A Recommendation Method for Cross-Device Adaptive RFID Scanning and Identification

Abstract

The present invention discloses a recommendation method for cross-device adaptive RFID scanning and identification, which is used in the field of radio frequency identification technology. The method includes: constructing and utilizing an RFID scanning device characteristic model to analyze the RFID scanning device and obtain a characteristic identification result of the RFID scanning device; based on the characteristic identification result of the RFID scanning device and combined with a parameter adjustment strategy, obtaining the optimal scanning parameters of the RFID scanning device; using the optimal scanning parameters of the RFID scanning device to scan the RFID tag and obtaining the scanning effect of the RFID scanning device; constructing an intelligent RFID scanning and identification recommendation model according to the characteristics and scanning effect of the RFID scanning device; and inputting the scanning-related requirements into the intelligent RFID scanning and identification recommendation model to obtain a scanning recommendation result. The present invention can optimize the scanning strategy, achieve higher identification accuracy, and realize the cross-device adaptive application of RFID scanning technology.

Description

Technical Field

[0001] The present invention relates to the field of radio frequency identification technology. Specifically, it particularly relates to a recommendation method for cross-device adaptive RFID scanning and identification. Background Art

[0002] With the rapid development of radio frequency identification (RFID) technology, it has been widely applied in many fields such as supply chain management, inventory tracking, and asset monitoring. The RFID technology scans and identifies the tags attached to items through radio waves, realizing the rapid reading and processing of item information. Traditional RFID scanning systems are usually designed for a single device dedicated, lacking flexibility and adaptability, and it is difficult to meet the changing business environment and diverse device requirements.

[0003] In practical applications, RFID scanning devices face various challenges, including compatibility issues between different devices, the influence of environmental factors on the scanning effect, and the high requirements of users for scanning efficiency and accuracy. For example, different RFID scanning devices may have different scanning frequencies, powers, and algorithms, resulting in the inability to achieve the best scanning effect in specific application scenarios. In addition, environmental factors such as metal, liquid, and the human body can all interfere with RFID signals, affecting the identification accuracy of tags.

[0004] To overcome these challenges, a method capable of cross-device adaptive scanning and identifying RFID tags is needed. This method should be able to automatically adjust scanning parameters according to device characteristics and environmental conditions, optimize the scanning strategy, so as to achieve higher identification accuracy and efficiency. At present, there is still a lack of an intelligent cross-device adaptive RFID scanning and identification solution in the market that can meet the comprehensive needs of users for flexibility, accuracy, and efficiency.

[0005] Regarding the problems in the related technology, no effective solution has been proposed yet. Summary of the Invention

[0006] In order to overcome the above problems, the present invention aims to propose a recommendation method for cross-device adaptive RFID scanning and identification, aiming to solve the problem of the inability to intelligently recommend a solution for the cross-device adaptive RFID scanning and identification method.

[0007] For this reason, the specific technical solution adopted by the present invention is as follows:

[0008] A recommendation method for cross-device adaptive RFID scanning and identification, the method includes:

[0009] S1. Construct and utilize an RFID scanning device characteristic model to analyze the RFID scanning device, and obtain the characteristic identification result of the RFID scanning device;

[0010] S2. Based on the characteristic recognition results of the RFID scanning device, match the corresponding scanning module algorithm of the RFID scanning device, and combine the parameter adjustment strategy to obtain the optimal scanning parameters of the RFID scanning device;

[0011] S3. Use the optimal scanning parameters and scanning module algorithm of the RFID scanning device to scan the RFID tag to obtain the scanning effect of the RFID scanning device;

[0012] S4. According to the characteristics and scanning effect of the RFID scanning device, construct an intelligent RFID scanning and recognition recommendation model; input the scanning-related requirements into the intelligent RFID scanning and recognition recommendation model to obtain the intelligent scanning recommendation result;

[0013] S2 includes:

[0014] S23. Use the parameter adjustment strategy to dynamically adjust the scanning parameters of the RFID scanning device according to the characteristic recognition results of the RFID scanning device to obtain the optimal scanning parameters of the RFID scanning device;

[0015] S23 includes:

[0016] S231. Based on the performance evaluation function, evaluate the scanning performance of the RFID scanning device corresponding to different scanning parameter settings to obtain the scanning performance evaluation result.

[0017] Optionally, constructing and using the RFID scanning device characteristic model to analyze the RFID scanning device to obtain the characteristic recognition result of the RFID scanning device includes the following steps:

[0018] S11. According to the characteristics of different RFID scanning devices, construct the RFID scanning device characteristic model to obtain the RFID scanning device feature database;

[0019] S12. Perform a standardized test on the RFID scanning device and obtain the response data of the RFID scanning device;

[0020] S13. Use the feature matching algorithm to compare the response data of the RFID scanning device with the RFID scanning device feature database to obtain the characteristic recognition result of the RFID scanning device.

