Sample monitoring method, system, equipment and medium based on RFID technology

Abstract

The present invention discloses a sample monitoring method, system, device, and medium based on RFID technology. The present invention relates to the field of RFID technology. By preprocessing sample signal data, the preprocessed RFID signal data can provide accurate and reliable raw data for sample monitoring, providing support for subsequent path loss prediction and positioning prediction. By establishing a path loss prediction model, the loss of signals during transmission can be accurately predicted, providing the necessary data support for the precise positioning of samples. The positioning network model established using Bi-LSTM technology can combine RFID signal data and path loss data to perform real-time and accurate positioning prediction of samples, thereby improving the efficiency and accuracy of sample monitoring.

Description

Technical Field

[0001] The present invention relates to the field of RFID technology, and in particular to a sample monitoring method, system, equipment and medium based on RFID technology. Background Art

[0002] In the sample management industry, RFID technology is currently the primary method of managing samples. RFID (Radio Frequency Identification) technology, a key enabler of the "Internet of Things," is a wireless communication technology that uses radio signals to identify specific targets without direct contact and reads and writes relevant data. High-frequency (HF) RFID devices operate in the 13.56 MHz frequency band. The system transmits energy through inductive coupling between antenna coils, which results in a rapid drop in magnetic field energy.

[0003] However, after investigating the relevant industries, it was found that the number of samples was huge, the categories were numerous, and the storage was not standardized; during the process of collecting and returning samples, there was a phenomenon of multiple people taking or storing samples at the same time, which easily caused chaos in sample management; during the sample return process, the samples were not stored in the prescribed location, which made it difficult to find the samples.

[0004] In summary, the current application of RFID technology in sample monitoring makes it difficult to identify and monitor the sample collection trajectory. At the same time, current RFID technology is difficult to distinguish similar or similar samples, which leads to samples not being returned or not being returned to the corresponding location.

[0005] Due to the above technical problems, there is an urgent need for a sample monitoring method that uses RFID technology to accurately locate and track samples, thereby solving the problem of chaotic sample management. Summary of the Invention

[0006] Based on the problems raised by the above background technology, the purpose of the present invention is to provide a sample monitoring method, system, equipment and medium based on RFID technology, which solves the problem that it is difficult to identify and monitor the sample taking trajectory when applying RFID technology to sample monitoring. At the same time, the current RFID technology is difficult to distinguish similar and similar samples, which leads to the problem that samples are not returned or are not returned to the corresponding location.

[0007] The present invention is achieved through the following technical solutions:

[0008] The first aspect of the present invention provides a sample monitoring method based on RFID technology, comprising the following steps:

[0009] Step S1: Acquire sample signal data, pre-process the sample signal data, and obtain RFID signal data;

[0010] Step S2: establishing a path loss prediction model, and using the path loss prediction model to process the RFID signal data to obtain path loss data;

[0011] Step S3: Establish a positioning network model based on Bi-LSTM technology, input the RFID signal data and the path loss data into the positioning network model to perform positioning prediction and obtain sample coordinates.

[0012] In the above technical solution, during the sample monitoring process, it is necessary to accurately obtain the location and status information of the sample. In this method, by preprocessing the sample signal data, the preprocessed RFID signal data can provide accurate and reliable raw data for sample monitoring, providing support for subsequent path loss prediction and positioning prediction.

[0013] During RFID signal transmission, environmental factors can affect the signal, leading to path loss. Traditional monitoring methods may not accurately predict path loss, thus affecting positioning accuracy. This method, by establishing a path loss prediction model, can accurately predict signal loss during transmission, providing the necessary data support for precise sample positioning.

[0014] Accurately locating samples is a key technical issue in sample monitoring. Traditional positioning methods can suffer from low positioning accuracy and an inability to update location information in real time. Because Bidirectional Long Short Term Memory (Bi-LSTM) technology can process time series data and has excellent predictive capabilities for dynamically changing location information, this method uses a positioning network model built using Bi-LSTM technology to combine RFID signal data and path loss data to accurately predict the location of samples in real time, improving the efficiency and accuracy of sample monitoring.

