Article positioning method, system, electronic device, and readable storage medium

Abstract

The present disclosure relates to the field of radio frequency identification technology, and particularly relates to an article positioning method and system, an electronic device and a readable storage medium. The article positioning method comprises: receiving a plurality of signals emitted by a plurality of articles arranged in a preset manner; extracting a specified feature from the plurality of signals; inputting the specified feature into a trained machine learning model to determine the positions of the articles for which the signals are received and / or not received. The technical solution of the present disclosure can accurately identify the positions of the articles for which the signals are received and / or not received on the premise that the arrangement manner of the articles is known, helping the staff to accurately and quickly locate the articles that are not identified. In the scenario of applying RFID to realize warehouse in-out identification, the technical solution of the present disclosure can accurately and quickly locate the positions of the articles with the missed RFID tags, improving the accuracy and efficiency of the in-out.

Description

Item location methods, systems, electronic devices and readable storage media Technical Field

[0001] This disclosure relates to the field of radio frequency identification technology, specifically to an item positioning method, system, electronic device, and readable storage medium. Background Technology

[0002] RFID (Radio Frequency Identification) is one of the most important wireless communication technologies in fields such as the Internet of Things (IoT). It can uniquely identify targets using radio frequency signals and transmit radio frequency information. UHF RFID technology has many advantages, including unique IDs, non-line-of-sight capability, low cost, and no battery required, and is widely used in smart warehousing, inventory management, and logistics tracking. RSSI (Received Signal Strength Indicator) information and the distance between transmitting and receiving devices have a definite correspondence in an ideal environment. Currently, indoor positioning technology based on RSSI information is widely used in indoor environments such as warehouses.

[0003] Existing smart warehouse inbound / outbound identification solutions often employ vision-based detection devices or RFID-based access gates. Vision-based devices typically only allow for single-item detection and identification, and are limited by line-of-sight distance, obstructions, and lighting conditions, making batch reading difficult and impacting inbound / outbound management efficiency. RFID-based access gates, on the other hand, can read items in batches, unaffected by obstructions or lighting conditions, significantly improving management efficiency. However, due to the electromagnetic shielding effect of metal, when densely packed items composed of numerous metal components pass through the access gate, the RFID tags attached to the goods may be missed. Summary of the Invention

[0004] To address the problems in the related technologies, embodiments of this disclosure provide an item positioning method, system, electronic device, and readable storage medium.

[0005] In a first aspect, this disclosure provides a method for locating an item, including:

[0006] Receives multiple signals emitted by multiple items arranged in a preset manner;

[0007] Extract specified features from the plurality of signals;

[0008] The specified features are input into a trained machine learning model to determine the locations of items among the plurality of items where signals are received and / or not received.

[0009] According to embodiments of this disclosure, the plurality of articles have the same shape;

[0010] The preset placement method is a neat arrangement;

[0011] The signal is transmitted by a signal transmitting device positioned at the same location as the item;

[0012] The signal is an RFID signal.

[0013] According to an embodiment of this disclosure, receiving multiple signals emitted by multiple items arranged in a preset manner includes: receiving the multiple signals through multiple antennas disposed on both sides of a channel for transporting the multiple items;

[0014] The step of extracting a specified feature from the plurality of signals includes extracting the specified feature based on the signal strength indication (RSSI) of the signals received by the plurality of antennas.

[0015] According to embodiments of this disclosure, the specified features include any one or more of the following: the ratio of the number of times each antenna in the plurality of antennas receives a signal to the total number of times the plurality of antennas receive a signal; the average RSSI value of the signals received by each antenna in the plurality of antennas; the maximum RSSI value of the signals received by each antenna in the plurality of antennas; the minimum RSSI value of the signals received by each antenna in the plurality of antennas; and the RSSI skewness value of the signals received by each antenna in the plurality of antennas.

[0016] According to an embodiment of this disclosure, the antenna is disposed on a passageway door for transporting the plurality of items;

[0017] The receiving of multiple signals emitted by multiple items arranged in a preset manner includes: receiving the multiple signals through the multiple antennas installed on the passage door when transporting the multiple items through the passage door.

[0018] According to embodiments of this disclosure, the passageway door includes a first passageway door and a second passageway door;

[0019] The first channel door includes a channel door frame, a wave-absorbing material, and a shell;

[0020] The second channel door includes a channel door frame, a wave-absorbing material, and a shell;

[0021] The antenna is disposed in the first channel gate and / or the second channel gate;

[0022] The antenna is connected to the communication module, and the communication module is connected to the industrial control computer;

[0023] The communication module is installed in the first channel door and / or the second channel door;

[0024] The industrial control computer is installed in either the first or second access door.

[0025] According to an embodiment of this disclosure, the absorbing material is installed on the outside of the channel door frame and covers the periphery of the channel door frame;

[0026] The antenna is exposed on the outside of the absorbing material;

[0027] The outer casing is made of non-metallic material and covers the channel door frame, the absorbing material, and the periphery of the antenna.

[0028] According to embodiments of this disclosure, the machine learning model includes any one of the following: random forest model, decision tree model, support vector machine model, and K-nearest neighbor model.

[0029] According to embodiments of this disclosure, the machine learning model is trained through the following process:

[0030] Receive multiple signals emitted by multiple items arranged in the preset manner, wherein the signals contain identification information of the corresponding items;

[0031] Extract the specified feature from the plurality of signals;

[0032] The location of the items whose signals were received and / or not received is determined based on the identification information contained in the multiple signals, and this is used as the positioning result.

[0033] The machine learning model is trained using the specified features and the localization results.

[0034] According to embodiments of this disclosure, the locations of items among the plurality of items where signals are received and / or not received are displayed in a visual manner.

[0035] Secondly, this disclosure provides a model training method, including:

[0036] Receive multiple signals emitted by multiple items arranged in a preset manner, wherein the signals contain identification information of the corresponding items;

[0037] Extract specified features from the plurality of signals;

[0038] The location of the items whose signals were received and / or not received is determined based on the identification information contained in the multiple signals, and this is used as the positioning result.

[0039] The machine learning model is trained using the specified features and the localization results.

[0040] Thirdly, this disclosure provides an item positioning system, including: a communication module, an industrial control computer, and multiple antennas, wherein:

[0041] The communication module is configured to receive multiple signals emitted by multiple items placed in a preset manner via the multiple antennas;

[0042] The industrial control computer is configured to extract specified features from the multiple signals, input the specified features into a trained machine learning model, and determine the location of the items among the multiple items whose signals are received and / or not received.

