An efficient inventory system and method based on RFID technology

Abstract

This application relates to the technical field of RFID inventory taking. Specifically, it discloses an efficient inventory taking system and method based on RFID technology. The RFID inventory taking area is divided into multiple logical blocks, and an RFID reader is used to scan the RFID tags in each logical block to obtain the preliminary feedback signals of each logical block. Then, a data processing technology based on deep learning is introduced to conduct global correlation analysis on the preliminary feedback signals of each logical block. Furthermore, based on this, the transmission power of the RFID reader in each logical block is dynamically adjusted for re-scanning, and an inventory taking list is generated according to the scanning results. In this way, the parameter configuration of the RFID reader can be adaptively optimized, thereby improving the reading rate and accuracy of RFID tags, reducing the phenomena of missed reading and multiple reading of tags, and further improving the accuracy and efficiency of inventory taking.

Description

Technical Field

[0001] This application relates to the technical field of RFID inventory taking, and more specifically, to an efficient inventory taking system and method based on RFID technology. Background Art

[0002] In many fields such as logistics warehousing, retail management, and asset tracking, inventory taking is a key link to ensure that the quantity of items matches the records. Traditional inventory taking methods rely on manual checking one by one, which is not only time-consuming and laborious, but also easily affected by human errors, resulting in low inventory taking efficiency and difficult to guarantee accuracy. With the rapid development of information technology, especially the wide application of radio frequency identification (RFID) technology, automated and intelligent inventory taking solutions have gradually become the goal pursued by the industry.

[0003] RFID (Radio Frequency Identification) technology communicates through radio waves, allowing non-contact automatic identification of RFID tags attached to items, greatly improving the speed and accuracy of data collection. However, although RFID technology provides the advantage of fast batch reading, it still faces many challenges in practical applications. Especially in complex and changeable environments (such as near dense shelves and metal obstacles), the propagation of RFID signals may be interfered, resulting in a decrease in the tag reading rate, that is, the so-called "tag missed reading" or "multiple reading" phenomena, which directly affect the accuracy and efficiency of inventory taking.

[0004] Therefore, an optimized efficient inventory taking system and method based on RFID technology is expected. Summary of the Invention

[0005] This application provides an efficient inventory taking system and method based on RFID technology, which can adaptively optimize the parameter configuration of RFID readers and writers, thereby improving the reading rate and accuracy of RFID tags, reducing tag missed reading and multiple reading phenomena, and further improving the accuracy and efficiency of inventory taking.

[0006] In the first aspect, an efficient inventory taking method based on RFID technology is provided, including:

[0007] Step 1: Divide the RFID inventory taking area into multiple logical blocks;

[0008] Step 2: Use an RFID reader and writer to preliminarily scan the RFID tags of the objects to be inventoried in each of the multiple logical blocks to obtain multiple logical block preliminary feedback signals, where the logical block initial feedback signal includes transmission power, the number of tags successfully read, signal strength, and reading range;

[0009] Step 3: Based on the multiple logical block preliminary feedback signals, dynamically adjust the transmission power when the RFID reader performs a second scan on the RFID tags of the objects to be inventoried in each logical block to obtain multiple logical block optimized feedback signals;

[0010] Step 4: Generate an inventory count list based on the multiple logical block optimized feedback signals.

[0011] In a second aspect, an efficient inventory system based on RFID technology is provided, including:

[0012] A logical block division module, configured to divide the RFID inventory area into multiple logical blocks;

[0013] A preliminary scan module, configured to use an RFID reader to perform a preliminary scan on the RFID tags of the objects to be inventoried in each logical block of the multiple logical blocks to obtain multiple logical block preliminary feedback signals, where the logical block initial feedback signals include transmission power, the number of successfully read tags, signal strength, and reading range;

[0014] A second scan module, configured to dynamically adjust the transmission power when the RFID reader performs a second scan on the RFID tags of the objects to be inventoried in each logical block based on the multiple logical block preliminary feedback signals to obtain multiple logical block optimized feedback signals;

[0015] An inventory count list generation module, configured to generate an inventory count list based on the multiple logical block optimized feedback signals.

[0016] An efficient inventory system based on RFID technology and its method provided in this application divide the RFID inventory area into multiple logical blocks, use an RFID reader to scan the RFID tags of each logical block to obtain the preliminary feedback signals of each logical block. Then, a data processing technology based on deep learning is introduced to perform global correlation analysis on the preliminary feedback signals of each logical block to reveal the essential propagation characteristics of the signals sent by the RFID reader in the inventory area under the current settings. Furthermore, based on this, according to the deviation degree of the preliminary feedback signals of each logical block from the essential signal propagation characteristics, dynamically adjust the transmission power of the RFID reader in each logical block for a second scan, and generate an inventory count list based on the scan results. In this way, the parameter configuration of the RFID reader can be adaptively optimized, thereby improving the reading rate and accuracy of RFID tags, reducing the phenomena of missed tag readings and multiple tag readings, and further improving the accuracy and efficiency of inventory counts. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] To more clearly illustrate the technical solutions of the embodiments of the present application, the accompanying drawings of the embodiments of the present application will be briefly introduced below. Obviously, the accompanying drawings in the following description only relate to some embodiments of the present application and do not limit the present application.

[0018] Figure 1 It is a schematic flowchart of the efficient inventory-taking method based on RFID technology in the embodiments of the present application.

[0019] Figure 2 It is a schematic flowchart of step 3 in the efficient inventory-taking method based on RFID technology in the embodiments of the present application.

[0020] Figure 3 It is a schematic diagram of data flow in step 3 of the efficient inventory-taking method based on RFID technology in the embodiments of the present application.