[0015] In an optional embodiment, preprocessing the sample signal data includes the following steps:

[0016] Step S11: constructing an RFID unified data format, extracting data from the sample signal data according to the RFID unified data format, and establishing a signal path tree based on the extracted data;

[0017] Step S12: constructing a multidimensional signal structure database based on the signal path tree, and performing data mining on the data in the multidimensional signal structure database to obtain sample signal features;

[0018] Step S13: using a sliding window technique to sequence the sample signal features to generate RFID signal data.

[0019] In an optional embodiment, performing data mining on the data in the multidimensional signal structure database includes the following steps:

[0020] Step S121: Acquire position data from the data in the multidimensional signal structure database, perform position mining on the position data, and obtain a sample position sequence;

[0021] Step S122: recursively search the sample position sequence and perform a closedness test on the recursive search results to obtain closed paths and non-closed paths;

[0022] Step S123: Acquire a sample processing path, compare the closed path with the sample processing path, and divide the closed path into a frequent closed path and an infrequent closed path according to the comparison result.

[0023] In an optional embodiment, the sample signal features are serialized using a sliding window technique, including the following steps:

[0024] Step S131: extracting the time data groups corresponding to the frequent closed paths, the infrequent closed paths, and the infrequent closed paths from the signal path tree, and forming corresponding spatiotemporal data sets;

[0025] Step S132: Serialize the spatiotemporal dataset in the form of a time axis to obtain a spatiotemporal sequence dataset;

[0026] Step S133: Find the minimum frequently closed period in the spatiotemporal sequence data set, and construct a sliding window and a sliding step, wherein the sliding window is larger than the minimum frequently closed period, and the sliding step is smaller than the minimum frequently closed section.

[0027] In an optional embodiment, processing the RFID signal data using the path loss prediction model includes the following steps:

[0028] Step S21: traverse the RFID signal data to obtain the identification data group and signal data group of the RFID signal data at time t;

[0029] Step S22: performing sample attribute matching on the identification data group. If the sample attributes match, the path loss value of the signal data group is calculated using the path loss prediction model.

[0030] In an optional embodiment, a positioning network model is established based on Bi-LSTM technology, including:

[0031] Establishing an input layer, wherein the input layer is used to obtain RFID signal data and path loss data;

[0032] Establishing a preprocessing layer, the preprocessing layer is used to preprocess the RFID signal data and the path loss data to generate a feature sequence;

[0033] Establishing a Bi-LSTM layer, wherein the Bi-LSTM layer is used to extract features from the feature sequence;

[0034] The Bi-LSTM layer includes a forward feature extraction layer and a reverse feature extraction layer. The forward feature extraction layer is used to perform forward feature extraction on the feature sequence in combination with the path loss data to obtain path loss features; the reverse feature extraction layer is used to perform reverse feature extraction on the feature sequence to obtain path features.

[0035] Establishing a pattern recognition layer, wherein the pattern recognition layer is used to perform pattern recognition on the path features to generate a path pattern;

[0036] Establishing a feature splicing layer, wherein the feature splicing layer is used to splice the path loss feature with the path feature based on the path pattern and using an attention mechanism to obtain a spliced ​​feature;

[0037] A DNN layer is established, where the DNN layer is used to perform positioning prediction based on the splicing features and generate sample coordinates.

[0038] In an optional embodiment, inputting the RFID signal data and the path loss data into the positioning network model to perform positioning prediction includes the following steps:

[0039] Step S31: pre-process the RFID signal data to generate a path feature sequence, and perform point-to-point splicing of the path feature sequence with the path loss data to generate a path loss feature sequence;

[0040] Step S32: using the forward feature extraction layer to extract loss features from each sequence point in the path loss feature sequence, and sequentially performing feature recognition on the loss features of each sequence point to obtain a path loss feature; using the reverse feature extraction layer to extract path features from each sequence point in the path feature sequence, and sequentially performing feature recognition on the path features of each sequence point to obtain a path feature;

