Judgment device, judgment system, judgment method, and program

Abstract

This determination device comprises: a signal information acquisition means that acquires signal information from an RFID tag provided to a commodity held by a customer when commodity identification information in the RFID tag is read by a reading device provided to a gate through which the customer passes; a location information acquisition means that acquires location information indicating the location of the customer by a motion sensor provided to the gate; and a determination means that uses a trained model subjected in advance to machine learning to determine whether or not the commodity has been purchased by the customer on the basis of a feature of the signal information and a feature of the location information.

Description

Determination device, determination system, determination method, and recording medium

[0001] The present disclosure relates to a determination device, a determination system, a determination method, and a recording medium.

[0002] There are devices that acquire product information for purchase by customers by reading products in a shopping cart with a determination device installed at a gate. Such devices are designed to improve the accuracy of reading the product information. For example, Patent Document 1 discloses a device that determines whether a product is eligible for purchase based on changes in received signal strength over time for each product.

[0003] International Publication No. 2019-065684

[0004] However, the determination method of Patent Document 1 may not be able to correctly determine whether or not an item is eligible for purchase.

[0005] An example of an object of the present disclosure is to provide a determination device that can improve the accuracy of determining whether a customer who has passed through a gate has purchased an item.

[0006] A determination device in one aspect of the present disclosure includes a signal information acquisition means that acquires signal information from an RFID tag when a reading device installed at a gate through which a customer passes reads product identification information in an RFID tag attached to a product held by the customer; a position information acquisition means that acquires position information indicating the customer's position using a human presence sensor installed at the gate; and a determination means that determines whether the product is purchased by the customer based on characteristics of the signal information and characteristics of the position information using a trained model that has been machine-learned in advance.

[0007] In one aspect of the present disclosure, a determination method involves a computer that, when reading product identification information in an RFID tag attached to a product held by a customer using a reading device installed at a gate through which the customer passes, acquires signal information from the RFID tag, acquires location information indicating the customer's location using a human presence sensor installed at the gate, and determines whether the product is purchased by the customer based on the characteristics of the signal information and the characteristics of the location information using a trained model that has been machine-learned in advance.

[0008] In one aspect of the present disclosure, a recording medium stores a program that causes a computer to execute a process in which, when a reading device installed at a gate through which a customer passes reads product identification information in an RFID tag attached to a product carried by the customer, signal information is acquired from the RFID tag, location information indicating the customer's location is acquired by a human presence sensor installed at the gate, and whether the product is purchased by the customer is determined based on characteristics of the signal information and characteristics of the location information using a trained model that has been machine-learned in advance.

[0009] According to the present disclosure, it is possible to provide a determination device that can improve the accuracy of determining whether or not a product is purchased by a customer who has passed through a gate.

[0010] FIG. 1 is a block diagram showing an example configuration of a determination system according to the present disclosure. FIG. 2 is a schematic diagram showing an example of a gate according to the present disclosure. FIG. 3 is a diagram showing a hardware configuration in which a determination device according to the present disclosure is realized using a computer device and its peripheral devices. FIG. 4 is a top view of a gate according to the present disclosure as seen from the ceiling side. FIG. 5 is a graph showing the relationship between location information and changes in RSSI values ​​when a customer passes through a gate according to the present disclosure. FIG. 6 is a flowchart showing a determination operation according to the present disclosure. FIG. 7 is a block diagram showing an example configuration of a determination system according to the present disclosure. FIG. 8 is a diagram for explaining relearning of a trained model for each reading condition of a reading device according to the present disclosure. FIG. 9 is a diagram for explaining relearning of a trained model for each reading condition of a reading device according to the present disclosure. FIG. 10 is a top view of a gate according to the present disclosure as seen from the ceiling side. FIG. 11 is a schematic diagram showing an example of a gate according to the present disclosure.

[0011] Hereinafter, with reference to the drawings, embodiments of a determination device, a determination system, a determination method, and a non-transitory recording medium for recording a program according to the present disclosure will be described in detail. The present embodiments do not limit the disclosed technology.

[0012] 1 is a block diagram showing an example of a configuration including a determination device 100 according to the present disclosure. As shown in Fig. 1, in the determination system 10, the determination device 100 is connected to a reading device 200, a human presence sensor 300, and an output device 400 via a network.