[0021] Figure 4 It is a schematic flowchart of step 33 in the efficient inventory-taking method based on RFID technology in the embodiments of the present application.

[0022] Figure 5 It is a schematic flowchart of step 331 in the efficient inventory-taking method based on RFID technology in the embodiments of the present application.

[0023] Figure 6 It is a schematic flowchart of step 332 in the efficient inventory-taking method based on RFID technology in the embodiments of the present application.

[0024] Figure 7 It is a schematic flowchart of step 34 in the efficient inventory-taking method based on RFID technology in the embodiments of the present application.

[0025] Figure 8 It is a schematic block diagram of the efficient inventory-taking system based on RFID technology in the embodiments of the present application. Detailed implementation manners

[0026] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts also belong to the scope of protection of the present application.

[0027] In view of the above technical problems, the technical concept of this application is as follows: The RFID inventory-taking area is divided into multiple logical blocks. The RFID reader is used to scan the RFID tags in each logical block to obtain the preliminary feedback signals of each logical block. Then, a data processing technology based on deep learning is introduced to perform global correlation analysis on the preliminary feedback signals of each logical block to reveal the essential propagation characteristics of the signals sent by the RFID reader in this inventory-taking area under the current settings. Furthermore, based on this, according to the deviation degree of the preliminary feedback signals of each logical block from the essential signal propagation characteristics, the transmission power of the RFID reader in each logical block is dynamically adjusted for re-scanning, and an inventory list is generated based on the scanning results. In this way, the parameter configuration of the RFID reader can be adaptively optimized, thereby improving the reading rate and accuracy of RFID tags, reducing the phenomena of missed reading and multiple reading of tags, and further improving the accuracy and efficiency of inventory-taking.

[0028] Based on this, in the technical solution of this application, as Figure 1 shown, the efficient inventory-taking method based on RFID technology includes: Step 1: Divide the RFID inventory-taking area into multiple logical blocks; Step 2: Use the RFID reader to preliminarily scan the RFID tags of the objects to be inventoried in each logical block of the multiple logical blocks to obtain multiple preliminary feedback signals of the logical blocks, where the initial feedback signal of the logical block includes transmission power, the number of successfully read tags, signal strength, and reading range; Step 3: Based on the multiple preliminary feedback signals of the logical blocks, dynamically adjust the transmission power when the RFID reader re-scans the RFID tags of the objects to be inventoried in each logical block to obtain multiple optimized feedback signals of the logical blocks; Step 4: Generate an inventory list based on the multiple optimized feedback signals of the logical blocks.

[0029] Exemplarily, in Step 1, the RFID inventory-taking area is divided into multiple logical blocks. It should be understood that in large-scale RFID inventory-taking scenarios, such as large warehouses, logistics centers, or industrial production workshops, the objects to be inventoried are often distributed in a vast space, and their layouts and storage methods are diverse, and different logical blocks may have different environmental characteristics (such as item stacking density, spatial obstacle distribution, electromagnetic interference degree, etc.). Therefore, by dividing the inventory-taking area into multiple logical blocks (such as dividing by shelves, by categories of stored items, by physical partitions of space, etc.) in this application, it helps to decompose complex inventory-taking tasks into multiple relatively independent sub-tasks, more effectively adjust inventory-taking strategies according to the characteristics of each area, perform refined management and optimization operations on the inventory-taking of different areas to meet different inventory-taking requirements, thereby improving the operability, scalability of inventory-taking, and the ability to specifically solve local problems.

[0030] In a specific embodiment, when performing logical block division, it is first necessary to comprehensively evaluate the entire inventory area to identify which areas may cause RFID signal attenuation or reflection due to the above reasons, thereby affecting tag reading. Based on this evaluation, each logical block can be defined according to the location and number of the shelves, the types and properties of the stored items (for example, metal products may shield the RFID signal), and physical partitions (such as walls or other fixed structures). This not only helps to break down complex inventory tasks into smaller and more manageable parts, but also ensures that the environmental conditions within each logical block are relatively consistent, thereby optimizing the operating parameters of the RFID reader, such as transmission power and frequency selection, to meet the requirements of a specific area. In addition, considering the flexibility in actual operation, the size of the logical block can also be flexibly adjusted according to the specific requirements of the inventory task. For areas with dense tags or prone to reading problems, smaller logical blocks can be set to manage and adjust more meticulously; while for areas with sparse tag distribution and good reading conditions, the scope of the logical block can be appropriately expanded to reduce the number of unnecessary partitions and simplify the inventory process. At the same time, as the inventory work progresses, the division of the logical block can also be dynamically adjusted according to the results of the preliminary scan. That is, if it is found that there are many tags that have not been correctly read in a certain area, this area can be further subdivided in subsequent rounds; conversely, adjacent small blocks can be merged to form a larger block to achieve the best inventory effect.

[0031] Exemplarily, in the step 2, an RFID reader is used to preliminarily scan the RFID tags of the objects to be inventoried in each of the multiple logical blocks to obtain multiple preliminary feedback signals of the logical blocks. Among them, the initial feedback signal of the logical block includes transmission power, the number of successfully read tags, signal strength, and reading range. It should be understood that the RFID reader is used to transmit radio frequency signals into each logical block for preliminary scanning. When the signal covers the objects to be inventoried with RFID tags in the logical block, the tags will reflect or respond to the signal. The RFID reader receives these reflected or responded signals and analyzes them to extract the transmission power, the number of successfully read tags, signal strength, and reading range information of the logical block, so as to generate the preliminary feedback signals of each logical block. Among them, the transmission power refers to the power of the RFID reader when transmitting signals, which determines the propagation distance and coverage range of the RFID signal; the number of successfully read tags reflects the total number of RFID tags successfully read in the logical block and is an important indicator for evaluating the performance of the RFID reader in this area; the signal strength represents the strength of the reflected or responded signal received by the RFID reader, which is affected by various factors such as distance, obstacles, and environmental noise; the reading range refers to the maximum distance or area within the logical block where the RFID reader can effectively read RFID tags, representing the effective spatial range covered by the current signal. These information together constitute a comprehensive description of the performance of the RFID inventory process in each logical block, providing a basis for subsequent data processing and parameter optimization.

