Document 1 proposes a method for estimating the position of a communication terminal using video information acquired from a camera and received electromagnetic wave information output from the communication terminal.

[0004] Alexandre Alahi, Albert Haque, Li Fei-Fei, “RGB-W: When Vision MeetsWireless,” 2015 IEEE International Conference on Computer Vision (ICCV) “Godard et.al, “ Unsupervised Monocular Depth Estimation with Left-Right Consistency”, IEEE / CVF CVPR, 2017, Internet <https: / / openaccess.thecvf.com / content_cvpr_2017 / papers / Godard_Unsupervised_Monocular_Depth_CVPR_2017_paper.pdf> Zhan et.al, “A Flexible New Technique for Camera Calibration”, IEEE TPAMI, 2000

[0005] The method of Non-Patent Document 1 assumes the installation of multiple access points, and has a problem of narrow applicability. Furthermore, Non-Patent Document 1 does not consider the influence of shadowing when using high-frequency millimeter-wave bands such as 5G. Therefore, in next-generation wireless communications where high-frequency bands will be the mainstream, there is a possibility that the estimation accuracy will decrease when estimating the location of a terminal using the method of Non-Patent Document 1.

[0006] The present disclosure has been made in consideration of the above circumstances, and the purpose of the present disclosure is to provide a technology that can identify a terminal with high accuracy, even in communications that utilize high frequency bands, taking into account the effects of shadowing.

[0007] In order to achieve the above object, one aspect of the present disclosure is a terminal identification device comprising: an acquisition unit that acquires communication information indicating the communication quality of the communication terminal over a predetermined period from a device that communicates wirelessly with the communication terminal, and acquires sensor information for the predetermined period from a sensor; a quality analysis unit that analyzes the communication information and identifies the point in time when the communication quality of the communication terminal suddenly deteriorates; a sensor analysis unit that, when object areas of multiple objects detected from the sensor information at the point in time overlap, determines whether or not there is occlusion based on the size of the overlapping area; and, when the sensor analysis unit determines that there is occlusion, associates the communication terminal with the object on the occluded side of the overlapping area.

[0008] One aspect of the present disclosure is a terminal identification method performed by a terminal identification device, which acquires communication information indicating the communication quality of the communication terminal over a predetermined period from a device that communicates wirelessly with the communication terminal, and acquires sensor information for the predetermined period from a sensor, analyzes the communication information, identifies a point in time when the communication quality of the communication terminal suddenly deteriorated, and, if the object areas of multiple objects detected from the sensor information at the point in time overlap, determines whether or not there is occlusion based on the size of the overlapping overlapping area, and, if it is determined that there is occlusion, associates the communication terminal with the object on the occluded side of the overlapping area.

[0009] One aspect of the present disclosure is a program that causes a computer to function as an acquisition unit that acquires communication information indicating the communication quality of the communication terminal over a predetermined period from a device that communicates wirelessly with the communication terminal and acquires sensor information for the predetermined period from a sensor; a quality analysis unit that analyzes the communication information and identifies the point in time when the communication quality of the communication terminal suddenly deteriorated; a sensor analysis unit that, when object areas of multiple objects detected from the sensor information at the point in time overlap, determines whether or not there is occlusion based on the size of the overlapping area; and, when the sensor analysis unit determines that there is occlusion, associates the communication terminal with the object on the occluded side of the overlapping area.

[0010] According to the present disclosure, it is possible to provide a technology that can identify a terminal with high accuracy even in communications that utilize high frequency bands, taking into account the effects of obstructions.

[0011] FIG. 1 is a functional block diagram showing a configuration for terminal identification according to this embodiment. FIG. 2A is a graph showing an example of communication information of a communication terminal. FIG. 2B is a diagram showing an example of sensor information at time Tb in FIG. 2A. FIG. 3 is a diagram showing an example of association between an object and a communication terminal. FIG. 4 is a flowchart showing terminal identification processing. FIG. 5 is a diagram showing an example of communication information and sensor information of a communication terminal. FIG. 6 is an explanatory diagram for explaining a process for generating virtual sensor information. FIG. 7 is an example of a hardware configuration.

