Adhesive tape production management system and method based on knowledge graph and RFID technology
The tape production management system based on knowledge graph and RFID technology has solved the problem of low management efficiency in tape production and transportation, and has realized real-time monitoring and anomaly detection of tape location, status and environment, thereby improving management efficiency.
Patent Information
- Application Number
- CN202311223362.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-09-21
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2043-09-21
AI Technical Summary
Existing technologies have failed to effectively manage the production and transportation of adhesive tapes, resulting in low management efficiency.
The tape production management system adopts knowledge graph and RFID technology. Through RFID readers, sensor modules, data processing modules, data analysis modules, knowledge graph database and mobile terminal, it monitors tape production and logistics transportation in real time and performs anomaly detection and early warning.
It improves the efficiency of tape production management, enables real-time monitoring of tape location, status, and environmental conditions, and allows for timely detection and handling of abnormal situations.
Smart Images

Figure CN117290516B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the application field of knowledge graphs, and in particular to a tape production management system and method based on knowledge graphs and RFID technology. Background Art
[0002] With the development of economy and the progress of science and technology, China has become a major processing and consumption country in the world's adhesive industry. Combined with many years of supporting experience, it has grown at an annual rate of 16%. In particular, adhesive tapes, protective films and self-adhesive labels are widely used in electronics, communications, packaging, construction, papermaking, woodworking, aerospace, automobiles, textiles, metallurgy, machinery manufacturing, medical industries, etc. The adhesive industry has become the most dynamic and important industry in my country's chemical industry;
[0003] There are many types of adhesive tapes, including BOPP tape, double-sided tape, masking tape, PVC electrical tape, cloth tape, aluminum foil tape, foam tape, washi tape, stationery tape, protective tape, kraft paper tape, etc. The storage of tapes needs to consider the influence of temperature and humidity, but the existing technology does not have a management plan for the production and transportation of tapes. Summary of the invention
[0004] In order to solve the above problems, the purpose of the present invention is to provide a tape production management system and method based on knowledge graph and RFID technology to effectively improve the efficiency of tape production management.
[0005] To achieve the above object, the present invention adopts the following technical solutions:
[0006] A tape production management system based on knowledge graph and RFID technology, including an RFID reader / writer, a sensor module, a data processing module, a data analysis module, a knowledge graph library, a mobile terminal and an early warning module; the RFID reader / writer is deployed on the tape production line and the logistics transportation line to collect RFID tag information set on the finished product, and interacts with the knowledge graph after being processed by the data processing module; the data analysis module analyzes and monitors the location, status and environmental conditions of the object through the sensor module based on the data processed by the data processing module and the data in the knowledge graph library, and sends an early warning message to the mobile terminal through the early warning module if an abnormality is found; the knowledge base data can be retrieved through the mobile terminal to intuitively view and manage the information of the object.
[0007] Furthermore, the knowledge graph library is constructed as follows:
[0008] Obtain data related to tape raw materials and finished products, including supplier information, production batches, production dates, quality inspection results, and logistics information;
[0009] Clean and organize the collected data, remove duplicate, missing or erroneous data, and standardize the data;
[0010] Based on the collected data, a knowledge graph model of tape raw materials and finished products is constructed.