[0032] Exemplarily, in the step 3, based on the multiple preliminary feedback signals of the logical blocks, the transmission power of the RFID reader when re-scanning the RFID tags of the objects to be inventoried in each of the logical blocks is dynamically adjusted to obtain multiple optimized feedback signals of the logical blocks. It should be understood that considering the differences in the environment of different logical blocks and the objects to be inventoried, a single transmission power may not meet the inventory requirements of all areas. Therefore, the present application further analyzes the data of the multiple preliminary feedback signals of the logical blocks to identify the signal propagation characteristics and tag reading effects of each logical block, and then dynamically adjusts the transmission power of the RFID reader in this logical block to adapt to its unique environment and item distribution, improving the accuracy and efficiency of the inventory. For example, in areas with weak signals or difficult tag reading, the transmission power is appropriately increased to enhance signal coverage and reading ability; in areas with strong signals and good tag reading effects, the transmission power can be appropriately reduced to reduce electromagnetic interference and energy consumption, thus avoiding energy waste and signal interference caused by over-strong signals in some areas due to a unified transmission power, or incomplete tag reading caused by over-weak signals in some areas.

[0033] In one embodiment, as Figure 2and Figure 3 As shown in Figure 3 , step 3 includes: Step 31: Structurally encode each of the preliminary feedback signals of the multiple logical blocks to obtain multiple embedded encoding vectors of the preliminary feedback signals of the logical blocks; Step 32: Extract the embedded encoding vector of the preliminary feedback signal corresponding to the first logical block from the multiple embedded encoding vectors of the preliminary feedback signals of the logical blocks; Step 33: Input the multiple embedded encoding vectors of the preliminary feedback signals of the logical blocks into a feedback signal eigenfeature extractor to obtain a feedback signal eigenfeature encoding vector; Step 34: Generate a transmission power dynamic adjustment value based on the feature offset between the embedded encoding vector of the preliminary feedback signal of the first logical block and the feedback signal eigenfeature encoding vector; Step 35: Adjust the transmission power of the RFID reader when scanning the first logical block based on the transmission power dynamic adjustment value.

[0034] Exemplarily, in step 31, each of the preliminary feedback signals of the multiple logical blocks is structurally encoded to obtain multiple embedded encoding vectors of the preliminary feedback signals of the logical blocks. It should be understood that considering that the preliminary feedback signals of the logical blocks contain various types of data, such as transmission power, the number of successfully read tags, signal strength, and reading range, which have different dimensions and value ranges and are difficult to directly use for the fusion analysis between multi-dimensional data. Therefore, in this application, a structural encoding method is adopted, and each of the preliminary feedback signals of the logical blocks is input into a multi-layer perceptron model to utilize the powerful non-linear mapping ability of the multi-layer perceptron model. Through the neurons and activation functions in the hidden layer, information is transmitted and integrated layer by layer, so that the multi-dimensional information contained in each of the preliminary feedback signals of the logical blocks is mapped to a low-dimensional embedding space and encoded into a common embedding vector to obtain multiple embedded encoding vectors of the preliminary feedback signals of the logical blocks. In this way, the correlation learning between multi-dimensional data in the original feedback signal can be realized, so that each embedded encoding vector of the preliminary feedback signal of the logical block can accurately reflect the RFID signal propagation characteristics and tag reading effect of the corresponding logical block, providing a unified and comparable data basis for subsequent parameter optimization and configuration.

[0035] In one embodiment, structurally encoding each of the preliminary feedback signals of the multiple logical blocks to obtain multiple embedded encoding vectors of the preliminary feedback signals of the logical blocks includes: Inputting each of the preliminary feedback signals of the multiple logical blocks into an embedding encoder based on a multi-layer perceptron to obtain the multiple embedded encoding vectors of the preliminary feedback signals of the logical blocks.

[0036] Exemplarily, in the step 32, the first logical block corresponding first logical block preliminary feedback signal embedded coding vector is extracted from the multiple logical block preliminary feedback signal embedded coding vectors. It should be understood that for each divided logical block, the corresponding logical block preliminary feedback signal embedded coding vector thereof is extracted as the basic data for subsequent targeted analysis and adjustment. Here, taking the first logical block as an example, by extracting the first logical block corresponding first logical block preliminary feedback signal embedded coding vector, the transmission power of this logical block can be specifically analyzed and optimized.

[0037] Exemplarily, in the step 33, the multiple logical block preliminary feedback signal embedded coding vectors are input into a feedback signal eigenfeature extractor to obtain feedback signal eigenfeature coding vectors. It should be understood that considering the layout, environment of the current inventory area and the distribution of the objects to be inventoried, the signals sent by the RFID reader / writer will be affected by various factors, such as the obstruction of spatial obstacles, the interference of the electromagnetic environment, and the stacking density of items, etc., resulting in significant differences in the propagation characteristics of RFID signals among different logical blocks. Therefore, in order to effectively compensate for this difference, the present application further performs signal eigenfeature extraction on the multiple logical block preliminary feedback signal embedded coding vectors to extract more essential and representative feature information, revealing the inherent propagation and attenuation law of the signals sent by the RFID reader / writer in the overall area, so as to provide a stable reference benchmark for the transmission power adjustment of each local logical block.