[0012] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. In the description of the drawings, the same parts are designated by the same reference numerals and the description thereof will be omitted.

[0013] 1 is a diagram illustrating an overall configuration of an example of a system according to the present embodiment. In this embodiment, assuming wireless communication using a high frequency band such as millimeter waves, where the influence of obstruction is strong, a communication terminal is identified using a sudden decrease in communication quality due to obstruction and obstruction information acquired from sensor information as feature quantities.

[0014] The illustrated system includes a processing device 1 (terminal identification device), an access point 2, a sensor 3, and at least one communication terminal 4. The processing device 1, the access point 2, and the sensor 3 are communicatively connected via a network.

[0015] The access point 2 is a device capable of communicating with the communication terminal 4. Specifically, the access point 2 is a device capable of transmitting and receiving Wi-Fi radio waves, and connects the communication terminal 4 to the Internet or the like. Note that the access point 2 is not limited to the access point 2, and may be a base station, a computer, or the like, as long as it is a device capable of communicating with the communication terminal 4.

[0016] The sensor 3 is a device capable of acquiring spatial information. The sensor 3 senses an area where the communication terminal 4 may be present. For example, an RGB-D camera can be used as the sensor 3 in this embodiment. The RGB-D camera is a type of depth camera. The RGB-D camera outputs RGB-D video including both color (RGB) and depth (D) as sensor information. The RGB-D camera will also be referred to as a "camera" in the following description. The sensor 3 may be a LiDAR. The LiDAR uses light to measure the distance and shape of an object with high spatial resolution, and outputs the measurement result, a LiDAR video, as sensor information.

[0017] The sensor 3 of this embodiment needs to acquire depth information, but Non-Patent Document 2 proposes a method of inferring depth information from camera images of an RGB camera. When using such a method, not only an RGB-D camera or LiDAR capable of acquiring depth information, but also a general RGB camera may be used as the sensor 3.

[0018] The communication terminal 4 is a wireless communication terminal such as a smartphone, etc. The communication terminal 4 of this embodiment performs wireless communication with the access point 2, such as a Wi-Fi connection.

[0019] The processing device 1 identifies the communication terminal 4 based on the communication information acquired from the access point 2 and the sensor information acquired from the sensor 3. Specifically, the processing device 1 associates the communication terminal 4 connected to the access point 2 with an object captured in the sensor information (such as a camera image).

[0020] In this way, in this embodiment, by combining rich features estimated from sensor information with communication information, the communication terminal 4 can be identified and applied to a variety of applications such as predicting wireless communication quality, recommendations based on understanding human behavior, and crime prevention.

[0021] The processing device 1 shown in the figure includes an acquisition unit 11, a quality analysis unit 12, a sensor analysis unit 13, an identification unit 14, and a storage unit 16. The processing device 1 may also include a generation unit 15.

[0022] The acquisition unit 11 acquires communication information indicating the communication quality of the communication terminal 4 for a predetermined period (predetermined time) from the access point 2 (device) that wirelessly communicates with the communication terminal 4, and also acquires sensor information for the predetermined period from the sensor 3. For example, received power (RSSI: Received Signal Strength Indicator, RSRP: Reference Signal Received Power) may be used as the communication information indicating the communication quality.

[0023] The sensor information in this embodiment may be RGB-D video captured in a communication area where the communication terminal 4 may be present. The sensor information may include objects. Examples of objects include people carrying the communication terminal 4 and mobile objects carrying or equipped with a communication terminal (robots, animals, vehicles, bicycles, motorcycles, etc.). The objects may also include people and mobile objects that do not carry the communication terminal 4.

[0024] Although an object is an object to be associated with a communication terminal 4, if the object exists between the communication terminal 4 and the access point 2, it becomes an obstacle that causes obstruction, and rapidly degrades the communication quality between the obstructed communication terminal 4 and the access point 2. In particular, in the case of wireless communication using a high frequency band such as millimeter waves, where the effects of obstruction are stronger, the rapid degradation of communication quality due to obstruction is significant.

