Storage method of electronic reel material

Abstract

The invention relates to the field of electronic reel materials, in particular to an electronic reel material storage method. The method comprises the steps that when an electronic reel material is transferred to a warehousing table, a sensor senses the existence of the electronic reel material according to an in-place signal, and a recognition device obtains material disc information; when a grabbing device carries the electronic reel material to a storage position, a sensor senses the action completion condition according to an action signal, and temporarily stores the information of the reel material in a storage system; and when the electronic reel material is placed at a specified storage location, the storage system archives the material tray information and binds the material tray information with the specified storage location in response to the fact that the sensor senses that the action is completed according to the action signal.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of electronic reel material, in particular to a storage method of electronic reel material. BACKGROUND

[0002] In the electronic manufacturing industry, material management is crucial to improve production efficiency and reduce costs. With the development of the electronic product manufacturing industry, the miniaturization trend of electronic components has made the use of electronic reel material more and more widespread. Electronic reel material, as a common form of electronic material packaging, is usually used to store and transport small electronic components such as resistors, capacitors, etc. Due to the characteristics of similar appearance, similar weight, and numerous types of such materials, it poses a series of challenges in storage and management.

[0003] The traditional storage method of electronic reel material often relies on manual operation and simple barcode recognition. This method requires manual entry of material in and out of the warehouse information, which requires a large amount of manpower and time, resulting in low management efficiency. Moreover, human errors can occur during the information input and barcode recognition process of electronic reel material, which can lead to inaccurate or missing data, affecting the accuracy of electronic reel material information storage and making it difficult to ensure consistency between physical materials and reel information.

[0004] Therefore, an optimized storage solution for electronic reel material is desired. SUMMARY

[0005] The present application is made in consideration of the above problems. One object of the present application is to provide a storage method of electronic reel material.

[0006] Embodiments of the present application provide a storage method of electronic reel material, which comprises: When the electronic reel material is transported to the warehouse station, the sensor senses the presence of the electronic reel material according to the arrival signal, and the identification device obtains the reel information; When the electronic reel material is transported to the storage location by the grabbing device, the sensor senses the completion of the action according to the action signal, and the reel information is temporarily stored in the storage system; When the electronic reel material is placed in the designated storage location, in response to the sensor sensing the completion of the action according to the action signal, the storage system archives the reel information and binds it with the designated storage location.

[0007] For example, according to the storage method of electronic reel material of the embodiments of the present application, during the storage process of the electronic reel material, the sensor is always in a monitoring state, monitoring all the operating positions of the electronic reel material, ensuring that the arrival position of the electronic reel material has a corresponding signal provided.

[0008] For example, in the electronic reel material storage method according to an embodiment of this application, there is a signal feedback in any storage state of the electronic reel material to report the state of the electronic reel material, and the state of the electronic reel material is assigned to the reel information to manage the physical electronic reel material, the signal and the reel information as a fixed correspondence.

[0009] For example, in the electronic reel material storage method according to an embodiment of this application, the reel information includes information about the electronic reel itself, information about the relationships between the electronic reels, location information of the electronic reel, material status attribute information of the electronic reel, and management information of the electronic reel.

[0010] For example, in the electronic reel material storage method according to an embodiment of this application, a lower-level computer controls a plurality of said sensors to collect signals from the sensors indicating that the electronic reel material is located at a mechanical position corresponding to the sensor, and processes the signals into a sequential signal stream, wherein the sequential signal stream is a large-small closed-loop method.

[0011] For example, in the electronic reel material storage method according to an embodiment of this application, a host computer generates a record corresponding to the mechanical position of the electronic reel based on a signal triggered by a sensor of the slave computer, and stores it in the storage system.

[0012] For example, the method for storing electronic reel stock according to an embodiment of this application further includes: Obtain information about the first and second material trays; Acquire sensor signals collected by the second sensor; Semantic encoding is performed on the first tray information and the second tray information to obtain the semantic encoding feature vector of the first tray information and the semantic encoding feature vector of the second tray information. Calculate the semantic difference feature vector between the semantic encoding feature vector of the first material tray information and the semantic encoding feature vector of the second material tray information; The sensor signal is input into a signal feature extractor based on a convolutional neural network model to obtain a semantic feature vector of the sensor signal waveform. Based on the semantic feature vector of the sensor signal waveform, the semantic difference feature vector is compensated by backpropagation of metric information to obtain an optimized semantic difference feature vector; Calculate the cosine similarity between the optimized semantic difference feature vector and the semantic difference feature vector; In response to the cosine similarity being greater than or equal to a preset threshold, an abnormal state update warning is generated.

[0013] For example, according to an embodiment of the present application, a method for storing electronic reel stock includes, based on the semantic feature vector of the sensor signal waveform, performing backpropagation expression compensation of metric information on the semantic difference feature vector to obtain an optimized semantic difference feature vector, including: Calculate the position-based difference between the semantic difference feature vector and the sensor signal waveform semantic feature vector to obtain the tray information-sensor signal semantic shift position-based difference representation vector; After concatenating the semantic difference feature vector and the sensor signal waveform semantic feature vector into a joint feature vector, the joint feature vector is input into a one-dimensional convolutional layer to obtain the material tray information-sensor signal semantic shift convolutional encoding representation vector. After linearly embedding and modulating the semantic differential feature vector into a semantic differential feature vector linear modulation vector using an embedding matrix and a bias vector, the positional difference between the semantic differential feature vector linear modulation vector and the sensor signal waveform semantic feature vector is calculated to obtain the tray information-sensor signal semantic shift modulation differential representation vector. Using the position difference representation vector of the material tray information-sensor signal semantic shift as the value vector, the convolutional coding representation vector of the material tray information-sensor signal semantic shift, and the modulation differential representation vector of the material tray information-sensor signal semantic shift as the query vector and key vector, the material tray information-sensor signal semantic shift representation vector of position difference, the material tray information-sensor signal semantic shift convolutional coding representation vector, and the material tray information-sensor signal semantic shift modulation differential representation vector are input into a shift corrector based on a pseudo-converter structure to obtain the material tray information-sensor signal semantic multi-scale hierarchical fusion shift modulation weight vector. The optimized semantic differential feature vector is obtained by multiplying the position-based shift modulation weight vector of the material tray information-sensor signal semantic multi-scale hierarchical fusion with the semantic differential feature vector.