[0038] In one embodiment, as Figure 4 shown, in the step 33, inputting the multiple logical block preliminary feedback signal embedded coding vectors into a feedback signal eigenfeature extractor to obtain feedback signal eigenfeature coding vectors includes: Step 331: performing feature complementary analysis based on the hub feature on the multiple logical block preliminary feedback signal embedded coding vectors to obtain multiple logical block preliminary feedback signal - hub feature complementary information embedded coding vectors; Step 332: performing significant dynamic modulation aggregation coding on the multiple logical block preliminary feedback signal - hub feature complementary information embedded coding vectors to obtain the feedback signal eigenfeature coding vectors.

[0039] In one embodiment, as Figure 5 shown, in the step 331, performing feature complementary analysis based on the hub feature on the multiple logical block preliminary feedback signal embedded coding vectors to obtain multiple logical block preliminary feedback signal - hub feature complementary information embedded coding vectors includes: Step 3311: inputting the multiple logical block preliminary feedback signal embedded coding vectors into a hub feature extraction network to obtain logical block preliminary feedback signal hub feature coding vectors. Specifically, this process can be expressed by the formula as:

[0040]

[0041] Among them, represents the hub feature extraction network, represents the embedded coding vectors of the preliminary feedback signals of the multiple logical blocks, , , and respectively represent the first, second, th, and th embedded coding vectors of the preliminary feedback signals of the multiple logical blocks, is the number of the embedded coding vectors of the preliminary feedback signals of the logical blocks, and respectively represent the weight parameter matrix and the bias term of the hub feature extraction network, represents the hub feature correlation score conversion vector of the hub feature extraction network, represents the corresponding hub feature correlation score factor, represents matrix multiplication, represents the normalized exponential function, represents the corresponding normalized hub feature correlation score factor, represents the hub feature coding vector of the preliminary feedback signal of the logical block.

[0042] That is, the present application first constructs a hub feature extraction network based on a neural network architecture to utilize the powerful feature extraction ability of the neural network to perform hub feature correlation analysis on the embedded coding vectors of the preliminary feedback signals of the multiple logical blocks, so as to learn the potential associations between the features of the preliminary feedback signals of each logical block, thereby identifying and extracting the representative feature representations in the embedded coding vectors of the preliminary feedback signals of the multiple logical blocks, that is, the hub feature coding vector of the preliminary feedback signal of the logical block, as the key descriptor for inventorying the overall propagation characteristics of the RFID signals in the area.

[0043] Step 3312: Extract the complementary information of each embedded coding vector of the preliminary feedback signals of the multiple logical blocks relative to the hub feature coding vector of the preliminary feedback signal of the logical block to obtain the multiple logical block-hub feature complementary information embedded coding vectors. Specifically, this process can be expressed by the formula:

[0044]

[0045] Among them, denotes the Sigmoid activation function, denotes the embedding encoding vector of the preliminary feedback signal of the normalization logic block, denotes the hub feature encoding vector of the preliminary feedback signal of the normalization logic block, and denotes different weight matrices, denotes point convolution encoding, denotes the hub feature differential feature encoding vector of the preliminary feedback signal of the logic block, denotes differential operation, is to take the absolute value, denotes the corresponding embedding encoding vector of the complementary information of the preliminary feedback signal - hub feature of the logic block.

[0046] That is, in order to avoid losing too much detailed information of the local logic block during the hub feature extraction process, the present application further takes the hub feature encoding vector of the preliminary feedback signal of the logic block as the core to integrate and associate the unique features of the preliminary feedback signals of each logic block, and introduces a dynamic compensation mechanism to compensate and repair the lost detailed information of the logic block, so as to achieve a comprehensive coverage and accurate description of the signal propagation characteristics of each logic block. In this process, first, by calculating the complementary information of each embedding encoding vector of the preliminary feedback signal of the logic block relative to the hub feature encoding vector of the preliminary feedback signal of the logic block, to capture the feature differences and unique contributions of the signal propagation characteristics of each logic block relative to the global core information of the signal propagation characteristics of the inventory area, and generate multiple embedding encoding vectors of the complementary information of the preliminary feedback signal - hub feature of the logic block, so as to restore the feature components that are not considered and absorbed in the hub feature extraction process of each embedding encoding vector of the preliminary feedback signal of the logic block.

[0047] In one embodiment, as Figure 6 shown, in the step 332, performing significant dynamic modulation aggregation encoding on the multiple embedding encoding vectors of the complementary information of the preliminary feedback signal - hub feature of the logic block to obtain the eigen - feature encoding vector of the feedback signal, includes: Step 3321: Inputting each embedding encoding vector of the complementary information of the preliminary feedback signal - hub feature of the logic block in the multiple embedding encoding vectors of the complementary information of the preliminary feedback signal - hub feature of the logic block into the complementary information significant identification module based on the attention mechanism to obtain multiple attention weights of the complementary information of the preliminary feedback signal of the logic block. Specifically, this process can be expressed by the formula as:

[0048]

[0049]

[0050] Among them, represents the corresponding significance identification factor, and respectively represent the weight parameter matrix and bias term of the complementary information significant identification module, represents the complementary information significance scoring conversion vector of the complementary information significant identification module, represents the exponential function operation with base e, represents the complementary information attention weight of the corresponding preliminary feedback signal of the logical block.

[0051] That is, since the complementary information of the preliminary feedback signal features of each logical block relative to the global core information of the inventory area signal propagation may be a beneficial supplement to it, or may be redundant or interfering. Therefore, the present application further uses the attention mechanism to evaluate the importance of the complementary information embedding coding vectors of the preliminary feedback signals of each logical block - hub feature, so as to dynamically adjust the weight distribution of the complementary information embedding coding vectors of the preliminary feedback signals of each logical block - hub feature in the information aggregation process.