[0043] According to an embodiment of this disclosure, the system further includes a channel gate in which the plurality of antennas are disposed.

[0044] According to embodiments of this disclosure, the passageway door includes a first passageway door and a second passageway door;

[0045] The first channel door includes a channel door frame, a wave-absorbing material, and a shell;

[0046] The second channel door includes a channel door frame, a wave-absorbing material, and a shell;

[0047] The antennas are symmetrically arranged in the first channel gate and the second channel gate;

[0048] The antenna is connected to the communication module, and the communication module is connected to the industrial control computer;

[0049] The communication module is installed in the first channel door and / or the second channel door;

[0050] The industrial control computer is installed in either the first or second access door.

[0051] According to embodiments of this disclosure, the industrial control computer is further configured to train the machine learning model through the following process:

[0052] Receive multiple signals emitted by multiple items arranged in the preset manner, wherein the signals contain identification information of the corresponding items;

[0053] Extract the specified feature from the plurality of signals;

[0054] The location of the items whose signals were received and / or not received is determined based on the identification information contained in the multiple signals, and this is used as the positioning result.

[0055] The machine learning model is trained using the specified features and the localization results.

[0056] Fourthly, this disclosure provides an article positioning device, including:

[0057] The first receiving module is configured to receive multiple signals emitted by multiple items placed in a preset manner;

[0058] The first extraction module is configured to extract specified features from the plurality of signals;

[0059] The first determining module is configured to input the specified features into a trained machine learning model to determine the location of the items among the plurality of items where the signal is received and / or not received.

[0060] Fifthly, this disclosure provides a model training apparatus, comprising:

[0061] The second receiving module is configured to receive multiple signals emitted by multiple items placed in a preset manner, wherein the signals contain identification information of the corresponding items;

[0062] The second extraction module is configured to extract specified features from the plurality of signals;

[0063] The second determining module is configured to determine the location of the items whose signals are received and / or not received among the multiple items based on the identification information contained in the multiple signals, as a positioning result;

[0064] The training module is configured to train the machine learning model using the specified features and the localization results.

[0065] In a sixth aspect, embodiments of this disclosure provide an electronic device including a memory and a processor, wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method as described in either the first or second aspect.

[0066] In a seventh aspect, this disclosure provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the methods described in the first and second aspects.

[0067] Eighthly, this disclosure provides a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program containing program code for performing the method of any one of the first and second aspects of claims.

[0068] According to the technical solution provided in this disclosure, a method for locating items is disclosed, comprising: receiving multiple signals emitted by multiple items arranged in a preset manner; extracting specified features from the multiple signals; inputting the specified features into a trained machine learning model to determine the location of items among the multiple items whose signals have been received and / or not received. This technical solution can accurately identify the location of items whose signals have been received and / or not received, given a known arrangement of items, helping staff to accurately and quickly locate unidentified items. In scenarios where RFID is used for warehouse entry and exit identification, this technical solution can accurately and quickly locate the location of items with missed RFID tags, facilitating accurate and rapid manual re-entry, greatly improving the usability and scalability of using RFID for warehouse entry and exit management, significantly reducing manpower and material resources consumed during inventory checks, and improving the accuracy and efficiency of entry and exit.

[0069] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0070] Other features, objects, and advantages of this disclosure will become more apparent from the following detailed description of non-limiting embodiments, taken in conjunction with the accompanying drawings. In the drawings:

[0071] Figure 1 shows a schematic diagram of an application scenario of an embodiment of this disclosure.

[0072] Figure 2 shows a flowchart of an article positioning method according to an embodiment of the present disclosure.

[0073] Figure 3 shows a flowchart of a model training method according to an embodiment of the present disclosure.

[0074] Figure 4 shows a schematic diagram of an article positioning system according to an embodiment of the present disclosure.

[0075] Figure 5 shows a structural block diagram of an article positioning device according to an embodiment of the present disclosure.

[0076] Figure 6 shows a structural block diagram of a model training apparatus according to an embodiment of the present disclosure.

[0077] Figure 7 shows a structural block diagram of an electronic device according to an embodiment of the present disclosure.

[0078] Figure 8 shows a schematic diagram of the structure of a computer system suitable for implementing the method according to an embodiment of the present disclosure. Detailed Implementation

[0079] In the following, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings to enable those skilled in the art to readily implement them. Furthermore, for clarity, portions unrelated to the description of exemplary embodiments have been omitted from the drawings.

[0080] In this disclosure, it should be understood that terms such as “comprising” or “having” are intended to indicate the presence of features, figures, steps, behaviors, components, parts or combinations thereof disclosed in this specification, and are not intended to exclude the possibility of the presence or addition of one or more other features, figures, steps, behaviors, components, parts or combinations thereof.

[0081] It should also be noted that, unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other. This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0082] In this disclosure, any operation involving the acquisition of user information or user data, or the display of user information or user data to others, is an operation authorized or confirmed by the user, or actively selected by the user.

[0083] As described in the background section, existing intelligent warehouse inbound / outbound identification solutions often employ vision-based detection devices or RFID-based access gates. Vision-based devices typically only allow for single-item detection and identification, and are limited by line-of-sight distance, obstructions, and lighting conditions, making batch reading difficult and impacting inbound / outbound management efficiency. RFID-based access gates, on the other hand, can read items in batches, unaffected by obstructions or lighting conditions, significantly improving management efficiency. However, due to the electromagnetic shielding effect of metal, when densely packed items composed of numerous metal components pass through the access gate, the RFID tags attached to the goods may be missed.

[0084] To address the aforementioned technical problems, this invention discloses an item positioning method, comprising: receiving multiple signals emitted by multiple items arranged in a preset manner; extracting specified features from the multiple signals; inputting the specified features into a trained machine learning model to determine the positions of items among the multiple items whose signals were received and / or not received.

[0085] The technical solution disclosed herein can accurately identify the location of items whose signals have been received and / or not received, given a known arrangement of items, helping staff to accurately and quickly locate unidentified items. In scenarios where RFID is used for warehouse inbound and outbound identification, this technical solution can accurately and quickly locate the location of items with missed RFID tags, facilitating accurate and rapid manual re-entry. This significantly improves the usability and scalability of using RFID for warehouse inbound and outbound management, greatly reduces the manpower and material resources consumed during inventory checks, and improves the accuracy and efficiency of inbound and outbound operations.