[0041] Step S33: performing a frequent closedness test on the path feature, and generating a path pattern according to the frequent closedness test result; wherein the path pattern includes a frequent closed pattern, a non-frequent closed pattern, and a non-closed pattern;

[0042] Step S34: If the path pattern is a frequent closed pattern, the path loss feature is concatenated with the path feature to obtain a standard concatenated feature, and the standard concatenated feature is compared with the sample processing path to generate a standard path feature; if the path pattern is an infrequent closed pattern or a non-closed pattern, the path loss feature is concatenated with the path feature layer by layer using an attention mechanism to obtain a non-standard concatenated feature;

[0043] Step S35: Use the DNN layer to perform positioning prediction on the standard splicing features to generate standard sample coordinates; extract several features from the non-standard splicing features as starting features, and use the DNN layer to perform positioning prediction on the RFID signal data based on the starting features to generate non-standard sample coordinates.

[0044] A second aspect of the present invention provides a sample monitoring system based on RFID technology, comprising:

[0045] A preprocessing module is used to obtain sample signal data and preprocess the sample signal data to obtain RFID signal data;

[0046] a loss processing module, configured to establish a path loss prediction model, and process the RFID signal data using the path loss prediction model to obtain path loss data;

[0047] The positioning prediction module is used to establish a positioning network model based on Bi-LSTM technology, input the RFID signal data and the path loss data into the positioning network model to perform positioning prediction and obtain sample coordinates.

[0048] A third aspect of the present invention provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements a sample monitoring method based on RFID technology when executing the computer program.

[0049] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which implements a sample monitoring method based on RFID technology when executed by a processor.

[0050] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0051] The present invention constructs a dual neural network model, and uses the path loss prediction model to predict the path loss coefficient under the current environment, thereby avoiding positioning errors caused by environmental changes; through the positioning network model, the occurrence of actual conflicts is reduced while accurately positioning, and at the same time, labels of different protocols can be taken into account. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the examples. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be considered as limiting the scope. A person of ordinary skill in the art can also derive other relevant drawings based on these drawings without inventive effort. In the drawings:

[0053] Figure 1 A schematic diagram of the flow of a sample monitoring method based on RFID technology provided in Example 1 of the present invention;

[0054] Figure 2 A schematic diagram of the structure of a sample monitoring system based on RFID technology provided in Example 2 of the present invention;

[0055] Figure 3 This is a structural diagram of an electronic device provided in Example 3 of the present invention. DETAILED DESCRIPTION

[0056] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with examples and drawings. The exemplary embodiments of the present invention and their descriptions are only used to explain the present invention and are not intended to limit the present invention.

[0057] Example 1 of the present invention provides a sample monitoring method based on RFID technology, such as Figure 1 As shown, the sample monitoring method based on RFID technology includes the following steps:

[0058] Step S1: Acquire sample signal data, pre-process the sample signal data, and obtain RFID signal data;

[0059] Step S2: establishing a path loss prediction model, and using the path loss prediction model to process the RFID signal data to obtain path loss data;

[0060] Step S3: Establish a positioning network model based on Bi-LSTM technology, input the RFID signal data and the path loss data into the positioning network model to perform positioning prediction and obtain sample coordinates.

[0061] It should be noted that during the sample monitoring process, it is necessary to accurately obtain the location and status information of the sample. In this method, by preprocessing the sample signal data, the preprocessed RFID signal data can provide accurate and reliable raw data for sample monitoring, providing support for subsequent path loss prediction and positioning prediction.

[0062] During RFID signal transmission, environmental factors can affect the signal, leading to path loss. Traditional monitoring methods may not accurately predict path loss, thus affecting positioning accuracy. This method, by establishing a path loss prediction model, can accurately predict signal loss during transmission, providing the necessary data support for precise sample positioning.

[0063] Accurately locating samples is a key technical challenge in sample monitoring. Traditional positioning methods can suffer from low positioning accuracy and an inability to update location information in real time. Because Bi-LSTM technology can process time series data and has excellent predictive capabilities for dynamically changing location information, this method, using Bi-LSTM technology to build a positioning network model, can combine RFID signal data and path loss data to accurately predict the location of samples in real time, improving the efficiency and accuracy of sample monitoring.