[0011] Furthermore, the knowledge graph model uses BiLSTM-CRF for entity recognition, as follows:
[0012] Obtain and annotate a text dataset with entities related to tape raw materials and finished products, and divide it into training, validation, and test sets;
[0013] Convert text datasets into digital feature representations;
[0014] Build a sequence labeling model based on BiLSTM-CRF:
[0015] BiLSTM layer:
[0016] Forward propagation: h_i^f = LSTM(x_i,h_{i-1}^f)
[0017] Backward propagation: h_i^b=LSTM(x_i,h_{i+1}^b)
[0018] Output layer: h_i = [h_i^f, h_i^b]
[0019] CRF layer:
[0020] Transition probability: P(y_i=s|y_{i-1}=r)=exp(score(r,s)) / Z
[0021] The score of each label: score(r,s)=W_r*h_i+T_r,s
[0022] Loss function: L = -log(P(y|x)) + λ||θ||^2
[0023] Where x_i represents the input feature vector, h_i represents the output of the BiLSTM layer, h_i^f represents the output of the forward propagation, h_i^b represents the output of the backward propagation, y_i represents the label sequence, W_r and T_r,s are model parameters, score(r,s) represents the score transferred from label r to label s, Z is the normalization factor, θ is the model parameter, and λ is the weight of the regularization term;
[0024] The model is trained using the training set, and the model parameters are optimized by maximizing the log-likelihood function. Further, the model parameters are optimized by maximizing the log-likelihood function as follows:
[0025] (1) Define the model parameters W_r and T_r,s, assume that the observed data are x1, x2, ..., xn, and the probability density function of the probability model is f(x; θ);
[0026] (2) In the entity recognition task, the log-likelihood function can be obtained by calculating the log-probability of the label sequence. The specific formula is as follows:
[0027] L(θ)=f(x1;θ)*f(x2;θ)*...*f(xn;θ);
[0028] log L(θ)=log[f(x1;θ)]+log[f(x2;θ)]+...+log[f(xn;θ)];
[0029] Where n represents the number of training samples, and θ represents the parameter vector θ=(W_r,T_r,s).
[0030] (3) In the entity recognition task, negative log likelihood can be used as the loss function. The specific formula is as follows:
[0031] NLL(θ)=-log L(θ)
[0032] (4) Use stochastic gradient descent to maximize the log-likelihood function, that is, minimize the loss function. The specific formula is as follows:
[0033] Give θ an initial value and randomly select a training sample. For the selected sample, update the parameters by making its gradient descent:
[0034]
[0035] Among them, θ_new represents the updated parameter value, θ_old represents the current parameter value, and learning_rate represents the learning rate. Represents the gradient of the loss function with respect to the parameters;
[0036] Iterate multiple times, each time using a batch of training samples to update the parameters until the preset conditions are met.
[0037] Furthermore, the RFID reader and sensor are arranged in the assembly line, packaging line and warehouse of tape production, an RFID tag is attached to the product packaging, and information related to the tag and the item is stored in the tag. The sensor includes a temperature sensor, a humidity sensor and a vibration sensor; the RFID reader and sensor transmits the collected data to the data processing module to process the data in real time.
[0038] Furthermore, the data analysis module performs anomaly detection based on the support vector machine anomaly detector according to the information obtained by the sensor and the RFID reader and the data of the knowledge graph library.
[0039] Furthermore, the support vector machine anomaly detector is constructed as follows:
[0040] (1) Obtain historical RFID reader and sensor data, annotate them, and obtain a training set;
[0041] (2) Based on the training set, train the support vector machine and find the vector x containing all the training samples i The constraints are set, and the samples not included in the constraints are regarded as outliers, that is,
[0042]
[0043] ||x i -c|| 2 ≤R 2 +ξ i ≥0
[0044] Where i=1,...,m, m is the number of training samples, ξ is the error, c is the center of the constraint, R is the distance of the constraint, and C is the error adjustment parameter;
[0045] (3) Convert the above constraints into Lagrangian functions:
[0046]
[0047] Among them, α, β are Lagrange multipliers;
[0048] (4) Based on
[0049]
[0050] α≥0,β≥0
[0051] Further, we get the Lagrange dual formula:
[0052]
[0053] 0≤α i ≤C
[0054] where j = 1, ..., m, For x i The transpose of
[0055] (5) Based on The KKT optimization conditions of the optimization problem give the solutions for R and c:
[0056]
[0057]
[0058] (7) When the real-time monitoring sample vector x satisfies , it is abnormal.