[0025] The quality analysis unit 12 analyzes the communication information for a predetermined period acquired by the acquisition unit 11, and identifies the point in time when the communication quality of the communication terminal 4 suddenly deteriorates.

[0026] 2A is a graph showing an example of communication information of a certain communication terminal 4. The horizontal axis represents time (time), and the vertical axis represents communication quality (e.g., received power). In the example shown, it can be seen that the communication quality suddenly deteriorates at time Tb. The quality analysis unit 12 analyzes the communication information, and when the communication quality suddenly deteriorates (changes), it infers that obstruction has occurred, and identifies the time when the communication quality suddenly deteriorated.

[0027] For example, if the communication quality drops below a predetermined value within a short time span and then falls below a threshold, the quality analysis unit 12 may determine that the communication quality has dropped sharply, and identify the time point at which the sharp drop reached its minimum value as the time point at which occlusion occurred. In the illustrated example, the quality analysis unit 12 estimates that occlusion occurred at time Tb when the sharp change in communication quality occurred.

[0028] The sensor analysis unit 13 estimates the occurrence of occlusion using the sensor information. That is, when the object areas of multiple objects detected from the sensor information at the time specified by the quality analysis unit 12 overlap, the sensor analysis unit 13 determines whether or not occlusion occurs based on the size of the overlapping area.

[0029] In this embodiment, the sensor analysis unit 13 detects objects from the sensor information at the time specified by the quality analysis unit 12 and determines whether occlusion occurs based on the overlap (overlap) of each detected object. Specifically, the sensor analysis unit 13 detects each object shown in the sensor information and identifies an object area surrounding each detected object. Then, for each object area, the sensor analysis unit 13 determines whether the object area overlaps partially or entirely with other object areas. If none of the object areas overlap, the sensor analysis unit 13 determines that no occlusion occurs at that time.

[0030] On the other hand, when an object region overlaps at least a part of another object region, the sensor analysis unit 13 determines whether or not occlusion has occurred based on the size of the overlapping region (degree of overlap).

[0031] 2B is a diagram showing an example of sensor information (camera images) at time Tb in FIG. 2A. From the sensor information, the sensor analysis unit 13 detects three human objects A, B, and C. Here, the sensor analysis unit 13 sets rectangular regions including each of the objects A, B, and C as object regions AR, BR, and CR, respectively.

[0032] In the illustrated example, the object region BR of object B does not overlap with the other object regions AR and CR. On the other hand, the object region AR of object A partially overlaps with the object region CR of another object C. In this case, the sensor analysis unit 13 determines whether or not there is occlusion based on the size of the overlapping region where the object regions AR and CR overlap.

[0033] The size of the overlapping region may be, for example, the actual area of ​​the overlapping region, or may be the ratio of the overlapping region to the entire object region AR that is occluded (the degree of overlap). If the size of the overlapping region is equal to or greater than a predetermined threshold, the sensor analysis unit 13 determines that object A is occluded, and if the size of the overlapping region is less than the predetermined threshold, it determines that object A is not occluded.

[0034] In addition, the sensor analysis unit 13 uses the depth information contained in the sensor information to determine that object C is the occluding object on the occluding side (access point 2 side) and that object A is the occluded object on the occluded side (back side).

[0035] When the sensor analysis unit 13 determines that occlusion occurs, the identification unit 14 associates the communication terminal 4 in the communication information with the object on the occluded side of the overlapping area. For example, in the case of the communication information of FIG. 2A and the sensor information of FIG. 2B, when the sensor analysis unit 13 detects the occurrence of occlusion in the sensor information at time Tb identified by the quality analysis unit 12, the identification unit 14 determines that the occluded object A is carrying the communication terminal 4 in the communication information of FIG. 2A. Then, the identification unit 14 associates the communication terminal 4 with the object A and stores the association result in the storage unit 16.

[0036] 3 is a diagram showing an example of association between objects detected by sensor information and communication terminals 4, which is stored in storage unit 16. The example shown in the figure indicates that object A possesses a communication terminal 4 (referred to as "communication terminal 4A" in FIG. 3) with the communication quality shown in FIG. 2A.