[0013] The determination device 100 is a device that determines whether or not a customer will purchase a product based on product identification information in an RFID (Radio Frequency Identification) tag attached to the product read by the reading device 200. The reading device 200 is a sensor for reading product identification information of a product to be purchased by a customer passing through a gate. The human presence sensor 300 is a sensor for detecting a customer. The output device 400 is a device that outputs information on a product determined to be purchased by the customer, and is, for example, a display installed near the exit of the gate or a display on the customer's terminal. The store in the present disclosure is, for example, a retail store that sells clothing, food, etc. The product identification information is not particularly limited as long as it is information that can identify a product, and is, for example, a product name or a product ID.

[0014] (Gate) FIG. 2 is a schematic diagram illustrating an example of a gate in the present disclosure. Gate 1 as shown in FIG. 2 is provided, for example, near the exit of a store. Gate 1 is provided with a reading device 200 and a motion sensor 300. Reading device 200 has an RFID reader 201 and an antenna 202. The gate shown in FIG. 2 is a walk-through gate in which, as a customer walks through aisle A within gate 1, reading device 200 reads product identification information stored in an RFID tag attached to a product P carried by the customer. Note that, in the example gate shown in FIG. 2, two antennas 202 are provided on each side of aisle A near the center of the gate in the direction of travel, but the number of antennas 202 is not limited to the example configuration shown in FIG. 2.

[0015] The RFID reader 201 of the reading device 200 is a control circuit that communicates with an RFID tag via an antenna 202 in accordance with a predetermined communication protocol and reads the product identification information stored in the RFID tag of the product. The RFID reader 201 outputs the product identification information read from the RFID tag to the determination device 100. The antenna 202 is installed in a position where it can transmit and receive radio waves to and from the RFID tag passing through the gate 1, and transmits radio waves to the RFID tag and receives radio waves transmitted from the RFID tag. The antenna 202 may be provided as a separate transmitting antenna and a separate receiving antenna.

[0016] The human presence sensor 300 is a sensor that detects people inside the gate 1 without contact and is composed of, for example, an infrared sensor or a camera. The human presence sensor may be configured to include a vector sensor that can detect the movement vector of a customer and may also be configured to detect the direction or amount of movement of the customer. If the human presence sensor 300 includes a vector sensor, it can detect whether the customer is moving along the movement direction of the gate 1. Furthermore, if the area including the gate 1 is divided into multiple areas, the human presence sensor 300 may be capable of detecting the area through which the customer passes when signal information is acquired. In the configuration example shown in FIG. 2 , the human presence sensor is provided on the top surface of the gate 1, but the location of the human presence sensor is not limited thereto. For example, the human presence sensor may be provided on the ceiling of the location above the gate 1.

[0017] In the example of gate 1 in Fig. 2, output device 400 may be provided near the exit of gate 1 and configured to output a list of products determined to be purchased. Also, when it is detected that a customer has exited gate 1, a payment process for the products determined to be purchased may be performed using pre-registered payment information.

[0018] (Trained Model) The determination device 100 stores a trained model for determining whether or not a product is purchased by a customer in the storage device 505. This trained model is a model generated by machine learning of the correlation between the characteristics of signal information from the RFID tag and the characteristics of location information from the human presence sensor 300 when the reading device 200 reads the product identification information, and whether or not the product is purchased, and is expressed, for example, by a mathematical formula using the signal information and location information as variables.

[0019] Examples of the signal information include a received signal strength indicator (RSSI) value indicating the signal strength of the read product identification information, the number of times the product identification information has been read, the reading time, the number of antennas 202 that have read the electronic product code (EPC) stored in the RFID tag, and attribute information of the product to which the RFID tag is attached. The trained model may be a model expressed as a mathematical formula using multiple pieces of signal information selected from the above-mentioned signal information. The determination device 100 uses this trained model to estimate whether the read product identification information is a purchased product based on the signal information from the reading device 200 and the position information from the human presence sensor 300.

[0020] The trained model is generated by a learning means of the determination device 100 or a device other than the determination device 100. The learning algorithm may be any machine learning method such as a neural network, a support vector machine (SVM), or logistic regression.