[0059] A tape production management method based on knowledge graph and RFID technology includes the following steps:
[0060] Step S1: Obtain data related to the raw materials and finished products of the adhesive tape, pre-process them, and then construct a knowledge graph model of the raw materials and finished products of the adhesive tape and store it in the knowledge graph library;
[0061] Step S2: RFID readers and sensors are installed in the assembly line, packaging line and warehouse of the tape production, RFID tags are attached to the product packaging, and information related to the tag and the item is stored in the tag;
[0062] Step S3: Collecting information related to the item and environmental data in real time through RFID readers and sensors;
[0063] Step S4: transmitting the collected data to the data processing module, processing the data in real time, and transmitting the processed data to the data analysis module;
[0064] Step S5: The data analysis module performs anomaly detection based on the support vector machine anomaly detector according to the information obtained by the sensor and RFID reader and the data of the knowledge graph library. If an anomaly is detected, an early warning message is sent to the mobile terminal through the early warning module.
[0065] The present invention has the following beneficial effects:
[0066] The present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0067] Figure 1 For the present invention;
[0068] Figure 2 In one embodiment of the present invention. DETAILED DESCRIPTION
[0069] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:
[0070] refer to Figure 1The present invention provides a tape production management system based on knowledge graph and RFID technology, including an RFID reader / writer, a sensor module, a data processing module, a data analysis module, a knowledge graph library, a mobile terminal and an early warning module; the RFID reader / writer is deployed on the tape production line and the logistics transportation line to collect RFID tag information set on the finished product, and interacts with the knowledge graph after being processed by the data processing module; the data analysis module analyzes and monitors the location, status and environmental conditions of the object through the sensor module according to the data processed by the data processing module and the data in the knowledge graph library, and sends an early warning message to the mobile terminal through the early warning module if an abnormality is found; the knowledge base data can be retrieved through the mobile terminal to intuitively view and manage the information of the object.
[0071] In this embodiment, preferably, the knowledge graph library is constructed as follows:
[0072] Obtain data related to tape raw materials and finished products, including supplier information, production batches, production dates, quality inspection results, and logistics information;
[0073] Clean and organize the collected data, remove duplicate, missing or erroneous data, and standardize the data;
[0074] Based on the collected data, a knowledge graph model of tape raw materials and finished products is constructed.
[0075] In this embodiment, the knowledge graph model uses BiLSTM-CRF for entity recognition, as follows:
[0076] Obtain and annotate a text dataset with entities related to tape raw materials and finished products, and divide it into training, validation, and test sets;
[0077] Preferably, for text data related to the raw materials and finished products of the tape, the BIO (Begin, Inside, Outside) marking scheme can be used to mark the entities. The specific marking scheme is as follows:
[0078] B (Begin): Indicates the starting position of the entity.
[0079] I (Inside): Indicates the middle position of the entity.
[0080] O (Outside): Indicates a non-entity location.
[0081] Let's take an example to illustrate the application of the marking scheme:
[0082] Suppose there is the following sentence: Tape is composed of polyethylene and adhesive.
[0083] The corresponding entity tag sequence is: OOOO B-Material I-Material OOOO B-MaterialI-Material OO
[0084] Among them, "Polyethylene" and "Adhesive" are marked as entities and annotated using the BIO marking scheme. The beginning of "Polyethylene" is marked as B-Material, and the subsequent positions are marked as I-Material. Similarly, the beginning of "Adhesive" is also marked as B-Material, and the subsequent positions are marked as I-Material. The other positions are marked as O, indicating non-entities.
[0085] The formula for the BIO labeling scheme is as follows:
[0086] B-{Entity}: Indicates the starting position of the entity.
[0087] I-{Entity}: Indicates the middle position of the entity.
[0088] O: Indicates a non-entity location.
[0089] In the training data set, different entity types can be defined according to actual conditions, such as Material, Product, etc., and the entities can be marked as B-Material, I-Material, B-Product, I-Product, etc.
[0090] The RFID tag stores the item's unique identifier (UID).