[0037] 3, it is assumed that object B possesses communication terminal 4B, and object C does not possess communication terminal 4. As with communication terminal 4A, processing device 1 also estimates object B possessing communication terminal 4B based on the communication information of communication terminal 4B acquired from access point 2 and sensor information acquired from sensor 3, and associates object B with communication terminal 4B. If access point 2 is communicating with only two communication terminals 4A and 4B, that is, if processing device 1 can only acquire communication information of two communication terminals 4A and 4B, processing device 1 determines that object C shown in the sensor information does not possess a communication terminal 4.

[0038] The generation unit 15 may generate virtual sensor information based on the sensor information, with a viewpoint near the location of the access point 2. The virtual sensor information generated by the generation unit 15 will be described later. When the virtual sensor information is generated, the sensor analysis unit 13 uses the virtual sensor information to determine whether or not there is occlusion at the time specified by the quality analysis unit 12.

[0039] FIG. 4 is a flowchart showing the terminal identification process of the processing device 1.

[0040] The processing device 1 acquires communication information of each communication terminal 4 for a predetermined period from the access point 2. The processing device 1 also acquires sensor information for the predetermined period from the sensor 3 (S11).

[0041] The processing device 1 selects communication information of any one of the communication terminals 4 from the communication information acquired for each communication terminal 4 (S12). Because a sudden change in communication quality is not necessarily caused only by obstruction, in this embodiment, multiple candidates for events that are estimated to be obstruction within a certain period of time are extracted, and the presence or absence of obstruction is determined by analyzing the sensor information at the time of each candidate, and the communication terminal 4 is identified.

[0042] FIG. 5 is a diagram showing an example of communication information of the communication terminal 4 selected in S12 and sensor information at a certain point in time during a predetermined period.

[0043] The processing device 1 identifies, in the illustrated communication information 51, each time point at which the communication quality suddenly drops to a minimum as a shielding candidate (S13). In the illustrated example, the processing device 1 identifies, as a shielding candidate, a time point at which the communication quality suddenly drops and the minimum is below a threshold (0.4 or less in the illustrated example). In the illustrated communication information 51, the time points identified as shielding candidates are indicated by black circles.

[0044] The processing device 1 analyzes the sensor information (camera image) of the occlusion candidate at each time point and determines whether occlusion has occurred based on the sensor information at each time point (S14). Specifically, the processing device 1 detects an object area (a rectangular area surrounding an object) from the sensor information and determines whether overlap has occurred in the detected object area. If overlap has occurred, the processing device 1 then determines whether occlusion has occurred based on the size of the overlapping area.

[0045] In the illustrated example, in sensor information 61 at time Tb when communication quality suddenly deteriorates, the processing device 1 determines that object A is occluded and that object A is carrying the communication terminal 4 selected in S12. Note that in sensor information 62 at time Tc when communication quality suddenly deteriorates, the object regions AR, BR, and CR of objects A, B, and C do not overlap. Therefore, the sudden deterioration in communication quality at time Tc is thought to be due to a factor other than occlusion by objects A, B, and C.

[0046] If the processing device 1 determines in S14 whether or not there is occlusion using the sensor information of the occlusion candidate at each time point and determines that there is occlusion in the sensor information at least at one time point (S15: YES), it determines that the object on the occluded side is carrying the communication terminal 4 selected in S12, and associates the object on the occluded side with the communication terminal 4 selected in S12 (S16).If it determines that there is no occlusion in the sensor information at any time point (S15: NO), the processing device 1 determines that the object carrying the communication terminal 4 selected in S12 is not present in the camera image, and proceeds to S17.

[0047] When the processing device 1 has processed all of the communication information acquired from the access point 2 (S17: YES), the processing device 1 ends the processing. On the other hand, when there is unprocessed (unselected) communication information (S17: NO), the processing device 1 returns to S12, selects the unselected communication information, and repeats the subsequent processing.

[0048] 6 is an explanatory diagram illustrating the process of generating virtual sensor information. When the access point 2 and the sensor 3 are located at significantly different positions, the processing device 1 may generate virtual sensor information from a different viewpoint based on the sensor information acquired from the sensor 3, and use the virtual sensor information to determine whether or not there is occlusion.