[0064] In an optional embodiment, preprocessing the sample signal data includes the following steps:

[0065] Step S11: constructing an RFID unified data format, extracting data from the sample signal data according to the RFID unified data format, and establishing a signal path tree based on the extracted data;

[0066] Step S12: constructing a multidimensional signal structure database based on the signal path tree, and performing data mining on the data in the multidimensional signal structure database to obtain sample signal features;

[0067] Step S13: using a sliding window technique to sequence the sample signal features to generate RFID signal data.

[0068] In this embodiment, a unified RFID data format is constructed as follows: an identification data group, a location data group, a time data group, and a signal data group. The signal data group includes signal strength, which is processed along with the time data group in subsequent processing of step S1, with each time data group corresponding to one signal strength. The unified RFID data format is constructed because existing high-frequency card readers, such as ISO 14443 and ISO 15693, can only read tags with a single protocol and cannot simultaneously read tags with two different protocols. To achieve compatibility between different RFID protocols, this embodiment of the present invention proposes converting data from different RFID tags into a unified format at the data level to facilitate subsequent data processing.

[0069] Specifically, extracting data from sample signal data in the unified RFID data format involves assigning a unique identifier to each selected sample. This globally unique identifier is used to subsequently identify the same sample and prevent misidentification during sample signal data processing. This identification data set also includes sample attribute data, such as sample shape, length, and name.

[0070] After assigning the sample identifier, the sample signal data with the same identifier are aggregated, and the time-related data are extracted from the sample signal data and placed in the time data group, such as: sample scanning time, sample retention time, sample destruction time, etc.; and the position-related data are extracted from the sample signal data and placed in the position data group, such as: sample coordinates.

[0071] The aggregated sample signal data is sorted according to the chronological order in the time data group, and then a signal path tree is constructed based on the position data in the position data group corresponding to the time data group. A signal path tree represents the same sample. The tree structure represents the sample's time information, and the tree nodes store the sample's location data, thus completing the construction of the signal path tree. This signal path tree can provide information on the sample's movement trends, providing accurate data for subsequent mining of sample signal features and improving data traversal efficiency.

[0072] Among them, the purpose of constructing a multidimensional signal structure database based on the signal path tree is that since the sample signal data has multidimensional attributes, sample attributes, time attributes and space attributes, in order to correctly mine the characteristic information of the data, it is necessary to construct a correct and reasonable multidimensional signal structure database.

[0073] In this embodiment, a multidimensional signal structure database is constructed using sample attributes. The multidimensional signal structure database of sample A is constructed as follows {(sample A)(sample parameter a)(storage warehouse a): 4; (sample A)(sample parameter a)(storage warehouse b): 1; ...; (sample A)(sample parameter d)(storage warehouse d): 2}. Among them, the sample parameters include sample size, material, etc. The storage warehouse is used to indicate the storage location of the sample type, and the following number indicates the number of samples of this category in the identification process. Establishing a multidimensional signal structure database based on sample attributes is beneficial to laboratory sample management, etc. For example, in an experiment, there are several groups of control samples, and the composition and content of each group of samples are slightly different. If the data of this type of attributes is not processed in the RFID data, it will cause sample mixing, which will lead to experimental errors.

[0074] In an optional embodiment, performing data mining on the data in the multidimensional signal structure database includes the following steps:

[0075] Step S121: Acquire position data from the data in the multidimensional signal structure database, perform position mining on the position data, and obtain a sample position sequence;

[0076] Step S122: recursively search the sample position sequence and perform a closedness test on the recursive search results to obtain closed paths and non-closed paths;

[0077] Step S123: Acquire a sample processing path, compare the closed path with the sample processing path, and divide the closed path into a frequent closed path and an infrequent closed path according to the comparison result.