[0021] Optionally, according to the characteristics of different RFID scanning devices, constructing the RFID scanning device characteristic model to obtain the RFID scanning device feature database includes the following steps:

[0022] S111. According to all the characteristic attributes of different RFID scanning devices, construct the RFID scanning device characteristic model;

[0023] S112. Based on the RFID scanning device characteristic model, extract the parameter values of the corresponding characteristic attributes of different RFID scanning devices respectively to obtain the characteristic vector of the RFID scanning device;

[0024] S113. Use the characteristic vectors of the RFID scanning device to integrate and obtain the RFID scanning device characteristic database.

[0025] Optionally, the standardization test includes factory test, single characteristic attribute test, and combined characteristic attribute test;

[0026] For the factory test, test using the initial configuration of the RFID scanning device when it leaves the factory in a fixed test scenario;

[0027] For the single characteristic attribute test, use the method of controlling variables, fix the parameter values of other characteristic attributes, dynamically adjust the parameter value of a single characteristic attribute according to a linear rule, and conduct tests;

[0028] For the combined characteristic attribute test, use related characteristic attributes, combine the parameter values of related characteristic attributes, and keep other characteristic attributes fixed for testing.

[0029] Optionally, before dynamically adjusting the scanning parameters of the RFID scanning device according to the characteristic recognition result of the RFID scanning device using the parameter adjustment strategy to obtain the optimal scanning parameters of the RFID scanning device, it further includes:

[0030] S21. According to the characteristic recognition result of the RFID scanning device, match the corresponding scanning module algorithm;

[0031] S22. Initialize the scanning parameters of the RFID scanning device.

[0032] Optionally, after evaluating the scanning performance of the RFID scanning device corresponding to different scanning parameter settings based on the performance evaluation function to obtain the scanning performance evaluation result, it further includes:

[0033] S232. According to the scanning performance evaluation result, use an optimization algorithm and combine a constrained optimization algorithm to optimize the scanning parameters of the RFID scanning device;

[0034] S233. Repeat steps S2231 - S232 until the optimal scanning parameters of the RFID scanning device are obtained.

[0035] Optionally, the expression of the performance evaluation function is:

[0036] ;

[0037] In the formula, Performance represents the summary result of all calculated performance indicators; m represents the total number of all performance indicators; SubPerformancei A function representing the i-th performance metric in the RFID scanning device; P represents the set of scanning parameters of the RFID scanning device including all characteristic attributes in the current state; C represents the set of scanning parameters of the RFID scanning device including all characteristic attributes in the target state;

[0038] The expression of the optimization algorithm is:

[0039] ;

[0040] In the formula, Optimized represents the optimization algorithm function; n represents the total number of various characteristic attributes; Pr represents the calculation of the success probability of TagRead of the RFID scanning device; P i represents the scanning parameter of the i-th characteristic attribute of the RFID scanning device in the current state; C i represents the scanning parameter of the i-th characteristic attribute of the RFID scanning device in the target state; P represents the set of scanning parameters of the RFID scanning device including all characteristic attributes in the current state; C represents the set of scanning parameters of the RFID scanning device including all characteristic attributes in the target state;

[0041] The expression of the constrained optimization algorithm is:

[0042] ;

[0043] In the formula, s.t. means restricting that the currently optimized scanning parameters are not allowed to exceed the allowable range of scanning parameter values; represents the set of scanning parameters of the RFID scanning device including all characteristic attributes after optimization; represents the value of the scanning parameter of the i-th characteristic attribute after optimization; Range i represents the parameter value range of the i-th characteristic attribute of the RFID scanning device; Optimized represents the optimization algorithm function; P represents the set of characteristic attributes of the RFID scanning device in the current state; C represents the set of characteristic attributes of the RFID scanning device in the target state.

[0044] Optionally, according to the characteristics of the RFID scanning device and the scanning effect, construct an intelligent RFID scanning and recognition recommendation model; input the scanning-related requirements into the intelligent RFID scanning and recognition recommendation model, and the steps to obtain the intelligent scanning recommendation result are as follows:

[0045] S41. Based on the characteristics of the RFID scanning device and the scanning effect, perform feature extraction and encoding to obtain a data set, and divide the data set into a training set and a test set according to a preset ratio;

[0046] S42. Construct a graph neural network model, and train the graph neural network model according to the training set to obtain an RFID scanning and recognition recommendation model;

[0047] S43. Evaluate and optimize the RFID scanning and recognition recommendation model using the test set to obtain an intelligent RFID scanning and recognition recommendation model;

[0048] S44. Input the RFID scanning-related requirements into the intelligent RFID scanning and recognition recommendation model, and obtain an intelligent scanning recommendation result through the intelligent RFID scanning and recognition recommendation model.