[0021] Returning to FIG. 1 , the determination device 100 includes a signal information acquisition unit 101, a position information acquisition unit 102, and a determination unit 103. FIG. 3 is a diagram illustrating an example of a hardware configuration in which the determination device 100 according to the present disclosure is realized by a computer device 500 including a processor. As illustrated in FIG. 3 , the determination device 100 includes a central processing unit (CPU) 501, memories such as a read-only memory (ROM) 502 and a random access memory (RAM) 503, a storage device 505 such as a hard disk for storing a program 504, a communication interface 508 for network connection, and an input / output interface 509 for inputting and outputting data. The determination device 100 is also connected to each component via a bus 510. The determination device 100 illustrated in FIG. 1 can also be configured using cloud computing or the like.

[0022] The CPU 501 runs an operating system to control the entire determination device 100 according to the present disclosure. The CPU 501 also reads programs and data into memory from a recording medium 506 attached to, for example, a drive device 507. The CPU 501 also functions as the signal information acquisition unit 101, the position information acquisition unit 102, and the determination unit 103 according to the present disclosure, or as part of these, and executes processing or commands in the flowchart shown in FIG. 6, which will be described later, based on the program.

[0023] The recording medium 506 is, for example, an optical disk, a flexible disk, a magneto-optical disk, an external hard disk, or a semiconductor memory. The semiconductor memory or the like that is part of the recording medium is a non-volatile storage device that stores the program. The program may also be downloaded from an external computer (not shown) that is connected to a communication network.

[0024] As described above, the first embodiment shown in Fig. 1 is realized by the computer hardware shown in Fig. 3. However, the means for realizing each unit included in the determination device 100 in Fig. 1 is not limited to the configuration described above. The determination device 100 may be realized by a single physically coupled device, or may be realized by a system consisting of two or more physically separated devices connected by wire or wirelessly.

[0025] The signal information acquisition unit 101 is a means for acquiring signal information from an RFID tag when the reading device 200 installed at the gate 1 through which a customer passes reads commodity identification information stored in an RFID tag attached to a commodity. The signal information acquisition unit 101 measures an RSSI value from the RFID tag received by the antenna 202 to acquire signal strength. While the present specification describes a case in which the signal information is an RSSI value, the signal information is not limited to an RSSI value and may also include a maximum RSSI value, a minimum RSSI value, etc. The signal information may also include the number of times the reading device 200 reads each unit of commodity identification information, the reading time, the number of antennas 202 that have read the EPC stored in the RFID tag, attribute information of the commodity to which the RFID tag is attached, etc. If there are multiple commodities, the signal information acquisition unit 101 acquires signal information from the RFID tag of each commodity. Upon acquiring signal information from the reading device 200, the signal information acquisition unit 101 outputs information indicating that the signal information has been acquired to the location information acquisition unit 102.

[0026] The location information acquisition unit 102 is a means for acquiring location information indicating the location of a customer by the human presence sensor 300 provided at the gate 1. When the location information acquisition unit 102 receives information indicating that signal information has been acquired from the reading device 200 by the signal information acquisition unit 101, the location information acquisition unit 102 acquires location information indicating the location of the customer by the human presence sensor 300 provided at the gate 1. For example, the location information acquisition unit 102 acquires information on whether or not the customer is within the gate 1 by the human presence sensor 300. If the area within the gate 1 where the customer is located or the direction of movement of the customer can be determined by the human presence sensor 300, the location information acquisition unit 102 may acquire such information.

[0027] The determination unit 103 is a means for determining whether or not a product is purchased by a customer based on the characteristics of the traffic light information and the characteristics of the location information, using a trained model that has been machine-learned in advance. The determination unit 103 inputs the traffic light information and the location information from the RFID tag of each product into the trained model stored in the storage device 505, and determines whether or not each product identification information that has been read is a product purchased by a customer passing through the gate 1.

[0028] The reading process by the reading device 200 in the present disclosure has the following characteristics: In other words, in the reading process by the reading device 200, since the reading device 200 is provided at the gate 1 that guides the movement route of customers, when reading the RFID tag of a product carried by a certain customer, there is a possibility that the reading device 200 may read an RFID tag carried by another customer outside the gate 1 or an RFID tag of a product displayed on a shelf near the gate 1. In other words, there is a possibility that the reading device 200 may read an RFID tag that should not be read.