[0052] Step 3322: Based on the complementary information attention weights of the multiple preliminary feedback signals of the logical blocks, perform attention modulation on the complementary information embedding coding vectors of the multiple preliminary feedback signals of the logical blocks - hub feature to obtain multiple significantly modulated complementary information embedding coding vectors of the preliminary feedback signals of the logical blocks - hub feature. Specifically, this process can be expressed by the formula:

[0053]

[0054] Among them, , , and respectively represent , , and the corresponding complementary information embedding coding vectors of the preliminary feedback signals of the logical blocks - hub feature, , , and respectively represent , , and the corresponding complementary information attention weights of the preliminary feedback signals of the logical blocks, represents multiple significantly modulated complementary information embedding coding vectors of the preliminary feedback signals of the logical blocks - hub feature.

[0055] That is, based on the complementary information attention weights of the preliminary feedback signals of multiple logical blocks calculated, the preliminary feedback signals of multiple logical blocks - the complementary information embedding coding vectors of hub features are subjected to attention modulation. Through the enhancement or suppression of features, the modulated preliminary feedback signals of multiple logical blocks - the complementary information embedding coding vectors of hub features can better reflect the key roles and contributions of the characteristics of each preliminary feedback signal of the logical block in the global core feature description of signal propagation in the inventory area.

[0056] Step 3323: Fuse the hub feature coding vector of the preliminary feedback signal of the logical block and the complementary information embedding coding vectors of the preliminary feedback signals of multiple significantly modulated logical blocks - hub features to obtain the eigen - feature coding vector of the feedback signal. Specifically, this process can be expressed by the formula:

[0057]

[0058] Wherein, represents the concatenation function, represents the eigen - feature coding vector of the feedback signal.

[0059] That is, through the concatenation operation, the complementary information embedding coding vectors of the preliminary feedback signals of multiple significantly modulated logical blocks - hub features and the hub feature coding vector of the preliminary feedback signal of the logical block are fused to comprehensively consider the global core information of the signal propagation characteristics in the RFID inventory area and the local detailed characteristics of the signal propagation characteristics of each logical block, improve the accuracy and comprehensiveness of feature expression, so as to achieve accurate description and modeling of the essential signal propagation characteristics in the RFID inventory area, and ensure that it can be used as a reliable reference basis in the subsequent process of optimizing the configuration of the transmission power.

[0060] In one embodiment, as Figure 7 shown, in step 34, based on the feature offset between the embedded coding vector of the preliminary feedback signal of the first logical block and the eigen - feature coding vector of the feedback signal, a transmission power dynamic adjustment value is generated, including: Step 341: Perform feature offset measurement on the preliminary feedback signal of the first logical block and the eigen - feature coding vector of the feedback signal to obtain the semantic offset coding vector of the feedback signal of the first logical block; Step 342: Input the semantic offset coding vector of the feedback signal of the first logical block into the transmission power dynamic adjuster based on the decoder to obtain the transmission power dynamic adjustment value.

[0061] Exemplarily, in the step 341, a feature offset metric is performed on the first logical block preliminary feedback signal and the feedback signal eigen - feature encoding vector to obtain a first logical block feedback signal semantic offset encoding vector. In one embodiment, performing a feature offset metric on the first logical block preliminary feedback signal and the feedback signal eigen - feature encoding vector to obtain a first logical block feedback signal semantic offset encoding vector includes: calculating a position - by - position difference vector between the first logical block preliminary feedback signal and the feedback signal eigen - feature encoding vector to obtain the first logical block feedback signal semantic offset encoding vector. It should be understood that after the above - mentioned processing, the feedback signal eigen - feature encoding vector represents the eigen - characteristics of the signal propagation in the RFID inventory area. Therefore, in order to quantitatively evaluate the signal performance difference of the first logical block relative to the overall inventory area, the present application further calculates the position - by - position difference vector between the first logical block preliminary feedback signal and the feedback signal eigen - feature encoding vector to reveal the offset direction and the degree of difference in signal propagation characteristics between the first logical block and the overall inventory area, and obtains the first logical block feedback signal semantic offset encoding vector. It should be understood that each element of the first logical block feedback signal semantic offset encoding vector represents the offset amount of the first logical block at the corresponding position relative to the eigen - characteristics of the overall signal propagation in the inventory area. A positive value indicates that the signal propagation effect is better than the overall level, and a negative value indicates that the signal propagation effect is worse than the overall level. Therefore, based on the first logical block feedback signal semantic offset encoding vector, the signal performance of this logical block during the RFID inventory process can be intuitively understood, providing targeted guidance for subsequent transmit power adjustment. For example, a logical block located in a dense area of metal shelves may have a much lower signal strength than logical blocks in other open areas, and its offset degree is large. Then, the transmit power can be adjusted according to its offset degree, by significantly increasing the transmit power to overcome the shielding effect of the metal.

[0062] Exemplarily, in the step 342, the first logical block feedback signal semantic offset encoding vector is input into a decoder - based transmit power dynamic adjuster to obtain the transmit power dynamic adjustment value. Specifically, the decoder is used to parse the offset information in the first logical block feedback signal semantic offset encoding vector to identify the strong - weak distribution of the signal propagation of this logical block during the RFID inventory process and its difference from the eigen - characteristics of the signal propagation in the inventory area, and based on the mapping relationship between the transmit power and the offset information learned during the training process, convert the parsed offset information into a transmit power dynamic adjustment value for the first logical block.