[0049] Optionally, constructing a graph neural network model and training the graph neural network model according to the training set to obtain the RFID scanning and recognition recommendation model includes the following steps:

[0050] S421. Construct a scanning and recognition relationship graph structure according to the training set;

[0051] S422. Based on the scanning and recognition relationship graph structure, initialize the parameters of the graph convolutional neural network model and use the stochastic gradient descent algorithm as the optimization algorithm;

[0052] S423. Train the graph convolutional neural network model according to the training set and in combination with the loss function to obtain the RFID scanning and recognition recommendation model.

[0053] Optionally, evaluating and optimizing the RFID scanning and recognition recommendation model using the test set to obtain an intelligent RFID scanning and recognition recommendation model includes the following steps:

[0054] S431. Evaluate the RFID scanning and recognition recommendation model based on the test set, take the accuracy rate of the RFID scanning and recognition recommendation model as the evaluation index, and obtain the evaluation result of the RFID scanning and recognition recommendation model;

[0055] S432. According to the evaluation result of the RFID scanning and recognition recommendation model, combine random search for hyperparameter optimization, and iteratively update the evaluation result of the RFID scanning and recognition recommendation model to obtain an intelligent RFID scanning and recognition recommendation model.

[0056] Compared with the prior art, the present application has the following beneficial effects: The present invention can automatically adjust the scanning parameters according to the characteristics of the RFID scanning device and environmental conditions, optimize the scanning strategy, achieve higher recognition accuracy and efficiency, and the cross-device adaptive application of RFID scanning technology, improve the intelligent level of asset management, and meet the needs of modern enterprises for high-efficiency and high-precision asset management. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Combined with the following description of the embodiments, the above characteristics, features, and advantages of the present invention and their implementation methods and means become more understandable. The embodiments are described in detail in conjunction with the drawings. Shown herein in schematic diagrams:

[0058] Figure 1 It is a flowchart of a recommendation method for cross-device adaptive RFID scanning and identification according to an embodiment of the present invention. Detailed implementation manners

[0059] In order to enable those skilled in the art to better understand the solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0060] According to an embodiment of the present invention, a recommendation method for cross-device adaptive RFID scanning and identification is provided.

[0061] The present invention will be further described below in conjunction with the accompanying drawings and specific implementation manners. As Figure 1 shown, the recommendation method for cross-device adaptive RFID scanning and identification according to an embodiment of the present invention includes:

[0062] S1. Construct and utilize an RFID scanning device characteristic model to analyze the RFID scanning device and obtain a characteristic identification result of the RFID scanning device.

[0063] In this alternative embodiment, constructing and utilizing an RFID scanning device characteristic model to analyze the RFID scanning device and obtain a characteristic identification result of the RFID scanning device includes the following steps:

[0064] S11. Construct an RFID scanning device characteristic model according to the characteristics of different RFID scanning devices to obtain an RFID scanning device feature database;

[0065] S12. Perform a standardized test on the RFID scanning device and obtain the response data of the RFID scanning device;

[0066] S13. Use a feature matching algorithm to compare the response data of the RFID scanning device with the RFID scanning device feature database to obtain a characteristic identification result of the RFID scanning device.

[0067] In this alternative embodiment, constructing an RFID scanning device characteristic model according to the characteristics of different RFID scanning devices to obtain an RFID scanning device feature database includes the following steps:

[0068] S111. Construct an RFID scanning device characteristic model according to all the characteristic attributes of different RFID scanning devices;

[0069] S112. Based on the RFID scanning device characteristic model, extract the parameter values of the corresponding characteristic attributes of different RFID scanning devices respectively to obtain the characteristic vector of the RFID scanning device;

[0070] S113. Use the characteristic vectors of the RFID scanning devices to integrate and obtain the RFID scanning device characteristic database.

[0071] In this alternative embodiment, the standardized test includes factory test, single characteristic attribute test, and combined characteristic attribute test;

[0072] For the factory test, use the initial configuration of the RFID scanning device when leaving the factory to conduct the test in a fixed test scenario;

[0073] For the single characteristic attribute test, use the method of controlling variables, fix the parameter values of other characteristic attributes, dynamically adjust the parameter values of a single characteristic attribute according to the linear law, and conduct the test;

[0074] For the combined characteristic attribute test, use the related characteristic attributes, combine the parameter values of the related characteristic attributes, and keep other characteristic attributes fixed for the test.

[0075] It should be noted that this step is responsible for automatically detecting and analyzing the hardware and software characteristics of different RFID scanning devices, including but not limited to terminal models, scanning frequencies, powers, etc., to determine the device's capability range and optimal working parameters.