[0029] Therefore, in the determination device 100 of the present disclosure, the determination unit 103 uses a trained model to determine whether the product is a purchased product from the customer based on the characteristics of the signal information and the characteristics of the location information. Specifically, when the signal information is an RSSI value, the determination unit 103 uses the trained model to determine whether the product is a purchased product from the customer based on a combination of the magnitude of the RSSI value and the customer's location. That is, when the magnitude of the RSSI value from the antenna 202 is equal to or greater than a predetermined value and the customer is inside gate 1, the determination unit 103 determines that the read product identification information is a purchased product. On the other hand, when the magnitude of the RSSI value from the antenna 202 is equal to or greater than the predetermined value but the customer is outside gate 1, the determination unit 103 determines that the product identified by the read product identification information is not a purchased product. Note that when the magnitude of the RSSI value from the antenna 202 is less than the predetermined value, the determination unit 103 determines that the product identified by the read product identification information is not a purchased product, regardless of whether the customer is inside or outside gate 1.

[0030] FIG. 4 is a top view of gate 1 in the present disclosure as seen from the ceiling side. Here, the determination method in the present disclosure will be described using FIG. 4. As shown in FIG. 4, in this determination method, an area including gate 1 is divided into multiple areas, and the area through which a customer is passing can be detected based on position information acquired from the human presence sensor 300. In the example of FIG. 4, position information indicating customer positions from near the entrance to near the exit of gate 1 is designated as positions 1 to 5. Customer positions are not limited to the example of FIG. 4, and may be divided into any position and any number of areas.

[0031] FIG. 5 is a graph showing the relationship between location information and changes in RSSI values ​​when a customer passes through gate 1 in the present disclosure. In FIG. 5, tag A represents the time-series change in the RSSI value from an RFID tag attached to a product carried by the customer as the customer moves through gate 1 from positions 1 to 3, and tag B represents the time-series change in the RSSI value from an RFID tag outside gate 1 as the customer moves through gate 1 from positions 1 to 3. When the RFID tag moves through gate 1, as shown by tag A in FIG. 5, the RSSI value exhibits a time-series change that exceeds a predetermined value. Note that while the graph in FIG. 5 shows the time-series change in the RSSI value as signal information, the signal information is not limited to the RSSI value as long as it is information that affects the reading accuracy of the product identification information stored in the RFID tag.

[0032] 6 is a flowchart showing an outline of the operation of the determination device 100 according to the present disclosure. The processing according to this flowchart may be executed based on program control by the processor described above. The determination device 100 may start the flow according to this flowchart, for example, when the signal information received from the reading device 200 is equal to or greater than a predetermined value.

[0033] 6 , first, the traffic light information acquisition unit 101 acquires traffic light information from the RFID tag when the reader 200 installed at the gate 1 through which the customer passes reads the product identification information in the RFID tag attached to the product carried by the customer (step S101). Next, the location information acquisition unit 102 acquires location information indicating the customer's location using the human presence sensor 300 installed at the gate 1 (step S102). Next, the determination unit 103 determines whether the product is purchased by the customer based on the characteristics of the traffic light information and the characteristics of the location information using a trained model that has been machine-learned in advance (step S103). The determination device 100 then ends the flow.

[0034] In the determination device 100 of this embodiment, the determination unit 103 uses a trained model that has been machine-learned in advance to determine whether or not a product has been purchased by a customer based on the characteristics of the traffic light information and the characteristics of the location information. As a result, even if the reading device 200 reads product identification information in an RFID tag outside the gate 1, for example, the product identification information of a product held by a customer inside the gate 1 can be determined based on the customer's location information. This improves the accuracy of determining whether or not a product has been purchased by a customer who has passed through the gate 1.

[0035] [First Modification of the First Embodiment] Next, a modification of the present disclosure will be described in detail with reference to the drawings. Fig. 7 is a block diagram showing a determination system 11 according to the present disclosure. In the determination system 11, similar to the determination system 10, a determination device 110 is connected to a reading device 210, a human presence sensor 310, and an output device 410 via a network.

[0036] In the following, the description of the present embodiment will be omitted to the extent that the description thereof does not become unclear. The functions of each component in each embodiment of the present disclosure can be realized not only by hardware but also by a computer device or software under program control, similar to the computer device shown in FIG.

[0037] 7, the determination device 110 includes a signal information acquisition unit 111, a position information acquisition unit 112, a learning unit 113, a determination unit 114, and an output unit 115. The configurations of the signal information acquisition unit 111 and the position information acquisition unit 112 are similar to the corresponding configurations in the first embodiment, and therefore description thereof will be omitted.