[0086] For ease of explanation, the embodiments of this disclosure are described below using a scenario employing RFID technology as an example. However, those skilled in the art will understand that the embodiments of this disclosure are not limited to RFID technology, but can be applied to various scenarios where information about an item is identified by signals emitted from the item.

[0087] Figure 1 shows a schematic diagram of an application scenario of an embodiment of this disclosure.

[0088] As shown in Figure 1, multiple items 110 (e.g., goods) are arranged on a truck in a known layout. For example, these items can be neatly arranged in one or more layers, with each layer arranged in one or more rows and columns. According to embodiments of this disclosure, the items can be densely arranged, i.e., placed one next to another, or with small gaps between them. Each item is affixed with an RFID tag 120. By reading the RFID tag 120, relevant information about the item can be obtained, such as the item name, model, parameters, batch number, etc., but not limited to these.

[0089] As mentioned above, when items contain metal components and are densely packed, the electromagnetic shielding effect of metal may cause missed reads when using an RFID reader to read the RFID tag 120 on the items. Applying the technical solution of this disclosure embodiment can quickly locate the positions of read and / or unread items, greatly reducing the manpower and material resources consumed during inventory checks.

[0090] Figure 2 shows a flowchart of an article positioning method according to an embodiment of the present disclosure. As shown in Figure 2, the article positioning method includes the following steps S101-S103.

[0091] In step S101, multiple signals emitted by multiple items arranged in a preset manner are received.

[0092] In step S102, specified features are extracted from the plurality of signals.

[0093] In step S103, the specified features are input into the trained machine learning model to determine the locations of items among the plurality of items where the signal is received and / or not received.

[0094] The specified features of signals received by a signal receiving device from multiple items are affected by factors such as the distance from the corresponding item to the signal receiving device, the placement relationship between the items, and shielding relationships, thus reflecting the position of the corresponding item. Based on this principle, embodiments of this disclosure extract specified features from the signals received from multiple items, and use these specified features and a trained machine learning model to determine the positions of items among the multiple items whose signals were received and / or not received. According to embodiments of this disclosure, the machine learning model is a pre-trained model used to determine the positions of items among the multiple items whose signals were received and / or not received based on the specified features, and may include, for example, any of the following: random forest model, decision tree model, support vector machine model, or K-nearest neighbor model.

[0095] According to embodiments of this disclosure, the plurality of items may have the same shape. For example, the plurality of items may be packaging boxes of the same size containing goods. Alternatively, the plurality of items may have different shapes, as long as their arrangement during model training is the same as their arrangement during actual use.

[0096] According to embodiments of this disclosure, the plurality of articles may or may not include metal or other materials that may shield radio signals.

[0097] According to embodiments of this disclosure, the preset arrangement can be a neat arrangement of one or more layers, with each layer of items neatly arranged in one or more rows and one or more columns. According to embodiments of this disclosure, items can be densely arranged, i.e., placed one next to another, or with small gaps between them. Alternatively, the preset arrangement may not be neat, as long as the arrangement of items during model training is the same as the arrangement during actual use.

[0098] According to embodiments of this disclosure, the signal transmitting device for each item is located at the same position. For example, when the signal transmitting device is an RFID tag, the RFID tag can be placed at the same position on the item (e.g., a packaging box). The RFID signal emitted by the RFID tag may contain relevant information about the item, such as the item name, model, parameters, batch number, etc., but is not limited thereto. Alternatively, the signal transmitting device for each item may be located at different positions, as long as the positions of the signal transmitting devices during model training are the same as the positions of the signal transmitting devices during actual use.

[0099] According to embodiments of this disclosure, receiving multiple signals emitted by multiple items arranged in a preset manner includes: receiving the multiple signals through multiple antennas disposed on both sides of a passage for transporting the multiple items. For example, multiple antennas can be disposed on the side walls of the passage, with one or more antennas disposed on each side wall. Alternatively, one or more antennas can be disposed at the bottom and / or top of the passage as needed.

[0100] According to embodiments of this disclosure, the plurality of antennas can be disposed on the channel gates. When there are two channel gates arranged opposite each other, one or more antennas can be disposed on each channel gate. In a specific example, two antennas are disposed on each channel gate, and the antenna positions on the two channel gates are symmetrical to each other.

[0101] According to embodiments of this disclosure, extracting a specified feature from the plurality of signals includes extracting the specified feature based on the signal strength indication (RSSI) of the signals received by the plurality of antennas.

[0102] According to embodiments of this disclosure, the specified features include any one or more of the following: the ratio of the number of times each antenna in the plurality of antennas receives a signal to the total number of times the plurality of antennas receive a signal; the average RSSI of the signals received by the plurality of antennas; the maximum RSSI of the signals received by the plurality of antennas; the minimum RSSI of the signals received by the plurality of antennas; and the RSSI skewness value of the signals received by the plurality of antennas.

[0103] Specifically, for a specific RFID tag i on one of multiple items, the RSSI sequence read by an antenna j can be represented as follows: The ratio of the number of times each antenna receives a signal to the total number of times the multiple antennas receive a signal is:

[0104]

[0105] The average RSSI of the signals received by each of the multiple antennas is:

[0106]

[0107] The maximum RSSI value of the signal received by each of the plurality of antennas is:

[0108]

[0109] The minimum RSSI value of the signal received by each of the multiple antennas is:

[0110]

[0111] The RSSI skewness value of the signal received by each of the plurality of antennas is:

[0112]

[0113] in

[0114] After obtaining the average RSSI, maximum RSSI, minimum RSSI, RSSI skewness, and the ratio of the number of times each antenna received a signal to the total number of times the multiple antennas received a signal, the above parameters are input into a trained machine learning model, which can accurately locate items whose signals were received and / or items whose signals were not received.

[0115] According to embodiments of this disclosure, when a positioning activation condition is met, the reception of the plurality of signals begins. Meeting the positioning activation condition means that multiple items are passing through a passageway sidewall or passageway door equipped with an antenna. The positioning activation condition may be that the number of tags continuously read by the RFID reader exceeds a preset threshold, at which point it is considered that multiple items are passing through the passageway sidewall or passageway door equipped with an antenna. Alternatively, a sensor can be installed at the passageway or passageway door; when the sensor's output signal indicates that multiple items are passing through the passageway sidewall or passageway door equipped with an antenna, the positioning activation condition is met. When the positioning activation condition is met, the continuous reading of RFID signals for a preset time period (e.g., 10 seconds) can begin as the plurality of signals. During this preset time period, the multiple items can remain stationary or in motion. When the multiple items remain stationary, stable signal reception can be achieved; when the multiple items are in motion, the positional relationship between the antenna and the items can be changed, allowing the antenna to receive signals from multiple angles, increasing the probability of signal reception by the antenna.