[0049] In the illustrated example, the location where sensor 3 (camera) is installed is significantly different from the location where access point 2 is installed. In this case, when viewed from the access point 2 side, object A is occluded by object C, and occlusion occurs. On the other hand, in the sensor information when viewed from sensor 3, object A does not appear to be occluded. Therefore, the generation unit 15 of the processing device 1 may virtually convert the viewpoint of the sensor information, generate sensor information when viewed from the virtual viewpoint of virtual sensor 3A near access point 2, and determine whether or not occlusion occurs.

[0050] Since the positional relationship between the access point 2 and the sensor 3 is generally unknown, calculations are performed at multiple positions while changing the position of the virtual sensor 3A in preprocessing to estimate a likely position. This makes it possible to determine whether the access point 2 and the sensor 3 are occluded by an object using sensor information, even if the access point 2 and the sensor 3 are installed in significantly different positions.

[0051] The virtual sensor information generation process will be specifically described below, including prerequisites, a method for generating a virtual image, and pre-processing and terminal identification processing according to this embodiment.

[0052] <Prerequisites> As a prerequisite, it is assumed that the actually placed sensor 3 (actual sensor) is placed at the origin of the sensor coordinate system Os = [0, 0, 0]. In the following explanation, it is assumed that the sensor 3 is a camera (RGB-D camera) and the sensor information is a camera image.

[0053] In this case, the three-dimensional position of the object in the camera image is described as Ps = [Xs, Ys, Zs] on the sensor coordinate system Os. Since the three-dimensional position Ps of the object is estimated from the camera image, it is assumed that the three-dimensional position Ps is associated with information indicating that it was captured by a camera.

[0054] <Method of generating virtual image> A virtual image obtained by virtually transforming the viewpoint can be generated, for example, as follows: It is necessary to calculate onto which pixel of the virtual camera image of virtual camera 3A the three-dimensional position Ps = [Xs, Ys, Zs] of an object described in sensor coordinate system Os is projected when viewed from the virtual viewpoint (virtual camera 3A).

[0055] First, to express the position and orientation of the virtual camera 3A, a rotation matrix R (a matrix expressed by real numbers with a size of 3 × 3) and a translation vector T (a matrix expressed by real numbers with a size of 3 × 1) are prepared. The rotation matrix R has the function of converting the orientation, and the translation vector T has the function of converting the position.

[0056] This allows the three-dimensional position Ps of the object on the sensor coordinate system Os to be converted into a three-dimensional position Pv=R*Ps+T when viewed from the virtual viewpoint.

[0057] To generate a virtual camera image seen from a virtual viewpoint, elements include the camera's focal length, center coordinates, etc. Using a 3D to 2D transformation matrix A (internal parameters), the pixel coordinate q of the virtual camera onto which the 3D position Ps is projected is calculated as q = A * Pv = A(R * Ps + T), and information indicating that the pixel value of pixel coordinate q is linked to the 3D position Ps is recorded.

[0058] The three-dimensional to two-dimensional transformation matrix A uses the same values ​​as the transformation matrix A of camera 3, and these values ​​are acquired by performing calibration in advance. Calibration is described in, for example, Non-Patent Document 3. If there are multiple objects in the camera image of camera 3, the above processing is performed for each object.

[0059] <Preprocessing> The generation unit 15 of the processing device 1 performs preprocessing in the following procedure to estimate the likely position T and orientation R of the virtual camera 3A.

[0060] 1. The generation unit 15 obtains the internal parameters A (a three-dimensional to two-dimensional transformation matrix A) of the installed camera 3 by calibration.

[0061] 2. The generation unit 15 acquires communication information indicating the communication quality of communication terminal X held by object X for a predetermined period from the access point 2, and acquires camera images for the predetermined period from the camera 3. The generation unit 15 then calculates the three-dimensional positions Ps_i (i = 1, 2, ..., N) of multiple objects from the camera images of the camera 3, where N is the number of objects. It is assumed here that which of the multiple objects owns which communication terminal.