[0038] The signal information received from the RFID tag when reading the product identification information stored in the RFID tag varies depending on the reading conditions of the reader 210. Using FIGS. 8 and 9 , we will explain the re-learning of the trained model for each reading condition of the reader 210. FIG. 8 shows a graph of the transition of RSSI values ​​when a customer passes through a gate when a trained model re-trained under the reading conditions of the reader 210 is used. FIG. 9 shows a graph of the transition of RSSI values ​​when a customer passes through a gate when a trained model similar to FIG. 8 is used under different reading conditions. As shown in FIG. 8 , when a trained model re-trained under the reading conditions of the reader 210 is used, it is easy to distinguish between the graphs of RSSI values ​​for tag A inside the gate and tag B outside the gate. In contrast, when a trained model not re-trained under the reading conditions of the reader 210 is used, as shown in FIG. 9 , it is difficult to distinguish between the graphs of RSSI values ​​for tag A inside the gate and tag B outside the gate. Similarly, even in the case of a trained model that uses signal information other than RSSI values, by optimizing the trained model by re-training it for each reading condition of the reading device 210, it becomes easier to distinguish between signal information when an RFID tag inside the gate is read and signal information when an RFID tag outside the gate is read.

[0039] That is, the trained model in this modification is a model that has been further retrained to learn the correlation between the characteristics of the traffic light information and the characteristics of the location information, and whether or not the product is for purchase, under the reading conditions of the reading device 210. More specifically, the trained model is a model that has been retrained to learn the correlation by inputting multiple sets of training datasets that include the traffic light information and location information, and the results of whether or not the product is for purchase, obtained under the reading conditions of the reading device 210.

[0040] The reading conditions include, for example, the store environment where the reader 210 is installed, the weather, and the radio wave environment. The store environment includes, for example, the store location (e.g., a street store or a store in a mall), the store ceiling height, the material of the store floor or ceiling, the layout of the product shelves, or how crowded the store is. The radio wave environment includes, for example, the position of the antenna inside the gate, the number of antennas, the material of the shopping cart, or the material of the product to which the RFID tag is attached.

[0041] The learning unit 113 re-learns the trained model based on the reading conditions of the reading device 210. Then, the determination unit 114 uses the re-trained trained model to determine whether or not the product is a purchase product for the customer based on the characteristics of the traffic light information and the characteristics of the location information.

[0042] The output unit 115 is a means for outputting a screen listing the products determined to be purchased by the customer. The output unit 115 outputs the list of purchased products to, for example, the output device 410 or the customer's mobile terminal.

[0043] In this modified example, the trained model used by the determination unit 114 to determine whether or not an item is a purchased item is a trained model in which correlations have been further re-trained under the reading conditions of the reading device 210. In this case, as shown by comparing Figures 8 and 9, it becomes easier to distinguish between the product identification information of an RFID tag inside the gate 2 and the product identification information of an RFID tag outside the gate 2.

[0044] [Variation 2 of First Embodiment] Next, another variation of the present disclosure will be described in detail with reference to the drawings. In this variation, the antenna provided at the gate is different from that in the first embodiment. FIGS. 10 and 11 are diagrams for explaining the installation position of the antenna 212 in this variation. FIG. 10 is a top view of the gate 2 in this disclosure as seen from the ceiling side. FIG. 11 is a schematic diagram showing an example of the gate 2 in this disclosure. As shown in FIGS. 10 and 11 , in this variation, four antennas 212 are provided on each side of passage A. Furthermore, in this variation, the customer's location information acquired by the location information acquisition unit 112 is either near the entrance, near the center, or near the exit, as shown in FIG. 10 .

[0045] The trained model of this modification is a model generated using customer location information and signal information from multiple antennas located at different positions as features. The multiple antennas located at different positions are, for example, multiple antennas located at different positions along the direction of travel of gate 2. In the example configuration of antenna 212 shown in FIGS. 10 and 11 , antenna 212a is located near the gate entrance, antenna 212b is located near the center of the gate, and antenna 212c is located near the gate exit. Two antennas 212b located near the center of the gate are located on each side of passage A. Note that the locations of the antennas are not limited to the examples shown in FIGS. 10 and 11 . For example, antenna 212a near the gate entrance or antenna 212c near the gate exit do not need to be located at gate 2 itself, as long as they are located in positions that allow detection of a customer near the gate entrance or gate exit, respectively.