[0116] Within the aforementioned preset time period, each antenna may receive signals from each item multiple times, thus allowing the generation of an RSSI sequence based on the signal received from any item by each antenna. For example, for an RFID tag i on one of multiple items, the RSSI sequence read by antenna j can be represented as follows: Then, specified features are extracted from multiple signals read during the preset time period, and the specified features are input into a trained machine learning model to determine the location of the items whose signals were received and / or not received among the multiple items.

[0117] According to embodiments of this disclosure, the machine learning model includes any one of the following: random forest model, decision tree model, support vector machine model, and K-nearest neighbor model. Random forest belongs to a type of ensemble learning model in machine learning. Its essential characteristic is that it votes on the classification results of several weak classifiers to generate a strong classifier. Specifically, assuming the random forest model uses the results of T decision tree models for final voting, and the training set size is N with M features, then for each decision tree model, n sets are selected as input from the N sets of training data with replacement, and m (m < M) features are selected as branches of the decision tree. The results of all decision tree models are statistically analyzed, and the classification results are voted on. The category with the highest votes is the output result of the random forest model. This model can effectively avoid the influence of abnormal training input data on the training results and improve the accuracy of the model's judgment. Support vector machine is a binary classification model that finds a hyperplane in the feature space to separate data of different categories. The core of support vector machine is to maximize the classification margin, that is, to maximize the distance of the data point closest to the hyperplane. The K-Nearest Neighbors (KNN) model is an instance-based learning algorithm that predicts the label of an input sample based on the labels of its K nearest neighbors in the training dataset. The choice of K and the distance metric are key factors affecting KNN performance. By inputting specified features into a trained machine learning model, the model can determine the location of items that received signals and / or the location of items that did not receive signals, accurately locating missed items and facilitating the statistical analysis of incoming and outgoing items.

[0118] According to embodiments of this disclosure, the locations of items among the plurality of items whose signals were received and / or not received can also be displayed visually. Visualization allows the user to intuitively understand the locations of items that did not receive signals, facilitating subsequent supplementary information entry for these items.

[0119] Figure 3 shows a flowchart of a model training method according to an embodiment of the present disclosure. As shown in Figure 3, the model training method includes the following steps S201–S204.

[0120] In step S201, multiple signals emitted by multiple items arranged in a preset manner are received, and the signals contain identification information of the corresponding items.

[0121] In step S202, specified features are extracted from the plurality of signals.

[0122] In step S203, the location of the items whose signals were received and / or not received is determined based on the identification information contained in the plurality of signals, and is used as the positioning result.

[0123] In step S204, the machine learning model is trained using the specified features and the localization results.

[0124] According to embodiments of this disclosure, a machine learning model can be trained through the following process.

[0125] First, multiple items with known identification information are placed on the truck according to a preset method (i.e., the arrangement during actual transportation), where the correspondence between the identification information of each item and its placement position is known. When a positioning activation condition is met, the multiple signals are received. Meeting the positioning activation condition means that multiple items are passing through a passageway sidewall or passageway door equipped with antennas. The positioning activation condition can be that the number of tags continuously read by the RFID reader exceeds a preset threshold, at which point it is considered that multiple items are passing through a passageway sidewall or passageway door equipped with antennas. Alternatively, sensors can be installed at the passageway or passageway door; when the sensor's output signal indicates that multiple items are passing through a passageway sidewall or passageway door equipped with antennas, the positioning activation condition is met. When the positioning activation condition is met, RFID signals for a preset time period can be continuously read as the multiple signals. During this preset time period, the multiple items can remain stationary or in motion, but must maintain consistency with their actual state during the preset time period during transportation.

[0126] Different items correspond to different identification information. Therefore, by parsing the received signal, the identification information of the item whose signal was received can be obtained, and then the location of the item whose signal was received can be obtained. Furthermore, the location of the item whose signal was not received can be inferred. The location of the item whose signal was received and / or the item whose signal was not received can be used as the positioning result.

[0127] Then, specified features are extracted from the received signals, such as any one or more of the following: the ratio of the number of times each antenna in the plurality of antennas receives signals to the total number of times the plurality of antennas receive signals; the average RSSI of the signals received by each antenna in the plurality of antennas; the maximum RSSI of the signals received by each antenna in the plurality of antennas; the minimum RSSI of the signals received by each antenna in the plurality of antennas; and the RSSI skewness value of the signals received by each antenna in the plurality of antennas.

[0128] In this way, multiple sample data can be obtained, each of which includes corresponding specified features and localization results.

[0129] A portion of the sample data is used as training data, and another portion is used as test data to train the machine learning model. Specifically, specified features are used as input to the machine learning model, and the localization results are used as output. When the sample data is limited, cross-validation can be used to evaluate the model's accuracy.

[0130] Figure 4 shows a schematic diagram of an article positioning system according to an embodiment of the present disclosure.

[0131] As shown in Figure 4, the item positioning system according to an embodiment of this disclosure includes a first channel door 310 and a second channel door 320. The first channel door includes a channel door frame 311, a wave-absorbing material 400, and a housing 312. The second channel door includes a channel door frame 321, a wave-absorbing material 400, and a housing 322. The first channel door 310 and the second channel door 320 can be arranged parallel to each other on both sides of the channel, facilitating the placement of antennas from different angles.

[0132] The first and second passage doors are located on either side of a passageway for transporting the multiple items, which pass between them. An antenna is disposed within the first and / or second passage doors. The antenna is positioned at multiple locations within the first and / or second passage doors to receive multiple signals emitted by the multiple items arranged in a predetermined manner from different angles. In the example shown in Figure 4, two antennas are symmetrically disposed within each of the first and second passage doors.

[0133] According to embodiments of this disclosure, the antenna is connected to a communication module, which in turn is connected to an industrial control computer. For example, the communication module may be an RFID reader / writer used to read RFID signals received by the antenna. The industrial control computer is used to obtain relevant information about the item based on the signals received by the antenna, and to locate the item using the method according to embodiments of this disclosure based on the signals received by the antenna.