[0076] Construct a device characteristic model, including all recognizable characteristic attributes of the device, such as D = {Model, Batch, Quality, Frequency, Power, Distance, Hangle, Vangle...}; where D represents all recognizable characteristic attributes of the RFID scanning device; Model represents the RFID scanning device model; Bath represents the production batch; Quliy represents the signal strength; Frequency represents the scanning frequency; Power represents the power of the RFID scanning device; Distance represents the reading distance; Hangle represents the horizontal angle of the RFID scanning device; Vangle represents the vertical angle of the RFID scanning device, etc.

[0077] For each device, extract the parameter values of its characteristic attributes to form a characteristic vector: f = [f1, f2, …, f i , …, f n ; where f i is the parameter value of the i-th characteristic attribute of the device.

[0078] Establish a device characteristic database F to store the characteristic vectors f of known devices.

[0079] For the RFID scanning device to be tested, a series of standardized tests are performed to collect the device response data: r = TestResponse(f); where r represents the response data of the RFID scanning device; f represents the feature vector of the RFID scanning device; TestResponse represents the test response result of the RFID scanning device.

[0080] The standardized tests include factory tests, single feature attribute tests, combined feature attribute tests, etc.

[0081] For the factory test, it is carried out under the initial configuration when the device leaves the factory and a fixed test scenario, and the test results are collected.

[0082] For the single feature attribute test, the control variable method is used. The parameter values of other feature attributes are fixed, and the parameter value of a single feature attribute is dynamically adjusted according to a linear rule to collect the test results. For the parameter range with large fluctuations in the test results, the number of test samples can be increased. For example, for power and scanning frequency, they can be tested at several specific gears equally divided within the adjustable range of the device.

[0083] For the combined feature attribute test, for feature attributes with a certain correlation, combined parameter values for common usage scenarios are customized for them, and the test is carried out and the test results are collected while keeping other feature attributes fixed. For example, the horizontal angle and vertical angle of the device can be combined, and multiple common device usage situations are listed, and these related feature attributes are combined and adjusted, and the results are tested and collected in turn.

[0084] Using the feature matching algorithm, the response data r is compared with the feature vectors in the database F to find the most matching device features. The expression of the feature matching algorithm is:

[0085] ;

[0086] In the formula, Distance(r,f) represents a function that measures the difference between two feature vectors, such as Euclidean distance or cosine similarity; represents sequentially executing the subsequent functions on the parameter set f from the dataset F, and selecting the original parameter set f with the smallest result value according to the function results; f * represents the most matching RFID scanning device features.

[0087] The expression of the Euclidean distance metric of the RFID scanning device feature vector is:

[0088] ;

[0089] In the formula, Distance(r,f) represents a function that measures the difference between two feature vectors; r i represents the parameter value of the i-th response data of the RFID scanning device; fi The parameter value representing the i-th characteristic attribute of the RFID scanning device; n represents the dimension of the device feature vector, that is, the number of characteristic attributes.

[0090] The determination of the optimal operating parameters of the device may be based on finding the predefined parameter set corresponding to the most matching feature vector, or calculated through an optimization algorithm.

[0091] According to the matching feature vector f * , determine the capability range and recommended optimal operating parameters of the RFID scanning device. The expression for the capability range of the RFID scanning device is:

[0092] ;

[0093] In the formula, P * represents the capability range of the RFID scanning device; OptimalParams represents a function for filtering out the characteristic attributes of adjustable parameters and their target parameter values.

[0094] First, exclude the non-adjustable parameters, and then calculate the parameter values that meet the range constraints from f * among the adjustable parameter characteristic attributes according to the adjustment range. If f * exceeds the maximum parameter value of the adjustable range, it is constrained to the maximum parameter value; if f * is lower than the minimum parameter value of the adjustable range, it is constrained to the minimum parameter value; if f * is not among the optional parameter values of the adjustable range, it is constrained to the optional parameter value closest to f * among all the optional parameter values.

[0095] Through the above algorithm, the present invention can effectively identify the characteristics of the connected RFID scanning device and provide accurate parameter configuration for device operation, thereby improving the adaptability and efficiency of RFID scanning.

[0096] S2. Based on the characteristic recognition result of the RFID scanning device, match the corresponding scanning module algorithm of the RFID scanning device, and combine the parameter adjustment strategy to obtain the optimal scanning parameters of the RFID scanning device.

[0097] In this alternative embodiment, before obtaining the optimal scanning parameters of the RFID scanning device by dynamically adjusting the scanning parameters of the RFID scanning device according to the characteristic recognition result of the RFID scanning device using the parameter adjustment strategy, it further includes:

[0098] S21. According to the characteristic recognition result of the RFID scanning device, match the corresponding scanning module algorithm;

[0099] S22. Initialize the scanning parameters of the RFID scanning device.

[0100] In this optional embodiment, based on the performance evaluation function, the scanning performance of the RFID scanning device corresponding to different scanning parameter settings is evaluated, and after obtaining the scanning performance evaluation result, the following steps are further included:

[0101] S232. Optimize the scanning parameters of the RFID scanning device using an optimization algorithm combined with a constrained optimization algorithm based on the scanning performance evaluation results.