[0091] -Read the tag information through the RFID reader and writer to obtain the UID of the item.
[0092] -According to the UID, associate the RFID data with the item entity in the knowledge graph.
[0093] Convert text datasets into digital feature representations;
[0094] Build a sequence labeling model based on BiLSTM-CRF:
[0095] BiLSTM layer:
[0096] Forward propagation: h_i^f = LSTM(x_i,h_{i-1}^f)
[0097] Backward propagation: h_i^b=LSTM(x_i,h_{i+1}^b)
[0098] Output layer: h_i = [h_i^f, h_i^b]
[0099] CRF layer:
[0100] Transition probability: P(y_i=s|y_{i-1}=r)=exp(score(r,s)) / Z
[0101] The score of each label: score(r,s)=W_r*h_i+T_r,s
[0102] Loss function: L = -log(P(y|x)) + λ||θ||^2
[0103] Where x_i represents the input feature vector, h_i represents the output of the BiLSTM layer, h_i^f represents the output of the forward propagation, h_i^b represents the output of the backward propagation, y_i represents the label sequence, W_r and T_r,s are model parameters, score(r,s) represents the score transferred from label r to label s, Z is the normalization factor, θ is the model parameter, and λ is the weight of the regularization term;
[0104] The model is trained using the training set, and the model parameters are optimized by maximizing the log-likelihood function. In this embodiment, the model parameters are optimized by maximizing the log-likelihood function, as follows:
[0105] (1) Define the model parameters W_r and T_r,s, assume that the observed data are x1, x2, ..., xn, and the probability density function of the probability model is f(x; θ);
[0106] (2) In the entity recognition task, the log-likelihood function can be obtained by calculating the log-probability of the label sequence. The specific formula is as follows:
[0107] L(θ)=f(x1;θ)*f(x2;θ)*...*f(xn;θ);
[0108] log L(θ)=log[f(x1;θ)]+log[f(x2;θ)]+...+log[f(xn;θ)];
[0109] Where n represents the number of training samples, and θ represents the parameter vector θ=(W_r,T_r,s).
[0110] (3) In the entity recognition task, negative log likelihood can be used as the loss function. The specific formula is as follows:
[0111] NLL(θ)=-log L(θ)
[0112] (4) Use stochastic gradient descent to maximize the log-likelihood function, that is, minimize the loss function. The specific formula is as follows:
[0113] Give θ an initial value and randomly select a training sample. For the selected sample, update the parameters by making its gradient descent:
[0114]
[0115] Among them, θ_new represents the updated parameter value, θ_old represents the current parameter value, and learning_rate represents the learning rate. Represents the gradient of the loss function with respect to the parameters;
[0116] Iterate multiple times, each time using a batch of training samples to update the parameters until the preset conditions are met.
[0117] In this embodiment, RFID readers and sensors are set up in the assembly line, packaging line and warehouse of tape production. RFID tags are attached to the product packaging, and information related to the tag and the item is stored in the tag. The sensors include temperature sensors, humidity sensors and vibration sensors; the RFID readers and sensors transmit the collected data to the data processing module to process the data in real time.
[0118] In this embodiment, the data analysis module performs anomaly detection based on the support vector machine anomaly detector according to the information obtained by the sensor and the RFID reader and the data of the knowledge graph library.
[0119] In this embodiment, the support vector machine anomaly detector is constructed as follows:
[0120] (1) Obtain historical RFID reader and sensor data, annotate them, and obtain a training set;
[0121] (2) Based on the training set, train the support vector machine and find the vector x containing all the training samples i The constraints are set, and the samples not included in the constraints are regarded as outliers, that is,
[0122]
[0123] ||x i -c|| 2 ≤R 2 +ξ i ≥0
[0124] Where i=1,...,m, m is the number of training samples, ξ is the error, c is the center of the constraint, R is the distance of the constraint, and C is the error adjustment parameter;
[0125] (3) Convert the above constraints into Lagrangian functions:
[0126]
[0127] Among them, α, β are Lagrange multipliers;
[0128] (4) Based on
[0129]
[0130] α≥0,β≥0
[0131] Further, we get the Lagrange dual formula:
[0132]
[0133] 0≤α i ≤C
[0134] where j = 1, ..., m, For x i The transpose of
[0135] (5) Based on The KKT optimization conditions of the optimization problem give the solutions for R and c:
[0136]
[0137]
[0138] (8) When the real-time monitoring sample vector x satisfies , it is abnormal.