[0046] Here, we will explain how the characteristic quantities of the signal information from each antenna 212 change as the customer moves within the gate 2. For example, if the signal information is an RSSI value, as the customer approaches one of the antennas 212 installed within the gate 2, the RSSI value received from that antenna 212 increases. Specifically, as the customer approaches a position near the gate entrance, near the center, or near the exit, the RSSI values ​​received from each of the antennas 212a, 212b, and 212c increase. For example, when the customer is near the entrance, the RSSI value from the antenna 212a near the gate entrance in FIG. 10 is the largest. Next, as the customer approaches near the center of the gate, the RSSI value from the antenna 212b near the gate center in FIG. 10 is the largest, and the RSSI value from the antenna 212a near the gate entrance is the smallest. As the customer approaches the exit, the RSSI value from antenna 212c near the gate exit in FIG. 10 becomes the largest, and the RSSI value from antenna 212b near the center of the gate becomes small.

[0047] Therefore, using this trained model, the determination unit 114 determines whether the read product identification information is a purchased product based on whether the customer's position corresponds to the magnitude of the RSSI values ​​from multiple antennas arranged at different positions along the travel direction of gate 2. That is, if the customer is located near the center and the magnitude of the RSSI value from antenna 212a near the gate entrance or antenna 212c near the exit is equal to or greater than a predetermined value, the determination unit 114 determines that the read product identification information has been read from an RFID tag outside gate 2. Similarly, if the customer is located near the entrance or exit and the magnitude of the RSSI value from antenna 212b near the gate center is equal to or greater than a predetermined value, the determination unit 114 determines that the read product identification information has been read from an RFID tag located outside the aisle near the gate center.

[0048] The trained model may also be a model that uses the customer's location information and the difference in signal information from multiple antennas located at different positions as features. An example will be described below, taking the change in the feature values ​​of the RSSI values ​​from each antenna 212 as the customer moves within the gate 2. This model utilizes the tendency for the RSSI value from the antenna closest to the customer's location to be largest, while the RSSI values ​​from other antennas to be small. That is, when a customer is near the gate entrance, the difference between the RSSI value from the antenna 212a near the gate entrance and the RSSI value from the antenna 212b near the center of the gate or the antenna 212c near the gate exit is greater than or equal to a predetermined value. Similarly, when a customer is near the center of the gate, the difference between the RSSI value from the antenna 212b near the gate center and the RSSI value from the antenna 212a near the gate entrance or the antenna 212c near the gate exit is greater than or equal to a predetermined value. When a customer is near the gate exit, the difference between the RSSI value from the antenna 212c near the gate exit and the RSSI value from the antenna 212a near the gate entrance or the antenna 212b near the center of the gate is equal to or greater than a predetermined value.

[0049] Using this trained model, for example, when a customer is near the center of the gate, if the difference between the RSSI value from antenna 212b near the center of the gate and the RSSI value from antenna 212a near the gate entrance or antenna 212c near the gate exit is equal to or greater than a predetermined value, the determination unit 114 determines that the customer is reading commodity identification information from an RFID tag inside gate 2. On the other hand, when a customer is near the center of the gate, if the difference between the RSSI value from antenna 212b near the center of the gate and the RSSI value from antenna 212a near the gate entrance or antenna 212c near the gate exit is less than a predetermined value, the determination unit 114 determines that the customer is reading commodity identification information from an RFID tag outside gate 2.

[0050] In this modification, the trained model used by the determination unit 114 to determine whether or not an item is purchased may be a model generated using the customer's location information and signal information from multiple antennas located at different locations as features. Also, in this modification, the trained model may be a model using the customer's location information and the difference in signal information from multiple antennas located at different locations as features. In these cases, even if the product identification information of an RFID tag outside gate 2 is read when the customer is inside gate 2, it may be possible to distinguish between the product identification information of the RFID tag inside gate 2 and the product identification information of the RFID tag outside gate 2.