[0102] S233. Repeat steps S2231-S232 until the optimal scanning parameters of the RFID scanning device are obtained.

[0103] It should be explained that, based on the results of device feature identification, this step automatically adjusts the RFID scanning module algorithm and scanning parameters, such as signal strength, scanning frequency, and reading distance, to adapt to different devices and optimize scanning performance.

[0104] Adapt the device's built-in scanning module algorithm interface according to the device model.

[0105] Receive the RFID scanning device characteristic vector C=[c1,c2,...,c i ,...,c n ]; where c i It represents the parameter value of each characteristic attribute in the characteristic model of the corresponding RFID scanning device; i represents the parameter value of the i-th characteristic attribute of the device; n represents the total number of characteristic attributes defined in the characteristic model of the RFID scanning device.

[0106] Define a set of initial scanning parameters P o =[P o1 ,P o2 ,...,P oi ,...,P on ], including device model, production batch, signal strength, scanning frequency, reading distance, device horizontal angle, device vertical angle, etc.; Among them, P oi Indicates the parameter value of each characteristic attribute in the corresponding device characteristic model in the initial state of the current device; o indicates the initial state; i indicates the parameter value of the i-th characteristic attribute of the device; n indicates the total number of characteristic attributes defined in the device characteristic model.

[0107] Design a parameter adjustment strategy to dynamically adjust scanning parameters based on device characteristics. Repeat the following adjustment steps until the performance of the optimized parameter set exceeds a threshold (which can be set between 50% and 100%), or until the maximum number of repetitions is reached.

[0108] 1. Performance evaluation function:

[0109] Define the performance evaluation function Performance(P k , C), which evaluates the scanning performance under different parameter settings. The performance evaluation function should obtain a performance metric within the range (0%, 100%). Among them, k represents the number of times that have been repeatedly executed; P k represents the set of device parameters in the current latest state.

[0110] The expression of the performance evaluation function is:

[0111] ;

[0112] In the formula, Performance represents the summary result of all calculated performance metrics; m represents the total number of all performance metrics; SubPerformance i represents the function of the i-th performance metric in the RFID scanning device; P represents the set of scanning parameters including all characteristic attributes of the RFID scanning device in the current state; C represents the set of scanning parameters including all characteristic attributes of the RFID scanning device in the target state.

[0113] If the signal-to-noise ratio is used as an indicator, the expression of the signal-to-noise ratio indicator is:

[0114] ;

[0115] In the formula, SubPerformance i represents the function of the i-th performance metric in the RFID scanning device; SNR represents the signal-to-noise ratio function; (P frequency , C frequency ) represents the frequency of the RFID scanning device; (P power , C power ) represents the power of the RFID scanning device; P represents the set of scanning parameters including all characteristic attributes of the RFID scanning device in the current state; C represents the set of scanning parameters including all characteristic attributes of the RFID scanning device in the target state.

[0116] In the above expression, the corresponding signal-to-noise ratio is calculated through the frequency (P frequency , C frequency ) of the RFID scanning device and the power (P power )]], C power ) of the RFID scanning device, and the ratio of the current state indicator to the target state indicator of the returned indicator is used as the result of performance evaluation.

[0117] 2. Parameter optimization:

[0118] Use the optimization algorithm to search for the local optimal parameter set P k+1 = Optimized(P k, C); where k represents the number of times already repeatedly executed; P k represents the set of device parameter sets in the current latest state; P k+1 represents the optimized set of device parameter sets.

[0119] During the parameter optimization process, consider the constraint conditions of the device, such as power limitation, scanning frequency range, etc.

[0120] Optimize the parameter P using an approximation function k to obtain P k the optimization result P approaching C k+1 .

[0121] For example, when using maximum likelihood estimation, the expression of the optimization algorithm is:

[0122] ;

[0123] In the formula, Optimized represents the optimization algorithm function; n represents the total number of each feature attribute; Pr represents the calculated successful probability of TagRead of the RFID scanning device; P i represents the scanning parameter of the i-th feature attribute of the RFID scanning device in the current state; C i represents the scanning parameter of the i-th feature attribute of the RFID scanning device in the target state; P represents the set of scanning parameters of the RFID scanning device including all feature attributes in the current state; C represents the set of scanning parameters of the RFID scanning device including all feature attributes in the target state.

[0124] Combined with the constraint conditions, perform optimization. The expression of the constraint optimization algorithm is:

[0125] ;

[0126] In the formula, s.t. means restricting that the currently optimized scanning parameters are not allowed to exceed the allowable scanning parameter value range; represents the set of scanning parameters of the RFID scanning device including all feature attributes after optimization; represents the scanning parameter value of the i-th feature attribute after optimization; Range i represents the parameter value range of the i-th feature attribute of the RFID scanning device; Optimized represents the optimization algorithm function; P represents the set of feature attributes of the RFID scanning device in the current state; C represents the set of feature attributes of the RFID scanning device in the target state.