[0134] According to an embodiment of this disclosure, the industrial control computer is further configured to train the machine learning model through the following process: receiving multiple signals emitted by multiple items placed in the preset manner, the signals containing identification information of the corresponding items; extracting the specified features from the multiple signals; determining the positions of the items whose signals are received and / or not received among the multiple items based on the identification information contained in the multiple signals as positioning results; and training the machine learning model using the specified features and the positioning results.

[0135] According to embodiments of this disclosure, the communication module is disposed in the first passage door 310 and / or the second passage door 320. Disposing the communication module in the first passage door 310 and / or the second passage door 320 facilitates co-deployment with both the first passage door 310 and the second passage door 320. Optionally, the communication module can be disposed in a location outside the first passage door 310 and the second passage door 320, such as a control room.

[0136] According to embodiments of this disclosure, the industrial control computer is installed in either the first passageway door 310 or the second passageway door 320. Installing the industrial control computer in the first passageway door 310 and / or the second passageway door 320 facilitates joint deployment with both doors. Optionally, the industrial control computer can be installed in a location other than the first passageway door 310 and the second passageway door 320, such as a control room.

[0137] According to embodiments of this disclosure, the absorbing material 400 is installed on the outside of the channel door frame and covers the perimeter of the channel door frame. The function of the absorbing material 400 is to reduce electromagnetic wave reflection and interference, and optimize the signal reception performance of the antenna. Specifically, when an RFID tag is attached to a conductive object such as metal, electromagnetic waves will be reflected and attenuated on the metal surface, causing the antenna to be unable to accurately read the RFID tag information. The absorbing material 400 can absorb electromagnetic waves, thereby reducing electromagnetic wave reflection and interference, and improving the sensitivity and accuracy of the antenna.

[0138] According to an embodiment of this disclosure, the antenna is exposed on the outside of the absorbing material 400.

[0139] According to an embodiment of this disclosure, the housing is made of a non-metallic material and covers the channel door frame, the absorbing material 400, and the periphery of the antenna.

[0140] The item positioning system shown in Figure 4 integrates an antenna, communication module, industrial control computer, etc. in the passage door. However, the item positioning system according to the embodiments of this disclosure may also not include the passage door. For example, the antenna can be deployed on the side wall and / or top surface and / or ground of the passage, and the signal received by the antenna can be sent to the communication module in a wired or wireless manner, and then sent to the industrial control computer.

[0141] The method and system according to the embodiments of this disclosure can accurately and efficiently locate missed items. To verify the feasibility and accuracy of the embodiments of this disclosure, an experimental platform was built and the algorithm was tested. A simple passageway gate was built with a width of 1.7m. The installation heights of the RFID directional antennas on one side were 35cm and 75cm, respectively. Electricity meters were used as goods, densely stacked in a 3×4×2 pattern, ensuring a center height of 55cm. The stacks were placed on a handcart and passed through the passageway gate, and data was collected for 10 seconds. A total of 200 sets of data were collected as the training set. Due to the small amount of data, five-fold cross-validation was used to evaluate the model accuracy. The results show that the random forest method achieved a judgment accuracy of 99.3%, the ordinary decision tree model achieved 96.9%, the support vector machine model achieved the highest accuracy of 98.4%, and the K-nearest neighbor model achieved the highest accuracy of 96.4%.

[0142] Figure 5 shows a structural block diagram of an article positioning device according to an embodiment of the present disclosure. This device can be implemented as part or all of an electronic device through software, hardware, or a combination of both.

[0143] As shown in Figure 5, the item positioning device 500 includes a first receiving module 510, a first extraction module 520, and a first determining module 530.

[0144] The first receiving module 510 is configured to receive multiple signals emitted by multiple items placed in a preset manner;

[0145] The first extraction module 520 is configured to extract specified features from the plurality of signals;

[0146] The first determining module 530 is configured to input the specified features into a trained machine learning model to determine the location of the items among the plurality of items where the signal is received and / or not received.

[0147] The specific characteristics of signals received by a signal receiving device from multiple items are affected by factors such as the distance from the corresponding item to the signal receiving device, the placement relationship between the items, and the shielding relationship, thereby reflecting the position of the corresponding item. Based on this principle, embodiments of this disclosure extract specific features from signals received from multiple items, and use these specific features and a trained machine learning model to determine the positions of items among the multiple items whose signals were received and / or not received.

[0148] According to an embodiment of this disclosure, the item positioning device 500 includes multiple items having the same shape; the preset placement method is a neat arrangement; and a signal is transmitted by a signal transmitting device located at the same position on the items; the signal is an RFID signal.

[0149] According to an embodiment of this disclosure, the item positioning device 500, wherein receiving multiple signals emitted by multiple items placed in a preset manner includes: receiving the multiple signals through multiple antennas disposed on both sides of a channel for transporting the multiple items; and extracting specified features from the multiple signals includes extracting the specified features based on the signal strength index RSSI of the signals received by the multiple antennas.

[0150] According to embodiments of this disclosure, the item positioning device 500 includes the following specified features: the ratio of the number of times each antenna receives a signal to the total number of times the multiple antennas receive a signal; the average RSSI of the signals received by each antenna; the maximum RSSI of the signals received by each antenna; the minimum RSSI of the signals received by each antenna; and the RSSI skewness value of the signals received by each antenna. By inputting the above parameters into a trained machine learning model, the device can accurately locate items for which signals are received and / or items for which signals are not received.

[0151] According to an embodiment of this disclosure, the item positioning device 500 includes an antenna disposed on a passage door for transporting the plurality of items; receiving the plurality of signals emitted by the plurality of items placed in a preset manner includes: receiving the plurality of signals through the plurality of antennas disposed on the passage door when transporting the plurality of items through the passage door.

[0152] According to an embodiment of this disclosure, the item positioning device 500 includes a channel door comprising a first channel door and a second channel door; the first channel door includes a channel door frame, a wave-absorbing material, and a housing; the second channel door includes a channel door frame, a wave-absorbing material, and a housing; an antenna is disposed in the first channel door and / or the second channel door; the antenna is connected to a communication module, and the communication module is connected to an industrial control computer; the communication module is disposed in the first channel door and / or the second channel door; and the industrial control computer is disposed in the first channel door or the second channel door.

[0153] According to an embodiment of this disclosure, the item positioning device 500 includes an absorbing material installed on the outside of a channel door frame and covering the periphery of the channel door frame; an antenna exposed on the outside of the absorbing material; and a housing made of a non-metallic material covering the channel door frame, the absorbing material, and the periphery of the antenna.