[0063] Here, since the first logical block preliminary feedback signal and the feedback signal eigenfeature encoding vector respectively represent the structured encoding features of the first logical block preliminary feedback signal and the full-sample domain distribution eigenfeatures of the preliminary feedback signals of all logical blocks, when calculating the position-wise difference vector between the first logical block preliminary feedback signal and the feedback signal eigenfeature encoding vector to obtain the first logical block feedback signal semantic offset encoding vector, the population attributes of the feedback signal semantic features corresponding to different sample numbers will have fine-grained differential interaction fairness differences, thereby affecting the feature distribution interaction inclusiveness of the first logical block feedback signal semantic offset encoding vector, and causing the feature distribution of the first logical block feedback signal semantic offset encoding vector to have a feature manifold edge strengthening architecture, reducing the accuracy of the decoding result obtained through the decoder-based transmit power adjuster.

[0064] In a preferred example, in the process of obtaining the transmit power dynamic adjustment value by passing the first logical block feedback signal semantic offset encoding vector through the decoder-based transmit power adjuster, feature manifold edge architecture modulation is performed on the first logical block feedback signal semantic offset encoding vector. The process includes:

[0065] Using a random permutation operator to perform dynamic topological recombination on the first logical block feedback signal semantic offset encoding vector to generate a logical block feedback signal semantic offset feature dynamic recombination encoding tensor, denoted as , where represents the first logical block feedback signal semantic offset encoding vector, is the random permutation operator, represents matrix multiplication, represents the logical block feedback signal semantic offset feature dynamic recombination encoding tensor;

[0066] Constructing a complete linear field encoding tensor through the outer product tensor generation operation of the first logical block feedback signal semantic offset encoding vector, denoted as: , where represents the transposed vector of the first logical block feedback signal semantic offset encoding vector, represents the complete linear field encoding tensor;

[0067] Performing field projection on the logical block feedback signal semantic offset feature dynamic recombination encoding tensor and the complete linear field encoding tensor to generate a logical block feedback signal semantic offset feature scalar correlation distribution heterogeneous encoding tensor, denoted as: , where represents the logical block feedback signal semantic offset feature scalar correlation distribution heterogeneous encoding tensor;

[0068] Perform topological slice analysis on the conjugate tensor of the semantic offset feature scalar correlation distribution heterogeneous coding tensor of the logical block feedback signal and the semantic offset coding vector of the first logical block feedback signal to obtain the semantic offset feature heterogeneous topological tensor of the logical block feedback signal, expressed as: , where represents truncated decomposition that retains the first singular values, represents the semantic offset feature heterogeneous topological tensor of the logical block feedback signal;

[0069] Perform field fusion on the semantic offset coding vector of the first logical block feedback signal and the semantic offset feature heterogeneous topological tensor of the logical block feedback signal to output the optimized semantic offset coding vector of the first logical block feedback signal.

[0070] Here, by performing dynamic topological recombination on the semantic offset coding vector of the first logical block feedback signal through a random permutation operator and embedding it into a complete linear field generated by an outer product tensor, the scalar correlation distribution of the feature set can be analytically measured in a heterogeneous topological structure. Based on this, a heterogeneous topological tensor is constructed using conjugate tensor transformation, and in combination with topological slice analysis, the dimensionality-reduced manifold of the feature space is reconstructed within the field to avoid the compression effect of the edge enhancement architecture on the feature representation probability, thereby improving the accuracy of the transmit power dynamic adjustment value obtained by it through a decoder-based transmit power dynamic adjuster.

[0071] Exemplarily, in step 35, based on the transmit power dynamic adjustment value, adjust the transmit power of the RFID reader when scanning the first logical block. That is, based on the transmit power dynamic adjustment value, adjust the transmit power of the RFID reader when scanning the first logical block, and use the adjusted transmit power to rescan the first logical block to obtain an optimized feedback signal for this logical block. By performing the above steps for each logical block, optimized feedback signals for each logical block can be obtained in sequence.

[0072] Exemplarily, in the step 4, based on the optimized feedback signals of the multiple logical blocks, an inventory count list is generated. It should be understood that by analyzing the optimized feedback signals of each logical block and extracting the item-related information therein, that is, the information stored in the RFID tags of the items, the detailed information of the items, such as item numbers, names, specifications, etc., can be determined, and the item information of all logical blocks is summarized into a unified table or database to generate the final inventory count list. In this way, the actual situation of the current inventory can be accurately and clearly reflected, helping enterprises quickly perform operations such as inventory verification, replenishment decision-making, and logistics arrangement, improving the efficiency of enterprise operations and the scientific nature of decision-making. At the same time, data is provided for the subsequent inventory count history records, facilitating long-term tracking and analysis of inventory changes.

[0073] In summary, the efficient inventory counting method based on RFID technology according to the embodiments of the present application is elucidated. It divides the RFID inventory counting area into multiple logical blocks, uses an RFID reader to scan the RFID tags of each logical block to obtain the preliminary feedback signals of each logical block, and then introduces a data processing technology based on deep learning to perform global correlation analysis on the preliminary feedback signals of each logical block to reveal the essential propagation characteristics of the signals sent by the RFID reader in this inventory counting area under the current settings. Furthermore, based on this, according to the deviation degree of the preliminary feedback signals of each logical block from the essential signal propagation characteristics, the transmission power of the RFID reader in each logical block is dynamically adjusted for re-scanning, and an inventory count list is generated according to the scanning results. In this way, the parameter configuration of the RFID reader can be adaptively optimized, thereby improving the reading rate and accuracy of RFID tags, reducing the phenomenon of missed reading and multiple reading of tags, and further improving the accuracy and efficiency of inventory counting.