[0127] Parameter values exceeding the range should be corrected to values within the range. The OptimalParams function in the previous text can also be referred to. For example, considering the power constraint, according to the above constraint optimization algorithm, ; The power constraint is shown in the example, i.e. Indicates the optimized power parameter value; Maxpower indicates the maximum power parameter value of the device.

[0128] S3. Scan the RFID tag using the optimal scanning parameters and scanning module algorithm of the RFID scanning device to obtain the scanning effect of the RFID scanning device.

[0129] It should be explained that, in cooperation with the parameter adaptation module, this module performs the scanning and identification tasks of RFID tags, and uses the scanning module's own algorithm and scanning parameters to improve the accuracy and robustness of tag recognition.

[0130] S4. Build an intelligent RFID scanning recognition recommendation model based on the characteristics of the RFID scanning device and the scanning effect; input the scanning-related requirements into the intelligent RFID scanning recognition recommendation model to obtain the intelligent scanning recommendation results.

[0131] In this optional embodiment, an intelligent RFID scanning recognition recommendation model is constructed based on the characteristics of the RFID scanning device and the scanning effect; scanning-related requirements are input into the intelligent RFID scanning recognition recommendation model to obtain intelligent scanning recommendation results, including the following steps:

[0132] S41. Based on the characteristics of the RFID scanning device and the scanning effect, feature extraction and encoding are performed to obtain a data set, and the data set is divided into a training set and a test set according to a preset ratio;

[0133] S42. Build a graph neural network model and train the graph neural network model based on the training set to obtain an RFID scanning recognition recommendation model;

[0134] S43, using the test set to evaluate and optimize the RFID scanning recognition recommendation model to obtain an intelligent RFID scanning recognition recommendation model;

[0135] S44. Input the RFID scanning related requirements into the intelligent RFID scanning recognition recommendation model, and obtain the intelligent scanning recommendation results through the intelligent RFID scanning recognition recommendation model.

[0136] In this optional embodiment, constructing a graph neural network model and training the graph neural network model based on a training set to obtain an RFID scanning recognition recommendation model includes the following steps:

[0137] S421. Construct a scanning recognition relationship graph structure based on the training set;

[0138] S422. Initialize the graph convolutional neural network model parameters based on the scanned recognition relationship graph structure, and use the stochastic gradient descent algorithm as the optimization algorithm;

[0139] S423. Train the graph convolutional neural network model according to the training set and in combination with the loss function to obtain the RFID scanning recognition recommendation model.

[0140] In this optional embodiment, evaluating and optimizing the RFID scanning recognition recommendation model using the test set to obtain the intelligent RFID scanning recognition recommendation model includes the following steps:

[0141] S431. Based on the test set, evaluate the RFID scanning recognition recommendation model, use the accuracy rate of the RFID scanning recognition recommendation model as the evaluation index, and obtain the evaluation result of the RFID scanning recognition recommendation model;

[0142] S432. According to the evaluation result of the RFID scanning recognition recommendation model, combine random search for hyperparameter optimization, and iteratively update the evaluation result of the RFID scanning recognition recommendation model to obtain the intelligent RFID scanning recognition recommendation model.

[0143] It should be noted that this step comprehensively considers the device characteristics and scanning effects, provides intelligent scanning recommendations for users, including device selection, parameter configuration, and scanning strategies, etc., to help users achieve the best asset management practices.

[0144] Use the RFID scanning device to collect data, obtain the original data of the scanning effect, extract relevant features, such as signal strength, scanning frequency, etc., and perform encoding and normalization processing on the extracted features to obtain the data set. The division ratio of the data set is to divide 70% of the image type data set into the training set, and the remaining 30% into the test set.

[0145] According to the training set, construct the scanning recognition relationship graph structure. Take the entities (such as RFID tags, scanning devices, etc.) in the RFID scanning scenario as the nodes in the graph, and the relationships between the entities (such as scanning relationships, association relationships, etc.) as the edges in the graph.

[0146] According to the scanning recognition relationship graph structure, determine the input layer, hidden layer, and output layer of the graph convolutional neural network model. Initialize hyperparameters such as the learning rate, batch size, and number of training epochs in the graph convolutional neural network model, and select the stochastic gradient descent (SGD) algorithm as the optimization algorithm for updating the initialized graph convolutional neural network model parameters.

[0147] Input the training set into the graph convolutional neural network model, perform forward propagation to calculate the predicted output of the model, calculate the error between the predicted output and the true label according to the loss function (cross-entropy loss function), and then use the backpropagation algorithm and the optimization algorithm to update the graph convolutional neural network model. Repeat multiple rounds of iterative training until the model converges or reaches the preset number of training epochs to obtain the RFID scanning recognition recommendation model.