[0154] According to embodiments of this disclosure, the item positioning device 500 includes a machine learning model comprising any one of the following: a random forest model, a decision tree model, a support vector machine model, or a K-nearest neighbor model.

[0155] According to an embodiment of this disclosure, the item positioning device 500 trains the machine learning model through the following process: receiving multiple signals emitted by multiple items placed in a preset manner, the signals containing identification information of the corresponding items; extracting the specified features from the multiple signals; determining the positions of the items whose signals were received and / or not received based on the identification information contained in the multiple signals as positioning results; and training the machine learning model using the specified features and the positioning results.

[0156] Figure 6 shows a structural block diagram of a model training apparatus according to an embodiment of the present disclosure. This apparatus can be implemented as part or all of an electronic device through software, hardware, or a combination of both.

[0157] As shown in Figure 6, the model training device 600 includes a second receiving module 610, a second extraction module 620, a second determining module 630, and a training module 640.

[0158] The second receiving module 610 is configured to receive multiple signals emitted by multiple items placed in a preset manner, wherein the signals contain identification information of the corresponding items;

[0159] The second extraction module 620 is configured to extract specified features from the plurality of signals;

[0160] The second determining module 630 is configured to determine the location of the items whose signals are received and / or not received among the multiple items based on the identification information contained in the multiple signals, as a positioning result;

[0161] Training module 640 is configured to train the machine learning model using the specified features and the localization results.

[0162] According to embodiments of this disclosure, a machine learning model can be trained through the following process.

[0163] First, multiple items with known identification information are placed on the truck according to a preset method (i.e., the arrangement during actual transportation), where the correspondence between the identification information of each item and its placement position is known. When a positioning activation condition is met, the multiple signals are received. Meeting the positioning activation condition means that multiple items are passing through a passageway sidewall or passageway door equipped with antennas. The positioning activation condition can be that the number of tags continuously read by the RFID reader exceeds a preset threshold, at which point it is considered that multiple items are passing through a passageway sidewall or passageway door equipped with antennas. Alternatively, sensors can be installed at the passageway or passageway door; when the sensor's output signal indicates that multiple items are passing through a passageway sidewall or passageway door equipped with antennas, the positioning activation condition is met. When the positioning activation condition is met, RFID signals for a preset time period can be continuously read as the multiple signals. During this preset time period, the multiple items can remain stationary or in motion, but must maintain consistency with their actual state during the preset time period during transportation.

[0164] Different items correspond to different identification information. Therefore, by parsing the received signal, the identification information of the item whose signal was received can be obtained, and then the location of the item whose signal was received can be obtained. Furthermore, the location of the item whose signal was not received can be inferred. The location of the item whose signal was received and / or the item whose signal was not received can be used as the positioning result.

[0165] Then, specified features are extracted from the received signals, such as any one or more of the following: the ratio of the number of times each antenna in the plurality of antennas receives signals to the total number of times the plurality of antennas receive signals; the average RSSI of the signals received by each antenna in the plurality of antennas; the maximum RSSI of the signals received by each antenna in the plurality of antennas; the minimum RSSI of the signals received by each antenna in the plurality of antennas; and the RSSI skewness value of the signals received by each antenna in the plurality of antennas.

[0166] In this way, multiple sample data can be obtained, each of which includes corresponding specified features and localization results.

[0167] A portion of the sample data is used as training data, and another portion is used as test data to train the machine learning model. Specifically, specified features are used as input to the machine learning model, and the localization results are used as output. When the sample data is limited, cross-validation can be used to evaluate the model's accuracy.

[0168] This disclosure also discloses an electronic device, and FIG7 shows a structural block diagram of the electronic device according to an embodiment of the present disclosure. The electronic device according to an embodiment of the present invention can be implemented as an industrial control computer in the present invention.

[0169] As shown in FIG7, the electronic device includes a memory and a processor, wherein the memory is used to store computer instructions, wherein the computer instructions are executed by the processor to implement the method according to embodiments of the present disclosure.

[0170] A method for locating an item, comprising:

[0171] Receives multiple signals emitted by multiple items arranged in a preset manner;

[0172] Extract specified features from the plurality of signals;

[0173] The specified features are input into a trained machine learning model to determine the locations of items among the plurality of items where signals are received and / or not received.

[0174] The multiple items have the same shape;

[0175] The preset placement method is a neat arrangement;

[0176] The signal is transmitted by a signal transmitting device positioned at the same location as the item;

[0177] The signal is an RFID signal.

[0178] The receiving of multiple signals emitted by multiple items arranged in a preset manner includes: receiving the multiple signals through multiple antennas set on both sides of the channel used for transporting the multiple items;

[0179] The step of extracting a specified feature from the plurality of signals includes extracting the specified feature based on the signal strength indication (RSSI) of the signals received by the plurality of antennas.

[0180] The specified features include any one or more of the following: the ratio of the number of times each antenna in the plurality of antennas receives a signal to the total number of times the plurality of antennas receive a signal; the average RSSI of the signals received by each antenna in the plurality of antennas; the maximum RSSI of the signals received by each antenna in the plurality of antennas; the minimum RSSI of the signals received by each antenna in the plurality of antennas; and the RSSI skewness value of the signals received by each antenna in the plurality of antennas.

[0181] The antenna is mounted on the passageway door used for transporting the plurality of items;

[0182] The receiving of multiple signals emitted by multiple items arranged in a preset manner includes: receiving the multiple signals through the multiple antennas installed on the passage door when transporting the multiple items through the passage door.

[0183] The passageway door includes a first passageway door and a second passageway door;

[0184] The first channel door includes a channel door frame, a wave-absorbing material, and a shell;

[0185] The second channel door includes a channel door frame, a wave-absorbing material, and a shell;

[0186] The antenna is disposed in the first channel gate and / or the second channel gate;

[0187] The antenna is connected to the communication module, and the communication module is connected to the industrial control computer;

[0188] The communication module is installed in the first channel door and / or the second channel door;

[0189] The industrial control computer is installed in either the first or second access door.

[0190] The absorbing material is installed on the outside of the channel door frame and covers the periphery of the channel door frame;

[0191] The antenna is exposed on the outside of the absorbing material;

[0192] The outer casing is made of non-metallic material and covers the channel door frame, the absorbing material, and the periphery of the antenna.

[0193] The machine learning model includes any one of the following: random forest model, decision tree model, support vector machine model, and K-nearest neighbor model.