[0062] 3. The generation unit 15 generates a plurality of candidates for the translation vector T and rotation matrix R, which are parameters of the position and orientation of the virtual camera 3A. For each generated candidate, the generation unit 15 generates a virtual camera image using T and R of the candidate and the transformation matrix A obtained in 1 above.

[0063] 4. Using the communication information of communication terminal X possessed by object X, the generation unit 15 estimates the time when communication quality suddenly deteriorates as the time when occlusion occurred, similar to the quality analysis unit 12. Then, at the time when occlusion is estimated in each virtual camera video generated in 3 above, the generation unit 15 determines whether occlusion exists, similar to the sensor analysis unit 13. That is, the generation unit 15 determines whether object X possessing communication terminal X is also occluded in the virtual camera video at the time when occlusion is estimated to have occurred based on the communication information. Then, the generation unit 15 estimates T and R of the virtual camera video in which object X is occluded as the likely position and orientation of virtual camera 3A.

[0064] 5. The generation unit 15 also determines T and R of the final virtual camera image for other communication terminals Y held by other objects Y, taking into consideration the results of the processes 2 to 4 above. That is, the generation unit 15 estimates, at the time when it is estimated that occlusion occurred at each communication terminal, T and R of the virtual camera image in which the object holding that communication terminal is also occluded, as the likely position and orientation of virtual camera 3A.

[0065] <Terminal Identification Process> In S14 of the terminal identification process shown in Fig. 4, the generation unit 15 generates virtual camera information (virtual sensor information) based on camera images (sensor information). Specifically, the generation unit 15 calculates three-dimensional positions Ps_i (i = 1, 2, ..., N) of multiple objects from the camera images of the installed cameras 3 (sensors). Here, N is the number of objects.

[0066] Then, the generation unit 15 generates a virtual camera image (virtual sensor information) viewed from a virtual viewpoint using the parameters of the position T and orientation R of the virtual camera 3A (virtual sensor) estimated in the preprocessing, and the internal parameter A.

[0067] The sensor analysis unit 13 determines whether or not there is occlusion at the time specified by the quality analysis unit 12, using the virtual camera image generated by the generation unit 15 (S14). The determination of whether or not there is occlusion by the sensor analysis unit 13 is as described above.

[0068] The processing device 1 (terminal identification device) of this embodiment described above includes an acquisition unit 11 that acquires communication information indicating the communication quality of the communication terminal 4 over a predetermined period from an access point 2 (device) that communicates wirelessly with the communication terminal 4, and acquires sensor information for the predetermined period from a sensor 3; a quality analysis unit 12 that analyzes the communication information and identifies the point in time when the communication quality of the communication terminal 4 suddenly deteriorated; a sensor analysis unit 13 that, when the object areas of multiple objects detected from the sensor information at the time point overlap, determines whether or not there is occlusion based on the size of the overlapping area; and, when the sensor analysis unit 13 determines that there is occlusion, associates the communication terminal 4 with the object on the occluded side of the overlapping area.

[0069] The terminal identification method performed by the processing device 1 (terminal identification device) of this embodiment acquires communication information indicating the communication quality of the communication terminal 4 over a predetermined period of time from an access point 2 (device) that communicates wirelessly with the communication terminal 4, and acquires sensor information for the predetermined period of time from a sensor 3, analyzes the communication information, identifies the point in time when the communication quality of the communication terminal suddenly deteriorated, and if the object areas of multiple objects detected from the sensor information at that point in time overlap, determines whether or not there is occlusion based on the size of the overlapping area, and if it is determined that there is occlusion, associates the communication terminal with the object on the occluded side of the overlapping area.

[0070] According to this embodiment, the influence of obstructions is taken into consideration, and even in communications using high frequency bands, terminals can be identified with high accuracy.

[0071] Specifically, by associating objects captured in sensor information (camera footage) with communication terminal 4 and identifying communication terminal 4, rich features estimated from the sensor information can be combined with communication information to identify communication terminal 4, which can be applied to a variety of applications such as predicting wireless communication quality, recommendations based on understanding human behavior, and crime prevention.