[0148] Use the test set to evaluate the RFID scanning and identification recommendation model, calculate the accuracy of the RFID scanning and identification recommendation model and use it as an evaluation index to obtain the evaluation result of the RFID scanning and identification recommendation model.

[0149] According to the evaluation results, use the random search method to optimize the hyperparameters of the model. Random search can randomly select parameter combinations in the predefined hyperparameter space to find the optimal hyperparameter configuration. Evaluate the model with optimized hyperparameters on the test set again to ensure that the performance of the RFID scanning and identification recommendation model is improved. According to the above process, iteratively update the performance of the RFID scanning and identification recommendation model to obtain an intelligent RFID scanning and identification recommendation model.

[0150] Input the requirements related to RFID scanning (such as the tag information to be scanned, scanning environment parameters, etc.) into the intelligent RFID scanning and identification recommendation model to obtain intelligent scanning recommendation results (such as the recommended best scanning position, scanning time, or tag priority, etc.) through the intelligent RFID scanning and identification recommendation model.

[0151] In summary, with the above technical solutions of the present invention, the present invention can automatically adjust scanning parameters according to the characteristics of RFID scanning devices and environmental conditions, optimize scanning strategies, achieve higher identification accuracy and efficiency, and cross-device adaptive application of RFID scanning technology, improve the intelligent level of asset management, and meet the needs of modern enterprises for high-efficiency and high-precision asset management.

[0152] Although the present invention has been disclosed above with preferred embodiments, the embodiments are only for the purpose of illustration and are not intended to limit the present invention. Those skilled in the art can make several modifications and refinements without departing from the spirit and scope of the present invention. The scope of protection claimed by the present invention shall be subject to what is described in the claims.

Claims

1. A recommended method for cross-device adaptive RFID scanning and identification, characterized in that, The method includes: S1. Construct and utilize the RFID scanning device characteristic model, analyze the RFID scanning device, and obtain the characteristic recognition result of the RFID scanning device; S2. Based on the characteristic recognition result of the RFID scanning device, match the corresponding scanning module algorithm of the RFID scanning device, and combine the parameter adjustment strategy to obtain the optimal scanning parameters of the RFID scanning device; S3. Utilize the optimal scanning parameters and scanning module algorithm of the RFID scanning device to scan the RFID tag and obtain the scanning effect of the RFID scanning device; S4. Construct an intelligent RFID scanning and recognition recommendation model according to the characteristics and scanning effect of the RFID scanning device; input the scanning-related requirements into the intelligent RFID scanning and recognition recommendation model to obtain the intelligent scanning recommendation result; The S2 includes: S23. Utilize the parameter adjustment strategy to dynamically adjust the scanning parameters of the RFID scanning device according to the characteristic recognition result of the RFID scanning device to obtain the optimal scanning parameters of the RFID scanning device; The S23 includes: S231. Based on the performance evaluation function, evaluate the scanning performance of the RFID scanning device corresponding to different scanning parameter settings to obtain the scanning performance evaluation result; After obtaining the scanning performance evaluation result by evaluating the scanning performance of the RFID scanning device corresponding to different scanning parameter settings based on the performance evaluation function, it further includes: S232. According to the scanning performance evaluation result, utilize the optimization algorithm and combine the constrained optimization algorithm to optimize the scanning parameters of the RFID scanning device; S233. Repeat steps S231 - S232 until the optimal scanning parameters of the RFID scanning device are obtained; The expression of the performance evaluation function is: ； In the formula, Performance represents the summary result of all calculated performance indicators; m represents the total number of all performance indicators; SubPerformance i represents the function of the i-th performance indicator in the RFID scanning device; P represents the set of scanning parameters of the RFID scanning device including all characteristic attributes in the current state; C represents the set of scanning parameters of the RFID scanning device including all characteristic attributes in the target state; The expression of the optimization algorithm is: ； Where, Optimized represents the optimization algorithm function; n represents the total number of each feature attribute; Pr represents the successful probability of calculating TagRead of the RFID scanning device; P i represents the scanning parameter of the i-th feature attribute of the RFID scanning device in the current state; C i represents the scanning parameter of the i-th feature attribute of the RFID scanning device in the target state; P represents the set of scanning parameters of the RFID scanning device including all feature attributes in the current state; C represents the set of scanning parameters of the RFID scanning device including all feature attributes in the target state; The expression of the constrained optimization algorithm is: ； In the formula, s.t. means restricting that the currently optimized scanning parameters are not allowed to exceed the allowable range of scanning parameter values; represents the set of scanning parameters of the optimized RFID scanning device including all characteristic attributes; represents the scanning parameter value of the i-th characteristic attribute after optimization; Range i represents the parameter value range of the i-th characteristic attribute of the RFID scanning device; Optimized represents the optimization algorithm function; P represents the set of characteristic attributes of the RFID scanning device in the current state; C represents the set of characteristic attributes of the RFID scanning device in the target state.