[0078] It should be noted that the sample signal features in this method include the sample path characteristics, which specifically include frequent closed paths, infrequent closed paths, and non-closed paths. The purpose of extracting these features in this method is to study the sample's movement process and the location it reaches after movement, so as to better manage sample storage and access.

[0079] In an optional embodiment, the sample signal features are serialized using a sliding window technique, including the following steps:

[0080] Step S131: extracting the time data groups corresponding to the frequent closed paths, the infrequent closed paths, and the infrequent closed paths from the signal path tree, and forming corresponding spatiotemporal data sets;

[0081] Step S132: Serialize the spatiotemporal dataset in the form of a time axis to obtain a spatiotemporal sequence dataset;

[0082] Step S133: Find the minimum frequently closed period in the spatiotemporal sequence data set, and construct a sliding window and a sliding step, wherein the sliding window is larger than the minimum frequently closed period, and the sliding step is smaller than the minimum frequently closed section.

[0083] In this embodiment of the present invention, RFID signal data is a multidimensional data set of temporal and spatial sequences that includes a sliding window. Using the minimum frequently closed section as the demarcation point, the sliding window is set larger than the minimum frequently closed period, and the sliding step is set smaller than the minimum frequently closed section. This allows for the detection of frequently closed paths and avoids missed detections.

[0084] The spatiotemporal dataset is serialized and displayed as a timeline, which includes not only the time points of sample movement but also the time periods when the samples stayed at a certain location during a certain period. This is because the storage and retrieval of samples not only includes the movement time but also the sample retention time. Taking this retention time into account ensures that the window covers the entire sequence and helps the positioning network model to perform accurate positioning detection.

[0085] In an optional embodiment, processing the RFID signal data using the path loss prediction model includes the following steps:

[0086] Step S21: traverse the RFID signal data to obtain the identification data group and signal data group of the RFID signal data at time t;

[0087] Step S22: performing sample attribute matching on the identification data group. If the sample attributes match, the path loss value of the signal data group is calculated using the path loss prediction model.

[0088] It should be noted that, in order to ensure the identity of the samples, the sample attributes in the identification data set need to be matched in this step, namely, sample shape, sample length, sample name, etc., to prevent sample duplication.

[0089] In this embodiment, the path loss model may use a logarithmic path fading model or other models to calculate the signal strength in the signal data group, thereby obtaining a path loss value.

[0090] Furthermore, after calculating the path loss value at time t, if the spatial data of the RFID signal data at time t+1, i.e., the sample in the spatiotemporal sequence data set, is stagnant from time t to time t+1, the path loss value at time t is used as the path loss value at time t+1 until the stagnation disappears. This judgment reduces computing resource consumption and improves computational efficiency.

[0091] In an optional embodiment, a positioning network model is established based on Bi-LSTM technology, including:

[0092] Establishing an input layer, wherein the input layer is used to obtain RFID signal data and path loss data;

[0093] Establishing a preprocessing layer, the preprocessing layer is used to preprocess the RFID signal data and the path loss data to generate a feature sequence;

[0094] Establishing a Bi-LSTM layer, wherein the Bi-LSTM layer is used to extract features from the feature sequence;

[0095] The Bi-LSTM layer includes a forward feature extraction layer and a reverse feature extraction layer. The forward feature extraction layer is used to perform forward feature extraction on the feature sequence in combination with the path loss data to obtain path loss features; the reverse feature extraction layer is used to perform reverse feature extraction on the feature sequence to obtain path features.

[0096] Establishing a pattern recognition layer, wherein the pattern recognition layer is used to perform pattern recognition on the path features to generate a path pattern;

[0097] Establishing a feature splicing layer, wherein the feature splicing layer is used to splice the path loss feature with the path feature based on the path pattern and using an attention mechanism to obtain a spliced ​​feature;

[0098] A DNN layer is established, where the DNN layer is used to perform positioning prediction based on the splicing features and generate sample coordinates.