[0051] Although the present disclosure has been described above with reference to various embodiments, the present disclosure is not limited to the above embodiments. The configuration and details of each of the present disclosures may include embodiments to which various modifications that would be apparent to those skilled in the art are applied within the scope of the present disclosure. The present disclosure may also include embodiments in which the details described herein are appropriately combined or substituted as necessary. For example, details described using a particular embodiment may also be applied to other embodiments to the extent that no contradiction occurs. For example, although multiple operations are described in sequence in the form of a flowchart, the order of the descriptions does not limit the order in which the multiple operations are performed. Therefore, when implementing each embodiment, the order of the multiple operations may be changed as long as it does not interfere with the content.

[0052] Some or all of the above-described embodiments can be described as follows: However, some or all of the above-described embodiments are not limited to the following.

[0053] (Supplementary Note 1) A determination device comprising: a signal information acquisition means for acquiring signal information from an RFID tag attached to a product held by a customer when the product identification information in the RFID tag is read by a reading device installed at a gate through which the customer passes; a position information acquisition means for acquiring position information indicating the location of the customer by a human presence sensor installed at the gate; and a determination means for determining whether the product is purchased by the customer based on the characteristics of the signal information and the characteristics of the position information using a trained model that has been machine-learned in advance.

[0054] (Supplementary Note 2) The determination device according to Supplementary Note 1, wherein the trained model is a model generated by machine learning a correlation between a feature of the signal information and a feature of the location information and whether the product is a purchased product.

[0055] (Supplementary Note 3) The determination device according to Supplementary Note 2, wherein the trained model is a model obtained by further re-learning the correlation under the reading conditions of the reading device.

[0056] (Supplementary Note 4) The determination device according to Supplementary Note 3, wherein the trained model is a model that re-learns the correlation by inputting a plurality of sets of training datasets including the signal information and the location information obtained under the reading conditions of the reading device, and a result of whether the product is a purchased product.

[0057] (Supplementary Note 5) The determination device according to Supplementary Note 1 to 4, wherein the reading conditions of the reading device include at least one of a store environment in which the reading device is installed, weather, and a radio wave environment of the reading device.

[0058] (Supplementary Note 6) The determination device according to Supplementary Notes 1 to 5, wherein the human presence sensor includes a vector sensor and detects a direction in which the customer is moving.

[0059] (Supplementary Note 7) The determination device according to Supplementary Note 6, wherein the human presence sensor detects the area through which the customer passes when the traffic light information is acquired, when an area including the gate is divided into a plurality of areas.

[0060] (Supplementary Note 8) The determination device according to any one of Supplementary Notes 1 to 7, wherein the antennas of the reading devices are arranged near the entrance and exit of the gate and near the center of the gate.

[0061] (Supplementary Note 9) The determination device according to any one of Supplementary Notes 1 to 8, wherein the trained model is a model generated using location information of the customer and the signal information from a plurality of antennas arranged at different positions as features.

[0062] (Supplementary Note 10) The determination device according to Supplementary Note 9, wherein the trained model is a model generated using location information of the customer and a difference between signal information from an antenna closest to the customer's location and signal information from another antenna as features.

[0063] (Supplementary Note 11) The determination device according to any one of Supplementary Note 2 to Supplementary Note 4, further comprising a learning unit that re-learns the correlation in the reading conditions of the reading device.

[0064] (Supplementary Note 12) The determination device according to any one of Supplementary Notes 1 to 11, further comprising an output unit that outputs a screen showing a list of products determined to be purchased by the customer.

[0065] (Supplementary Note 13) A determination system comprising: a reading device provided at a gate through which a customer passes; a human presence sensor provided at the gate; and a determination device according to any one of Supplementary Notes 1 to 12.

[0066] (Supplementary Note 14) A determination method in which a computer acquires signal information from an RFID tag when reading product identification information in an RFID tag attached to a product held by a customer using a reading device installed at a gate through which the customer passes, acquires location information indicating the customer's location using a human presence sensor installed at the gate, and determines whether the product is purchased by the customer based on characteristics of the signal information and characteristics of the location information using a trained model that has been machine-learned in advance.

[0067] (Supplementary Note 15) The determination method according to Supplementary Note 14, wherein the trained model is a model generated by machine learning a correlation between the features of the signal information and the features of the location information and whether the product is a purchased product.

[0068] (Supplementary Note 16) The determination method according to Supplementary Note 15, further comprising relearning the correlation under the reading conditions of the reading device.