[0139] refer to Figure 2 In this embodiment, a method for managing tape production based on knowledge graph and RFID technology is also provided, comprising the following steps:
[0140] Step S1: Obtain data related to the raw materials and finished products of the adhesive tape, pre-process them, and then construct a knowledge graph model of the raw materials and finished products of the adhesive tape and store it in the knowledge graph library;
[0141] Step S2: RFID readers and sensors are installed in the assembly line, packaging line and warehouse of the tape production, RFID tags are attached to the product packaging, and information related to the tag and the item is stored in the tag;
[0142] Step S3: Collecting information related to the item and environmental data in real time through RFID readers and sensors;
[0143] Step S4: transmitting the collected data to the data processing module, processing the data in real time, and transmitting the processed data to the data analysis module;
[0144] Step S5: The data analysis module performs anomaly detection based on the support vector machine anomaly detector according to the information obtained by the sensor and RFID reader and the data of the knowledge graph library. If an anomaly is detected, an early warning message is sent to the mobile terminal through the early warning module.
[0145] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0146] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0147] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0148] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0149] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention still belongs to the protection scope of the technical solution of the present invention.
Claims
1. A tape production management system based on knowledge graph and RFID technology, characterized in that: It includes an RFID reader / writer, a sensor module, a data processing module, a data analysis module, a knowledge graph library, a mobile terminal and an early warning module; the RFID reader / writer is deployed on the tape production line and the logistics transportation line to collect RFID tag information set on the finished product, and interacts with the knowledge graph after being processed by the data processing module; the data analysis module analyzes and monitors the location, status and environmental conditions of the item through the sensor module based on the data processed by the data processing module and the data in the knowledge graph library, and sends an early warning message to the mobile terminal through the early warning module if an abnormality is found; the knowledge base data can be retrieved through the mobile terminal to intuitively view and manage the information of the item; The data analysis module is based on a support vector machine anomaly detector, and performs anomaly detection according to the information obtained by the sensor and the RFID reader and the data of the knowledge graph library; the support vector machine anomaly detector is constructed as follows: (1) Obtain historical RFID reader and sensor data, annotate them, and obtain a training set; (2) Based on the training set, train the support vector machine and find the vector containing all training samples The constraints of the support vector machine are: ; ; in, i=1,...,m , m is the number of training samples, is the error, c is the center of the constraint, R is the distance of the constraint, and C is the error adjustment parameter; (3) Convert the above constraints into Lagrangian functions: ; in, is the Lagrange multiplier; (4) Based on ; ; Further, we get the Lagrange dual formula: ; ; where j=1,...,m, for The transpose of (5) Based on The KKT optimization conditions of the optimization problem give the solutions for R and c: ; When the real-time monitoring sample vector x satisfies , it is abnormal.
2. The adhesive tape production management system based on knowledge graph and RFID technology according to claim 1 is characterized in that: The knowledge graph library is constructed as follows: Obtain data related to tape raw materials and finished products, including supplier information, production batches, production dates, quality inspection results, and logistics information; Clean and organize the collected data, remove duplicate, missing or erroneous data, and standardize the data; Based on the collected data, a knowledge graph model of tape raw materials and finished products is constructed.