[0099] In an optional embodiment, inputting the RFID signal data and the path loss data into the positioning network model to perform positioning prediction includes the following steps:

[0100] Step S31: pre-process the RFID signal data to generate a path feature sequence, and perform point-to-point splicing of the path feature sequence with the path loss data to generate a path loss feature sequence;

[0101] Step S32: using the forward feature extraction layer to extract loss features from each sequence point in the path loss feature sequence, and sequentially performing feature recognition on the loss features of each sequence point to obtain a path loss feature; using the reverse feature extraction layer to extract path features from each sequence point in the path feature sequence, and sequentially performing feature recognition on the path features of each sequence point to obtain a path feature;

[0102] Step S33: performing a frequent closedness test on the path feature, and generating a path pattern according to the frequent closedness test result; wherein the path pattern includes a frequent closed pattern, a non-frequent closed pattern, and a non-closed pattern;

[0103] Step S34: If the path pattern is a frequent closed pattern, the path loss feature is concatenated with the path feature to obtain a standard concatenated feature, and the standard concatenated feature is compared with the sample processing path to generate a standard path feature; if the path pattern is an infrequent closed pattern or a non-closed pattern, the path loss feature is concatenated with the path feature layer by layer using an attention mechanism to obtain a non-standard concatenated feature;

[0104] Step S35: Use the DNN layer to perform positioning prediction on the standard splicing features to generate standard sample coordinates; extract several features from the non-standard splicing features as starting features, and use the DNN layer to perform positioning prediction on the RFID signal data based on the starting features to generate non-standard sample coordinates.

[0105] It should be noted that the present invention requires forward and reverse feature extraction of RFID signal data through the Bi-LSTM layer. Forward extraction is used to extract the changing characteristics of RFID signals during forward propagation, and requires combining path loss data for feature extraction. Reverse extraction aims to extract path-related features in RFID sequence data, and only relies on time and location-related data in the RFID signal data for feature extraction. It should be noted that the preprocessing before forward and reverse extraction also includes the determination of sample attributes. If the sample identification and sample attributes are inconsistent, it indicates that the data is not the same sample, and the data is deemed abnormal and discarded.

[0106] Among them, point-to-point splicing is to splice the path loss value at a certain time in the spatiotemporal series data set at the sequence point, so that the corresponding path loss value can be extracted at the same time when performing spatiotemporal feature extraction. The feature node calculation of the spatiotemporal feature and the path loss value can be performed to obtain the path loss feature.

[0107] Specifically, the process of identifying the loss features of each sequence point in turn is as follows:

[0108]

[0109] In the above formula, is the path loss feature at time t in the forward extraction process, a represents forward extraction, f() is the activation function, w1 and w2 are the time series features X at time t respectively. t and path loss value N t The input weight of , b1 represents the forward extraction bias value.

[0110] In this embodiment, the input weights are randomly generated.

[0111] Specifically, the process of identifying the path features of each sequence point in turn is as follows:

[0112]

[0113] In the above formula, is the path loss feature at time t during the reverse extraction process, c represents reverse extraction, f() is the activation function, w3 and w4 are the time features at time t respectively and spatial characteristics The input weight of , b2 represents the reverse extraction bias value.

[0114] In step S1, the path data is divided into non-closed paths, frequently closed paths and non-frequently closed paths by preprocessing the data. The process of pattern matching of path features here is similar to that in step S12. Frequently closed paths belong to the regular paths of samples, which basically follow the predetermined paths of samples for storage and retrieval. In order to save computing resources, frequent closure detection is first performed here, and the data corresponding to the frequent closed patterns are standardized and spliced ​​using templates that match the sample processing path to generate standard splicing features. In the subsequent process of DNN layer recognition, the past sample processing paths are used as a reference for coordinate prediction.

[0115] For infrequent closed or non-closed modes, where the driving path is not a standard path, an attention mechanism is used to concatenate the path loss features with the path features layer by layer to generate non-standard concatenated features. In the subsequent DNN layer recognition process, several features from the non-standard concatenated features are extracted as starting features. Based on these starting features, the DNN layer uses the RFID signal data to predict the path, generating the coordinates of the possible subsequent movement.