[0069] (Supplementary Note 17) The determination method described in Supplementary Note 16, wherein in the relearning, the correlation is relearned by inputting multiple sets of learning datasets including the signal information, the location information, and the result of whether the product is a purchased product under the reading conditions of the reading device.

[0070] (Supplementary Note 18) The determination method according to Supplementary Note 16 or Supplementary Note 17, wherein the reading conditions of the reading device include at least one of the store environment in which the reading device is installed, the weather, and the radio wave environment of the reading device.

[0071] (Supplementary Note 19) The determination method according to any one of Supplementary Notes 14 to 18, wherein a screen showing a list of products determined to be purchased by the customer is output.

[0072] (Supplementary Note 20) A recording medium storing a program that causes a computer to execute a process of acquiring signal information from an RFID tag when reading product identification information in an RFID tag attached to a product carried by the customer using a reading device installed at a gate through which the customer passes, acquiring location information indicating the customer's location using a human presence sensor installed at the gate, and determining whether the product is a purchase of the customer based on characteristics of the signal information and characteristics of the location information using a trained model that has been machine-learned in advance.

[0073] Some or all of the configurations described in Supplements 2 to 12, which are dependent on Supplement 1 described above, may also be dependent on Supplement 14 and Supplement 20 in the same dependent relationship as Supplements 2 to 12. Not limited to Supplements 1, 14, and 20, some or all of the configurations described as Supplements may also be dependent on various hardware, software, various recording devices for recording software, or systems, within the scope of each of the above-mentioned embodiments.

[0074] 10, 11 Determination system 100, 110 Determination device 101, 111 Signal information acquisition unit 102, 112 Position information acquisition unit 103, 114 Determination unit 113 Learning unit 115 Output unit 200, 210 Reading device 300, 310 Human presence sensor 400, 410 Output device 500 Computer device 501 CPU 502 ROM 503 RAM 504 Program 505 Storage device 506 Recording medium 507 Drive device 508 Communication interface 509 Input / output interface 510 Bus

Claims

1. A signal information acquisition means that acquires signal information from an RFID tag when a reader device installed at a gate through which a customer passes reads product identification information from an RFID tag attached to a product held by the customer, A location information acquisition means that acquires location information indicating the customer's location using a human presence sensor installed at the gate, A determination device comprising: determination means for determining whether or not a product is a customer's purchased product based on the characteristics of the signal information and the characteristics of the location information, using a pre-trained model that has been machine-learned in advance.

2. The determination device according to claim 1, wherein the trained model is a model generated by machine learning the correlation between the characteristics of the signal information and the characteristics of the location information and whether or not it is a purchased product.

3. The determination device according to claim 2, wherein the trained model is a model that has been further retrained on the correlation relationship under the reading conditions of the reading device.

4. The determination device according to claim 3, wherein the trained model is a model that has been retrained on the correlation relationship by inputting multiple sets of training datasets, each containing the signal information and position information obtained under the reading conditions of the reading device, and the result of whether or not the product is purchased.

5. The determination device according to claim 1 or 2, wherein the reading conditions of the reading device include at least one of the store environment in which the reading device is installed, the weather, and the radio wave environment of the reading device.

6. The determination device according to claim 1 or 2, wherein the motion sensor includes a vector sensor and detects the direction in which the customer is moving.

7. The determination device according to claim 6, wherein the motion sensor detects the area through which the customer passes when acquiring the signal information, when the area including the gate is divided into multiple areas.

8. A determination system comprising a reading device installed at a gate through which customers pass, a human presence sensor installed at the gate, and a determination device as described in claim 1 or 2.

9. Computers When a reader installed at a gate through which a customer passes reads the product identification information in the RFID tag attached to the product the customer is carrying, the reader acquires signal information from the RFID tag. A motion sensor installed at the gate acquires location information indicating the customer's location. A determination method that uses a pre-trained machine learning model to determine whether or not a product is purchased by the customer, based on the characteristics of the signal information and the characteristics of the location information.

10. When a reader installed at a gate through which a customer passes reads the product identification information in the RFID tag attached to the product the customer is carrying, the reader acquires signal information from the RFID tag. A motion sensor installed at the gate acquires location information indicating the customer's location. A program that causes a computer to perform a process to determine whether or not a product is purchased by the customer, based on the characteristics of the signal information and the characteristics of the location information, using a pre-trained model that has been machine-learned in advance.