2. The recommended method for cross-device adaptive RFID scanning and recognition according to claim 1, characterized in that The steps of constructing and utilizing the RFID scanning device characteristic model, analyzing the RFID scanning device, and obtaining the characteristic recognition result of the RFID scanning device include the following: S11. Construct the RFID scanning device characteristic model according to the characteristics of different RFID scanning devices to obtain the RFID scanning device feature database; S12. Perform a standardized test on the RFID scanning device and obtain the response data of the RFID scanning device; S13. Utilize the feature matching algorithm to compare the response data of the RFID scanning device with the RFID scanning device feature database to obtain the characteristic recognition result of the RFID scanning device.

3. The recommended method for cross-device adaptive RFID scanning and identification according to claim 2, characterized in that, The steps of constructing the RFID scanning device characteristic model according to the characteristics of different RFID scanning devices to obtain the RFID scanning device feature database include the following: S111. Construct the RFID scanning device characteristic model according to all the characteristic attributes of different RFID scanning devices; S112. Based on the RFID scanning device characteristic model, extract the parameter values of the corresponding characteristic attributes of different RFID scanning devices respectively to obtain the feature vector of the RFID scanning device; S113. Integrate the feature vectors of the RFID scanning device to obtain the RFID scanning device feature database.

4. The recommended method for cross-device adaptive RFID scanning and identification according to claim 2, characterized in that, The standardized tests include factory tests, single feature property tests, and combined feature property tests; For the factory test, the initial configuration of the RFID scanning device when leaving the factory is used for testing in a fixed test scenario; For the single feature property test, using the method of controlling variables, fixing the parameter values of other feature properties, dynamically adjusting the parameter values of a single feature property according to a linear rule, and conducting tests; For the combined feature property test, using related feature properties, combining the parameter values of related feature properties, and keeping other feature properties fixed for testing.

5. The recommended method for cross-device adaptive RFID scanning and identification according to claim 1, wherein Before obtaining the optimal scanning parameters of the RFID scanning device by dynamically adjusting the scanning parameters of the RFID scanning device according to the characteristic recognition result of the RFID scanning device using the parameter adjustment strategy, it further includes: S21. Matching the corresponding scanning module algorithm according to the characteristic recognition result of the RFID scanning device; S22. Initializing the scanning parameters of the RFID scanning device.

6. The recommended method for cross-device adaptive RFID scanning and identification according to claim 1, characterized in that Constructing an intelligent RFID scanning and recognition recommendation model based on the characteristics and scanning effects of the RFID scanning device; inputting the scanning-related requirements into the intelligent RFID scanning and recognition recommendation model to obtain the intelligent scanning recommendation result includes the following steps: S41. Based on the characteristics and scanning effects of the RFID scanning device, performing feature extraction and encoding to obtain a data set, and dividing the data set into a training set and a test set according to a preset ratio; S42. Constructing a graph neural network model and training the graph neural network model according to the training set to obtain an RFID scanning and recognition recommendation model; S43. Using the test set to evaluate and optimize the RFID scanning and recognition recommendation model to obtain an intelligent RFID scanning and recognition recommendation model; S44. Inputting the RFID scanning-related requirements into the intelligent RFID scanning and recognition recommendation model, and obtaining the intelligent scanning recommendation result through the intelligent RFID scanning and recognition recommendation model.

7. The recommended method for cross-device adaptive RFID scanning and recognition according to claim 6, characterized in that, The steps of constructing a graph neural network model and training the graph neural network model according to the training set to obtain an RFID scanning and recognition recommendation model include: S421. Constructing a scanning and recognition relationship graph structure according to the training set; S422. Based on the scanning and recognition relationship graph structure, initializing the parameters of the graph convolutional neural network model and using the stochastic gradient descent algorithm as the optimization algorithm; S423. Training the graph convolutional neural network model according to the training set and in combination with the loss function to obtain an RFID scanning and recognition recommendation model.

8. A recommended method for cross-device adaptive RFID scanning and identification according to claim 7, characterized in that, The steps of using the test set to evaluate and optimize the RFID scanning and recognition recommendation model to obtain an intelligent RFID scanning and recognition recommendation model include: S431. Evaluating the RFID scanning and recognition recommendation model based on the test set, using the accuracy rate of the RFID scanning and recognition recommendation model as the evaluation index to obtain the evaluation result of the RFID scanning and recognition recommendation model; S432. According to the evaluation result of the RFID scanning and recognition recommendation model, combining random search for hyperparameter optimization, and iteratively updating the evaluation result of the RFID scanning and recognition recommendation model to obtain an intelligent RFID scanning and recognition recommendation model.

Images