[0116] Example 2 of the present invention provides a sample monitoring system based on RFID technology, such as Figure 2 As shown, the sample monitoring system based on RFID technology includes:

[0117] A preprocessing module is used to obtain sample signal data and preprocess the sample signal data to obtain RFID signal data;

[0118] a loss processing module, configured to establish a path loss prediction model, and process the RFID signal data using the path loss prediction model to obtain path loss data;

[0119] The positioning prediction module is used to establish a positioning network model based on Bi-LSTM technology, input the RFID signal data and the path loss data into the positioning network model to perform positioning prediction and obtain sample coordinates.

[0120] Embodiment 3 of the present invention provides an electronic device, such as Figure 3 As shown, the electronic device includes a processor 21, a memory 22, an input device 23 and an output device 24; the number of processors 21 in the computer device can be one or more. Figure 3 In the figure, a processor 21 is taken as an example; the processor 21, memory 22, input device 23 and output device 24 in the electronic device can be connected by a bus or other means. Figure 3 The bus connection is taken as an example.

[0121] Memory 22, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules. Processor 21 executes the software programs, instructions, and modules stored in memory 22 to perform various electronic device functions and data processing, thereby implementing the RFID-based sample monitoring method of Example 1.

[0122] The memory 22 may primarily include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function; the data storage area may store data created based on the use of the terminal, etc. Furthermore, the memory 22 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state memory device. In some instances, the memory 22 may further include a memory remotely located relative to the processor 21, and these remote memories may be connected to the electronic device via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0123] The input device 23 can be used to receive the ID and password input by the user, etc. The output device 24 is used to output the network configuration page.

[0124] Embodiment 4 of the present invention further provides a computer-readable storage medium, wherein the computer-executable instructions, when executed by a computer processor, are used to implement the sample monitoring method based on RFID technology as provided in embodiment 1.

[0125] An embodiment of the present invention provides a storage medium containing computer-executable instructions, and its computer-executable instructions are not limited to the method operations provided in Example 1, but can also execute related operations in the sample monitoring method based on RFID technology provided in any embodiment of the present invention.

[0126] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A sample monitoring method based on RFID technology, characterized in that: The steps include: Step S1: Acquire sample signal data, pre-process the sample signal data, and obtain RFID signal data; Step S2: establishing a path loss prediction model, and using the path loss prediction model to process the RFID signal data to obtain path loss data; Step S3: establishing a positioning network model based on Bi-LSTM technology, inputting the RFID signal data and the path loss data into the positioning network model to perform positioning prediction and obtain sample coordinates; Preprocessing the sample signal data includes the following steps: Step S11: constructing an RFID unified data format, extracting data from the sample signal data according to the RFID unified data format, and establishing a signal path tree based on the extracted data; Step S12: constructing a multidimensional signal structure database based on the signal path tree, and performing data mining on the data in the multidimensional signal structure database to obtain sample signal features; Step S13: using a sliding window technique to sequence the sample signal features to generate RFID signal data.

2. The sample monitoring method based on RFID technology according to claim 1, characterized in that: Performing data mining on the data in the multidimensional signal structure database comprises the following steps: Step S121: Acquire position data from the data in the multidimensional signal structure database, perform position mining on the position data, and obtain a sample position sequence; Step S122: recursively search the sample position sequence and perform a closedness test on the recursive search results to obtain closed paths and non-closed paths; Step S123: Acquire a sample processing path, compare the closed path with the sample processing path, and divide the closed path into a frequent closed path and an infrequent closed path according to the comparison result.

3. The sample monitoring method based on RFID technology according to claim 2, characterized in that: The sample signal features are serialized using a sliding window technique, comprising the following steps: Step S131: extracting the time data groups corresponding to the frequent closed paths, the infrequent closed paths, and the infrequent closed paths from the signal path tree, and forming corresponding spatiotemporal data sets; Step S132: Serialize the spatiotemporal dataset in the form of a time axis to obtain a spatiotemporal sequence dataset; Step S133: Find the minimum frequently closed period in the spatiotemporal sequence data set, and construct a sliding window and a sliding step, wherein the sliding window is larger than the minimum frequently closed period, and the sliding step is smaller than the minimum frequently closed section.