[0074] Figure 8 is a schematic block diagram of the efficient inventory counting system based on RFID technology according to the embodiments of the present application. As Figure 8As shown, the efficient inventory counting system 100 based on RFID technology includes: a logical block division module 110 for dividing the RFID inventory counting area into multiple logical blocks; a preliminary scanning module 120 for using an RFID reader / writer to preliminarily scan the RFID tags of the objects to be inventoried in each of the multiple logical blocks to obtain multiple logical block preliminary feedback signals, where the logical block initial feedback signals include transmission power, the number of successfully read tags, signal strength, and reading range; a re-scanning module 130 for dynamically adjusting the transmission power when the RFID reader / writer re-scans the RFID tags of the objects to be inventoried in each of the logical blocks based on the multiple logical block preliminary feedback signals to obtain multiple logical block optimized feedback signals; and an inventory counting list generation module 140 for generating an inventory counting list based on the multiple logical block optimized feedback signals.

[0075] In one embodiment, the re-scanning module includes: a preliminary feedback signal embedding and encoding unit for performing structured encoding on each of the multiple logical block preliminary feedback signals to obtain multiple logical block preliminary feedback signal embedding and encoding vectors; a first logical block vector extraction unit for extracting the first logical block preliminary feedback signal embedding and encoding vector corresponding to the first logical block from the multiple logical block preliminary feedback signal embedding and encoding vectors; a feedback signal eigenfeature encoding unit for inputting the multiple logical block preliminary feedback signal embedding and encoding vectors into a feedback signal eigenfeature extractor to obtain a feedback signal eigenfeature encoding vector; a transmission power dynamic adjustment value generation unit for generating a transmission power dynamic adjustment value based on the feature offset between the first logical block preliminary feedback signal embedding and encoding vector and the feedback signal eigenfeature encoding vector; and a transmission power adjustment unit for adjusting the transmission power of the RFID reader / writer when scanning the first logical block based on the transmission power dynamic adjustment value.

[0076] Here, those skilled in the art can understand that the specific operations of each module and unit in the above efficient inventory counting system based on RFID technology have been introduced in detail in the description of the Figures 1 to 7 efficient inventory counting method based on RFID technology above, and therefore, its repeated description will be omitted.

[0077] The embodiment of the present application also provides a computer program product, which includes computer program code. When the computer program code runs on a computer, it enables the computer to implement the methods in the above various embodiments of the present application.

[0078] The embodiments of the present application further provide a computer-readable storage medium storing computer instructions, which, when running on a computer, cause the computer to implement the methods in the above embodiments of the present application.

[0079] The embodiments of the present application further provide a chip, including a circuit for executing the methods in the above embodiments of the present application.

[0080] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0081] In the description of the embodiments of the present application, unless otherwise specified, " / " means "or". For example, A / B may represent A or B; herein, "and / or" is an association relationship describing associated objects, indicating that there can be three relationships. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. In the present application, "at least one" means one or more, and "a plurality" means two or more. "At least one (item)" or its similar expression refers to any combination of these items, including any combination of single (item) or plural items. For example, at least one (item) of a, b, or c may represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, c may be single or multiple.

[0082] In the embodiments of the present application, prefix words such as "first" and "second" are only used to distinguish different described objects, and have no limiting effect on the position, order, priority, quantity, or content of the described objects, etc. The use of ordinal numbers and other prefix words for distinguishing described objects in the embodiments of the present application does not constitute a limitation on the described objects. The description of the described objects refers to the description in the context of the claims or embodiments, and should not constitute an unnecessary limitation due to the use of such prefix words.

[0083] In several embodiments provided by the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there can be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be in electrical, mechanical, or other forms.

[0084] In various embodiments of the present application, without special instructions and logical conflicts, the terms and / or descriptions among the various embodiments are consistent and can be cross-referenced. The technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationships.

[0085] The unit described as a separate component may or may not be physically separated. The component shown as a unit may or may not be a physical unit, that is, it may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0086] In addition, in each embodiment of the present application, each functional unit may be integrated into one processing unit, may exist separately as individual physical units, or two or more units may be integrated into one unit.

[0087] As mentioned above, the above are only specific embodiments of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. An efficient inventory counting method based on RFID technology, characterized in that: include: Step 1: Divide the RFID inventory area into multiple logical blocks; Step 2: Using an RFID reader / writer to preliminarily scan the RFID tags of the objects to be counted in each of the multiple logical blocks to obtain multiple logical block preliminary feedback signals, wherein the logical block preliminary feedback signals include transmission power, number of tags successfully read, signal strength and reading range; Step 3: Based on the preliminary feedback signals of the multiple logical blocks, dynamically adjust the transmission power of the RFID reader when rescanning the RFID tags of the objects to be counted in the respective logical blocks to obtain multiple optimized feedback signals of the logical blocks; Step 4: Generate an inventory count list based on the optimization feedback signals of the multiple logic blocks; The step 3 comprises: Performing structured coding on each of the plurality of logic block preliminary feedback signals to obtain a plurality of logic block preliminary feedback signal embedding coding vectors; Extracting a first logic block preliminary feedback signal embedding code vector corresponding to a first logic block from the plurality of logic block preliminary feedback signal embedding code vectors; Embedding the preliminary feedback signals of the plurality of logic blocks into a coding vector and inputting it into a feedback signal intrinsic feature extractor to obtain a feedback signal intrinsic feature coding vector; Generate a dynamic adjustment value of transmit power based on a feature offset between a preliminary feedback signal embedding coding vector of the first logic block and an intrinsic feature coding vector of the feedback signal; Based on the dynamic adjustment value of the transmission power, the transmission power of the RFID reader when scanning the first logic block is adjusted.