3. The adhesive tape production management system based on knowledge graph and RFID technology according to claim 2 is characterized in that: The knowledge graph model uses BiLSTM-CRF for entity recognition, as follows: Obtain and annotate a text dataset with entities related to tape raw materials and finished products, and divide it into training, validation, and test sets; Convert text datasets into digital feature representations; Build a sequence labeling model based on BiLSTM-CRF: BiLSTM layer: Forward propagation: h_i^f = LSTM(x_i, h_{i-1}^f); Backward propagation: h_i^b = LSTM(x_i, h_{i+1}^b); Output layer: h_i = [h_i^f, h_i^b]; CRF layer: Transition probability: P(y_i = s|y_{i-1} = r) = exp(score(r, s)) / Z; The score of each label: score(r, s) = W_r * h_i + T_r,s; Loss function: L = -log(P(y|x)) + λ||θ||^2; Where x_i represents the input feature vector, h_i represents the output of the BiLSTM layer, h_i^f represents the output of the forward propagation, h_i^b represents the output of the backward propagation, y_i represents the label sequence, W_r and T_r,s are model parameters, score(r, s) represents the score transferred from label r to label s, Z is the normalization factor, θ is the model parameter, and λ is the weight of the regularization term; The model is trained using the training set and the model parameters are optimized by maximizing the log-likelihood function.
4. The adhesive tape production management system based on knowledge graph and RFID technology according to claim 3 is characterized in that: The model parameters are optimized by maximizing the log-likelihood function, as follows: (1) Define the parameters of the model W_r and T_r,s, assume that the observed data are x1, x2, ..., xn, and the probability density function of the probability model is f(x; θ); (2) In the entity recognition task, the log-likelihood function can be obtained by calculating the log-probability of the label sequence. The specific formula is as follows: L(θ) = f(x1; θ) * f(x2; θ) * ... * f(xn; θ); log L(θ) = log[f(x1; θ)] + log[f(x2; θ)] + ... + log[f(xn; θ)]; Where n represents the number of training samples, θ represents the parameter vector θ=(W_r, T_r,s); (3) In the entity recognition task, negative log-likelihood can be used as the loss function. The specific formula is as follows: NLL(θ) = -log L(θ); (4) Use stochastic gradient descent to maximize the log-likelihood function, that is, minimize the loss function. The specific formula is as follows: Give θ an initial value and randomly select a training sample. For the selected sample, update the parameters by making its gradient descent: θ_new=θ_old-learning_rate*∇_θloss(θ_old); Among them, θ_new represents the updated parameter value, θ_old represents the current parameter value, learning_rate represents the learning rate, and ∇_θloss(θ_old) represents the gradient of the loss function with respect to the parameter; Iterate multiple times, each time using a batch of training samples to update the parameters until the preset conditions are met.
5. The adhesive tape production management system based on knowledge graph and RFID technology according to claim 1 is characterized in that: The RFID reader and sensor are set up in the assembly line, packaging line and warehouse of tape production. RFID tags are attached to product packaging, and information related to the tags and items is stored in the tags. The sensors include temperature sensors, humidity sensors and vibration sensors; the RFID reader and sensor transmit the collected data to the data processing module to process the data in real time.
6. A management method for a tape production management system based on a knowledge graph and RFID technology according to any one of claims 1 to 5, characterized in that: The following steps are involved: Step S1: Obtain data related to the raw materials and finished products of the adhesive tape, pre-process them, and then construct a knowledge graph model of the raw materials and finished products of the adhesive tape and store it in the knowledge graph library; Step S2: RFID readers and sensors are installed in the assembly line, packaging line and warehouse of the tape production, RFID tags are attached to the product packaging, and information related to the tag and the item is stored in the tag; Step S3: Collecting information related to the item and environmental data in real time through RFID readers and sensors; Step S4: transmitting the collected data to the data processing module, processing the data in real time, and transmitting the processed data to the data analysis module; Step S5: The data analysis module performs anomaly detection based on the support vector machine anomaly detector according to the information obtained by the sensor and RFID reader and the data of the knowledge graph library. If an anomaly is detected, an early warning message is sent to the mobile terminal through the early warning module.