[0194] The machine learning model is trained using the following process:

[0195] Receive multiple signals emitted by multiple items arranged in the preset manner, wherein the signals contain identification information of the corresponding items;

[0196] Extract the specified feature from the plurality of signals;

[0197] The location of the items whose signals were received and / or not received is determined based on the identification information contained in the multiple signals, and this is used as the positioning result.

[0198] The machine learning model is trained using the specified features and the localization results.

[0199] The item positioning method further includes:

[0200] The locations of the items in the plurality of items where signals were received and / or not received are displayed in a visual manner.

[0201] A model training method, comprising:

[0202] Receive multiple signals emitted by multiple items arranged in a preset manner, wherein the signals contain identification information of the corresponding items;

[0203] Extract specified features from the plurality of signals;

[0204] The location of the items whose signals were received and / or not received is determined based on the identification information contained in the multiple signals, and this is used as the positioning result.

[0205] The machine learning model is trained using the specified features and the localization results.

[0206] Figure 8 shows a schematic diagram of a computer system suitable for implementing the method according to an embodiment of the present disclosure. The computer system according to an embodiment of the present invention can be implemented as an industrial control computer as described in this invention.

[0207] As shown in Figure 8, the computer system includes a processing unit that can execute various methods described above based on a program stored in a read-only memory (ROM) or a program loaded from a storage portion into a random access memory (RAM). The RAM also stores various programs and data required for the operation of the computer system. The processing unit, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0208] The following components are connected to the I / O interface: input sections including keyboards, mice, etc.; output sections including cathode ray tubes (CRTs), liquid crystal displays (LCDs), and speakers; storage sections including hard disks; and communication sections including network interface cards such as LAN cards and modems. The communication section performs communication via a network such as the Internet. Drives are also connected to the I / O interface as needed. Removable media, such as disks, optical disks, magneto-optical disks, semiconductor memories, etc., are installed on the drive as needed so that computer programs read from them can be installed into the storage section as needed. The processing unit can be implemented as a CPU, GPU, TPU, FPGA, NPU, etc. The input sections, output sections, and removable media are suitable for industrial control computers located in control rooms or other environments far from access doors.

[0209] In particular, according to embodiments of this disclosure, the methods described above can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program containing program code for performing the methods described above. In such embodiments, the computer program can be downloaded and installed from a network via a communication component, and / or installed from a removable medium.

[0210] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0211] The units or modules described in the embodiments of this disclosure can be implemented in software or programmable hardware. The described units or modules can also be located in a processor, and the names of these units or modules do not necessarily constitute a limitation on the unit or module itself.

[0212] In another aspect, this disclosure also provides a computer-readable storage medium, which may be a computer-readable storage medium included in the electronic device or computer system described above; or it may be a standalone computer-readable storage medium not assembled into a device. The computer-readable storage medium stores one or more programs, which are used by one or more processors to perform the methods described in this disclosure.

[0213] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

Claims

1. A method for locating an item, characterized in that, include: When the positioning activation conditions are met, multiple signals emitted by multiple items arranged in a preset layout are received by multiple antennas installed on both sides of the passage for transporting multiple items, or by multiple antennas installed on both sides of the passage for transporting the multiple items, and at the bottom and / or top of the passage. The multiple items have the same shape, and the preset layout is a neat arrangement. The signals are emitted by signal transmitting devices installed at the same position on the items. The positioning activation conditions include: the number of tags continuously read by the RFID reader exceeding a preset threshold, or the output signal of a sensor installed at the passage or passage door indicating that multiple items are passing through the side wall of the passage equipped with antennas. Alternatively, a channel gate can be used; based on the Signal Strength Indicator (RSSI) of the signals received by the multiple antennas, specified features are extracted from the multiple signals, wherein the specified features include: the ratio of the number of times each antenna in the multiple antennas receives a signal to the total number of times the multiple antennas receive a signal; the average RSSI of the signals received by each antenna in the multiple antennas; the maximum RSSI of the signals received by each antenna in the multiple antennas; the minimum RSSI of the signals received by each antenna in the multiple antennas; and the RSSI skewness value of the signals received by each antenna in the multiple antennas. The specified features are then input into a trained machine learning model to determine the location of the items whose signals were not received among the multiple items.

2. The method according to claim 1, wherein: The signal is an RFID signal.

3. The method according to claim 1, wherein: The antenna is mounted on a passageway door for transporting the multiple items; receiving multiple signals emitted by the multiple items arranged in a preset manner includes: receiving the multiple signals through the multiple antennas mounted on the passageway door when transporting the multiple items through the passageway door.

4. The method according to claim 3, wherein: The passageway includes a first passageway and a second passageway; the first passageway includes a passageway frame, a wave-absorbing material, and a shell; the second passageway includes a passageway frame, a wave-absorbing material, and a shell; the antenna is disposed in the first passageway and / or the second passageway; the antenna is connected to a communication module, and the communication module is connected to an industrial control computer; the communication module is disposed in the first passageway and / or the second passageway; the industrial control computer is disposed in the first passageway or the second passageway.

5. The method according to claim 4, wherein: The absorbing material is installed on the outside of the channel door frame and covers the periphery of the channel door frame; the antenna is exposed on the outside of the absorbing material; the housing is made of non-metallic material and covers the channel door frame, the absorbing material and the periphery of the antenna.

6. The method according to claim 1, wherein, The machine learning model includes any one of the following: random forest model, decision tree model, support vector machine model, and K-nearest neighbor model.

7. The method according to claim 1, wherein, The machine learning model is trained through the following process: receiving multiple signals emitted by multiple items arranged in the preset arrangement, the signals containing identification information of the corresponding items; extracting the specified features from the multiple signals; determining the positions of the items whose signals were received and / or not received based on the identification information contained in the multiple signals as the positioning results; and training the machine learning model using the specified features and the positioning results.

8. The method according to claim 1, further comprising: The location of the items whose signals were not received is displayed in a visual manner.