[0072] This embodiment makes it possible to link a communication terminal with the analysis results of the sensor information by identifying the combination of the ID of the communication terminal 4 and the object in the sensor information and tracking and analyzing the target object from the sensor information. Furthermore, this embodiment can be used in applications such as predicting communication quality and delivering content based on the analysis results to a desired communication terminal 4.

[0073] In this embodiment, a generation unit 15 is provided that generates, based on the sensor information, virtual sensor information with a viewpoint near the location of the access point 2. This makes it possible to determine whether an object is occluded by using the sensor information, even if the access point 2 and the sensor 3 are installed at positions that are significantly different from each other.

[0074] The processing device 1 (terminal identification device) described above can be, for example, a general-purpose computer system as shown in Fig. 6. The computer system shown in the figure includes a CPU (Central Processing Unit, processor) 901, a memory 902, a storage 903 (HDD: Hard Disk Drive, SSD: Solid State Drive), a communication device 904, an input device 905, and an output device 906. The memory 902 and the storage 903 are storage devices. In this computer system, the CPU 901 executes a predetermined program loaded on the memory 902, thereby realizing each function of the processing device 1.

[0075] The processing device 1 may be implemented in one computer or in multiple computers. The processing device 1 may be a virtual machine implemented in a computer. The program of the processing device 1 may be stored in a computer-readable recording medium such as a HDD, SSD, USB (Universal Serial Bus) memory, CD (Compact Disc), or DVD (Digital Versatile Disc), or may be distributed via a network. The computer-readable recording medium may be, for example, a non-transitory recording medium.

[0076] The present disclosure is not limited to the above-described embodiments, and various modifications are possible within the scope of the present disclosure.

[0077] 1: Processing device (terminal identification device) 11: Acquisition unit 12: Quality analysis unit 13: Sensor analysis unit 14: Identification unit 15: Generation unit 16: Storage unit 17: Display unit 2: Access point (device) 3: Sensor 4: Communication terminal

Claims

1. A terminal identification device comprising: an acquisition unit that acquires communication information indicating the communication quality of a communication terminal over a predetermined period from a device that wirelessly communicates with the communication terminal, and acquires sensor information for the predetermined period from a sensor; a quality analysis unit that analyzes the communication information and identifies the point in time when the communication quality of the communication terminal suddenly deteriorates; a sensor analysis unit that, when the object areas of multiple objects detected from the sensor information at the point in time overlap, determines whether or not there is occlusion based on the size of the overlapping area; and, when the sensor analysis unit determines that there is occlusion, associates the communication terminal with the object on the occluded side of the overlapping area.

2. A terminal identification device as described in claim 1, further comprising a generation unit that generates virtual sensor information based on the sensor information, with a viewpoint near the location of the device, and the sensor analysis unit uses the virtual sensor information to determine whether or not the device is obstructed at the time point.

3. A terminal identification method performed by a terminal identification device, comprising: acquiring communication information indicating the communication quality of a communication terminal for a predetermined period from a device that wirelessly communicates with the communication terminal; acquiring sensor information for the predetermined period from a sensor; analyzing the communication information to identify a time point at which the communication quality of the communication terminal suddenly decreased; if the object areas of multiple objects detected from the sensor information at the time point overlap, determining whether or not there is occlusion based on the size of the overlapping area; and if it is determined that there is occlusion, associating the communication terminal with the object on the occluded side of the overlapping area.

4. A program that causes a computer to function as: an acquisition unit that acquires communication information indicating the communication quality of a communication terminal over a predetermined period from a device that communicates wirelessly with the communication terminal, and acquires sensor information for the predetermined period from a sensor; a quality analysis unit that analyzes the communication information and identifies the point in time when the communication quality of the communication terminal suddenly deteriorated; a sensor analysis unit that, when the object areas of multiple objects detected from the sensor information at the point in time overlap, determines whether or not there is occlusion based on the size of the overlapping area; and, when the sensor analysis unit determines that there is occlusion, an identification unit that associates the communication terminal with the object on the occluded side of the overlapping area.

Images