4. The sample monitoring method based on RFID technology according to claim 1, characterized in that: Processing the RFID signal data using the path loss prediction model includes the following steps: Step S21, traversing the RFID signal data to obtain an identification data set and a signal data set of the RFID signal data at time t; wherein the signal data set includes signal strength, and the identification data set includes sample attributes, and the sample attributes include sample shape, sample length, and sample name; Step S22: performing sample attribute matching on the identification data group. If the sample attributes match, the path loss value of the signal data group is calculated using the path loss prediction model.

5. The sample monitoring method based on RFID technology according to claim 1, characterized in that: A positioning network model is established based on Bi-LSTM technology, including: Establishing an input layer, wherein the input layer is used to obtain RFID signal data and path loss data; Establishing a preprocessing layer, the preprocessing layer is used to preprocess the RFID signal data and generate a feature sequence; Establishing a Bi-LSTM layer, wherein the Bi-LSTM layer is used to extract features from the feature sequence; The Bi-LSTM layer includes a forward feature extraction layer and a reverse feature extraction layer. The forward feature extraction layer is used to perform forward feature extraction on the feature sequence in combination with the path loss data to obtain path loss features; the reverse feature extraction layer is used to perform reverse feature extraction on the feature sequence to obtain path features. Establishing a pattern recognition layer, wherein the pattern recognition layer is used to perform pattern recognition on the path features to generate a path pattern; Establishing a feature splicing layer, wherein the feature splicing layer is used to splice the path loss feature with the path feature based on the path pattern and using an attention mechanism to obtain a spliced ​​feature; A DNN layer is established, where the DNN layer is used to perform positioning prediction based on the splicing features and generate sample coordinates.

6. The sample monitoring method based on RFID technology according to claim 5, characterized in that: Inputting the RFID signal data and the path loss data into the positioning network model for positioning prediction includes the following steps: Step S31: pre-process the RFID signal data to generate a path feature sequence, and perform point-to-point splicing of the path feature sequence with the path loss data to generate a path loss feature sequence; Step S32: using the forward feature extraction layer to extract loss features from each sequence point in the path loss feature sequence, and sequentially performing feature recognition on the loss features of each sequence point to obtain a path loss feature; using the reverse feature extraction layer to extract path features from each sequence point in the path feature sequence, and sequentially performing feature recognition on the path features of each sequence point to obtain a path feature; Step S33: performing a frequent closedness test on the path feature, and generating a path pattern according to the frequent closedness test result; wherein the path pattern includes a frequent closed pattern, a non-frequent closed pattern, and a non-closed pattern; Step S34: If the path pattern is a frequent closed pattern, the path loss feature is concatenated with the path feature to obtain a standard concatenated feature, and the standard concatenated feature is compared with the sample processing path to generate a standard path feature; if the path pattern is an infrequent closed pattern or a non-closed pattern, the path loss feature is concatenated with the path feature layer by layer using an attention mechanism to obtain a non-standard concatenated feature; Step S35: Use the DNN layer to perform positioning prediction on the standard splicing features to generate standard sample coordinates; extract several features from the non-standard splicing features as starting features, and use the DNN layer to perform positioning prediction on the RFID signal data based on the starting features to generate non-standard sample coordinates.

7. The sample monitoring system based on RFID technology is characterized by: For implementing the sample monitoring method based on RFID technology according to any one of claims 1 to 6, the sample monitoring system comprises: A preprocessing module is used to obtain sample signal data and preprocess the sample signal data to obtain RFID signal data; a loss processing module, configured to establish a path loss prediction model, and process the RFID signal data using the path loss prediction model to obtain path loss data; The positioning prediction module is used to establish a positioning network model based on Bi-LSTM technology, input the RFID signal data and the path loss data into the positioning network model to perform positioning prediction and obtain sample coordinates.

8. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the sample monitoring method based on RFID technology as claimed in any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the sample monitoring method based on RFID technology as claimed in any one of claims 1 to 6 is implemented.

Images