2. The efficient inventory counting method based on RFID technology according to claim 1 is characterized in that: Performing structured coding on each of the plurality of logic block preliminary feedback signals to obtain a plurality of logic block preliminary feedback signal embedding coding vectors, including: Each logic block preliminary feedback signal of the plurality of logic block preliminary feedback signals is input into an embedding encoder based on a multi-layer perceptron to obtain an embedding coding vector of the plurality of logic block preliminary feedback signals.

3. The efficient inventory counting method based on RFID technology according to claim 2 is characterized in that: Embedding the preliminary feedback signals of the plurality of logic blocks into a coding vector and inputting the coding vector into a feedback signal intrinsic feature extractor to obtain a feedback signal intrinsic feature coding vector, comprising: Performing feature complementation analysis based on the hub feature on the plurality of logic block preliminary feedback signal embedded coding vectors to obtain a plurality of logic block preliminary feedback signal-hub feature complementary information embedded coding vectors; The preliminary feedback signals of the multiple logic blocks-hub feature complementary information embedded coding vectors are subjected to significant dynamic modulation aggregation coding to obtain the feedback signal intrinsic feature coding vector.

4. The efficient inventory counting method based on RFID technology according to claim 3 is characterized in that: Performing feature complementation analysis based on the hub feature on the plurality of logic block preliminary feedback signal embedded coding vectors to obtain a plurality of logic block preliminary feedback signal-hub feature complementary information embedded coding vectors, including: Embedding the plurality of logic block preliminary feedback signals into encoding vectors and inputting them into a pivot feature extraction network to obtain a logic block preliminary feedback signal pivot feature encoding vector; The complementary information of each logic block preliminary feedback signal embedded coding vector relative to the logic block preliminary feedback signal hub feature coding vector is extracted to obtain the multiple logic block preliminary feedback signal-hub feature complementary information embedded coding vectors.

5. The efficient inventory counting method based on RFID technology according to claim 4 is characterized in that: The preliminary feedback signals of the plurality of logic blocks and the hub feature complementary information embedding coding vectors are subjected to significant dynamic modulation aggregation coding to obtain the feedback signal intrinsic feature coding vector, including: Inputting each logic block preliminary feedback signal-hub feature complementary information embedding coding vector in the plurality of logic block preliminary feedback signal-hub feature complementary information embedding coding vectors into a complementary information significant identification module based on an attention mechanism to obtain a plurality of logic block preliminary feedback signal complementary information attention weights; Based on the attention weights of the complementary information of the preliminary feedback signals of the multiple logic blocks, the preliminary feedback signals of the multiple logic blocks-hub feature complementary information embedded coding vectors are subjected to attention modulation to obtain multiple significant modulated preliminary feedback signals of the logic blocks-hub feature complementary information embedded coding vectors; The feedback signal intrinsic feature coding vector is obtained by fusing the logic block preliminary feedback signal hub feature coding vector and the plurality of significantly modulated logic block preliminary feedback signal-hub feature complementary information embedding coding vectors.

6. The efficient inventory counting method based on RFID technology according to claim 5 is characterized in that: Generating a dynamic adjustment value of transmit power based on a feature offset between a preliminary feedback signal embedding coding vector of the first logic block and an intrinsic feature coding vector of the feedback signal, including: Performing feature offset measurement on the first logic block preliminary feedback signal and the feedback signal intrinsic feature coding vector to obtain a first logic block feedback signal semantic offset coding vector; The first logic block feedback signal semantic offset coding vector is input into a decoder-based transmit power dynamic adjuster to obtain the transmit power dynamic adjustment value.

7. The efficient inventory counting method based on RFID technology according to claim 6 is characterized in that: Performing feature offset measurement on the first logic block preliminary feedback signal and the feedback signal intrinsic feature coding vector to obtain a first logic block feedback signal semantic offset coding vector, including: A position difference vector between the first logic block preliminary feedback signal and the feedback signal intrinsic feature encoding vector is calculated to obtain the first logic block feedback signal semantic offset encoding vector.

8. An efficient inventory counting system based on RFID technology, characterized in that: include: A logical block division module is used to divide the RFID inventory area into multiple logical blocks; A preliminary scanning module, used to use an RFID reader to perform preliminary scanning on the RFID tags of the objects to be counted in each of the multiple logical blocks to obtain multiple logical block preliminary feedback signals, wherein the logical block preliminary feedback signals include transmission power, number of tags successfully read, signal strength and reading range; A re-scanning module, configured to dynamically adjust the transmission power of the RFID reader when re-scanning the RFID tags of the objects to be counted in the respective logical blocks based on the preliminary feedback signals of the plurality of logical blocks, so as to obtain the optimized feedback signals of the plurality of logical blocks; An inventory count list generating module, configured to generate an inventory count list based on the optimization feedback signals of the plurality of logic blocks; The re-scanning module comprises: A preliminary feedback signal embedding coding unit, configured to perform structured coding on each of the plurality of logic block preliminary feedback signals to obtain a plurality of logic block preliminary feedback signal embedding coding vectors; A first logic block vector extraction unit is used to extract a first logic block preliminary feedback signal embedding code vector corresponding to a first logic block from the plurality of logic block preliminary feedback signal embedding code vectors; A feedback signal intrinsic feature encoding unit, used for embedding the preliminary feedback signals of the plurality of logic blocks into a coding vector and inputting it into a feedback signal intrinsic feature extractor to obtain a feedback signal intrinsic feature coding vector; A transmit power dynamic adjustment value generating unit, configured to generate a transmit power dynamic adjustment value based on a feature offset between an embedded coding vector of a preliminary feedback signal of the first logic block and an intrinsic feature coding vector of the feedback signal; The transmission power adjustment unit is used to adjust the transmission power of the RFID reader when scanning the first logic block based on the transmission power dynamic adjustment value.

Images