9. A model training method, comprising: When the positioning activation conditions are met, multiple signals emitted by multiple items arranged in a preset arrangement are received by multiple antennas set on both sides of the passage for transporting multiple items, or by multiple antennas set on both sides of the passage for transporting the multiple items, and set at the bottom and / or top of the passage. The multiple items have the same shape, and the preset arrangement is a neat arrangement. The signals are emitted by signal transmitting devices set at the same position of the items. The positioning activation conditions include: the number of tags continuously read by the RFID reader exceeds a preset threshold, or the output signal of the sensor installed at the passage or passage door indicates that multiple items are passing through the side wall or passage door of the passage with antennas installed. The signal contains the identification information of the corresponding items. Based on the Signal Strength Indication (RSSI) of the signals received by the plurality of antennas, specified features are extracted from the plurality of signals, wherein the specified features include: the ratio of the number of times each antenna in the plurality of antennas receives a signal to the total number of times the plurality of antennas receive a signal; the average RSSI of the signals received by each antenna in the plurality of antennas; the maximum RSSI of the signals received by each antenna in the plurality of antennas; the minimum RSSI of the signals received by each antenna in the plurality of antennas; and the RSSI skewness value of the signals received by each antenna in the plurality of antennas. The location of the item whose signal was not received is determined based on the identification information contained in the plurality of signals, and used as the positioning result. A machine learning model is trained using the specified features and the positioning result.

10. An item positioning system, comprising a communication module, an industrial control computer, and multiple antennas, wherein: The communication module is configured to, when the positioning activation condition is met, receive multiple signals emitted by multiple items arranged in a preset layout via multiple antennas located on both sides of the passage for transporting multiple items, or via multiple antennas located on both sides of the passage for transporting the multiple items, and at the bottom and / or top of the passage, wherein the multiple items have the same shape, and the preset layout is a neat arrangement, and transmit the signals via signal transmitting devices located at the same position on the items. The positioning activation condition includes: the number of tags continuously read by the RFID reader exceeds a preset threshold, or the output signal of a sensor installed at the passage or passage door indicates that multiple items are passing through the passage side equipped with antennas. A wall or passageway door; the industrial control computer is configured to extract specified features from the multiple signals based on the Signal Strength Indication (RSSI) of the signals received by the multiple antennas, wherein the specified features include: the ratio of the number of times each antenna in the multiple antennas receives a signal to the total number of times the multiple antennas receive a signal; the average RSSI of the signals received by each antenna in the multiple antennas; the maximum RSSI of the signals received by each antenna in the multiple antennas; the minimum RSSI of the signals received by each antenna in the multiple antennas; and the RSSI skewness value of the signals received by each antenna in the multiple antennas. The specified features are input into a trained machine learning model to determine the location of the items whose signals were not received among the multiple items.

11. The system of claim 10 further includes a channel gate, wherein the plurality of antennas are disposed in the channel gate.

12. The system according to claim 11, wherein: The channel door includes a first channel door and a second channel door; the first channel door includes a channel door frame, a wave-absorbing material, and a shell; the second channel door includes a channel door frame, a wave-absorbing material, and a shell; the antenna is symmetrically arranged in the first channel door and the second channel door; the antenna is connected to a communication module, and the communication module is connected to an industrial control computer; the communication module is arranged in the first channel door and / or the second channel door; the industrial control computer is arranged in the first channel door or the second channel door.

13. The system according to claim 10, wherein, The industrial control computer is also configured to train a machine learning model through the following process: receiving multiple signals emitted by multiple items arranged in a preset arrangement by multiple antennas set on both sides of the channel for transporting the multiple items, or by multiple antennas set on both sides of the channel for transporting the multiple items, and set at the bottom and / or top of the channel, the signals containing identification information of the corresponding items; Based on the Signal Strength Indication (RSSI) of the signals received by the plurality of antennas, specified features are extracted from the plurality of signals, wherein the specified features include: the ratio of the number of times each antenna in the plurality of antennas receives a signal to the total number of times the plurality of antennas receive a signal; the average RSSI of the signals received by each antenna in the plurality of antennas; the maximum RSSI of the signals received by each antenna in the plurality of antennas; the minimum RSSI of the signals received by each antenna in the plurality of antennas; and the RSSI skewness value of the signals received by each antenna in the plurality of antennas. The location of the item whose signal was not received is determined based on the identification information contained in the plurality of signals, and used as the positioning result. A machine learning model is trained using the specified features and the positioning result.

14. An article positioning device, comprising: The first receiving module is configured to, when the positioning activation condition is met, receive multiple signals emitted by multiple items arranged in a preset arrangement via multiple antennas located on both sides of the channel for transporting multiple items, or via multiple antennas located on both sides of the channel for transporting the multiple items, and at the bottom and / or top of the channel, wherein the multiple items have the same shape, the preset arrangement is a neat arrangement, and the signals are emitted by signal transmitting devices located at the same position on the items. The positioning activation condition includes: the number of tags continuously read by the RFID reader exceeds a preset threshold, or the output signal of the sensor installed at the channel or channel door indicates that multiple items are passing through the channel sidewall or channel door with antennas installed. A first extraction module is configured to extract specified features from the plurality of signals based on the Signal Strength Indication (RSSI) of the signals received by the plurality of antennas. The specified features include: the ratio of the number of times each antenna receives a signal to the total number of times the plurality of antennas receive a signal; the average RSSI of the signals received by each antenna; the maximum RSSI of the signals received by each antenna; the minimum RSSI of the signals received by each antenna; and the RSSI skewness value of the signals received by each antenna. A first determination module is configured to input the specified features into a trained machine learning model to determine the location of items among the plurality of items whose signals were not received.

15. A model training device, comprising: The second receiving module is configured to receive multiple signals emitted by multiple items arranged in a preset arrangement via multiple antennas disposed on both sides of a channel for transporting multiple items, or via multiple antennas disposed on both sides of a channel for transporting the multiple items, and at the bottom and / or top of the channel, wherein the signals contain identification information of the corresponding items. The second extraction module is configured to extract specified features from the plurality of signals based on the Signal Strength Indication (RSSI) of the signals received by the plurality of antennas. The specified features include: the ratio of the number of times each antenna receives a signal to the total number of times the plurality of antennas receive a signal; the average RSSI of the signals received by each antenna; the maximum RSSI of the signals received by each antenna; the minimum RSSI of the signals received by each antenna; and the RSSI skewness value of the signals received by each antenna. The second determination module is configured to determine the location of an item whose signal was not received among the plurality of items based on the identification information contained in the plurality of signals, as a positioning result. The training module is configured to train a machine learning model using the specified features and the positioning result.

16. An electronic device, characterized in that, It includes a memory and a processor; wherein the memory is used to store computer instructions, wherein the computer instructions are executed by the processor to implement the steps of the method according to any one of claims 1-9.

17. A computer-readable storage medium storing computer instructions thereon, characterized in that, When executed by a processor, the computer instructions implement the steps of the method described in any one of claims 1-9.

18. A computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program containing program code for performing the method of any one of claims 1-9.

Images