[0014] For example, according to an embodiment of the present application, the method for storing electronic reel material includes using the reel information-sensor signal semantic shift by position difference representation vector as a value vector, the reel information-sensor signal semantic shift convolutional coding representation vector, and the reel information-sensor signal semantic shift modulation differential representation vector as a query vector and a key vector, respectively. The reel information-sensor signal semantic shift by position difference representation vector, the reel information-sensor signal semantic shift convolutional coding representation vector, and the reel information-sensor signal semantic shift modulation differential representation vector are input into a shift corrector based on a pseudo-converter structure to obtain a reel information-sensor signal semantic multi-scale hierarchical fusion shift modulation weight vector, including: Multiply the transpose of the material tray information-sensor signal semantic shift convolutional encoding representation vector and the material tray information-sensor signal semantic shift modulation differential representation vector, and then divide the transpose by the square root of the length of the material tray information-sensor signal semantic shift modulation differential representation vector to obtain the material tray information-sensor signal semantic shift multidimensional information interaction representation matrix. The Softmax function is used to perform soft maximum value normalization on the material tray information-sensor signal semantic shift multidimensional information interaction representation matrix to obtain the material tray information-sensor signal semantic shift multidimensional information interaction probabilistic matrix. The matrix multiplication between the position difference representation vector of the material tray information-sensor signal semantic shift and the multidimensional information interaction probability matrix of the material tray information-sensor signal semantic shift is calculated and then processed by the ReLU activation function to obtain the multi-scale hierarchical fusion shift modulation weight vector of the material tray information-sensor signal semantics.

[0015] The electronic reel material storage method according to an embodiment of this application collects information from a first reel and a second reel, and acquires sensor signals collected by a second sensor. Then, it introduces data processing and semantic understanding algorithms based on artificial intelligence and deep learning in the backend to capture semantic difference features between the first and second reel information, and analyzes the sensor signals. This allows for the detection of anomalies in the reel information update process using the alignment between the semantic difference features of the reel information and the sensor signal features. Upon detecting an anomaly, a status update anomaly warning is issued to ensure consistency between the physical reel and the reel information, thereby improving the accuracy and reliability of electronic reel material storage. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings of the embodiments of this application will be briefly described below. Obviously, the drawings described below only relate to some embodiments of this application, and are not intended to limit this application.

[0017] Figure 1 A schematic diagram showing the correspondence between the system and the physical object in the embodiments of this application is shown.

[0018] Figure 2 A schematic diagram of the material movement process in the coil is shown in an embodiment of this application.

[0019] Figure 3 A schematic diagram showing the correspondence between the host computer, the slave computer, and the mechanical positions in an embodiment of this application is provided.

[0020] Figure 4 A flowchart illustrating the method for storing electronic reel material in an embodiment of this application is shown.

[0021] Figure 5 A flowchart illustrating further steps included in the method for storing electronic reel material according to an embodiment of this application is shown.

[0022] Figure 6 A schematic diagram of the structure of the electronic reel material storage system in an embodiment of this application is shown.

[0023] Figure 7 The diagram illustrates an application scenario of the electronic reel material storage method according to an embodiment of this application. Detailed Implementation

[0024] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are also within the scope of protection of this application.

[0025] The terminology used in this specification is that which is currently widely used in the art in consideration of the functionality of this application; however, these terms may vary depending on the intent of a person skilled in the art, precedent, or new technology in the art. Furthermore, specific terms may be chosen, and in such cases, their detailed meanings will be described in the detailed description of this application. Therefore, the terms used in this specification should not be construed as simple names, but rather based on their meanings and the overall description of this application.

[0026] While this application makes various references to certain modules of the systems according to embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The modules described are merely illustrative, and different aspects of the systems and methods may use different modules.

[0027] This application uses flowcharts to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, various steps can be processed in reverse order or simultaneously, as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0028] Furthermore, the application architecture diagrams in this application are for the purpose of more clearly illustrating the technical solutions in this application and do not constitute a limitation on the technical solutions provided in this application. Of course, the technical solutions provided in this application are also applicable to similar problems for other application architectures and business applications.

[0029] The following examples or embodiments illustrate a non-limiting method for storing electronic reel stock according to at least one embodiment of this application. As described below, different features in these specific examples or embodiments can be combined with each other without conflict to obtain new examples or embodiments, which are also within the scope of protection of this application.

[0030] Reel packaging is a common packaging method used by electronics manufacturing companies. Its characteristics include: 1. Similar appearance; 2. Similar weight; 3. Diverse product categories; 4. Mostly small items; 5. Composed of reels and discs; 6. Strict control over material usage; 7. Strict control over consistency between records and physical inventory; 8. Complex quantity control; 9. Unused materials are repeatedly reused until exhausted. Given these characteristics and management methods, the core control lies in ensuring consistency between records and physical inventory.

[0031] This application provides a method for storing electronic reel materials, and the implementation of this method is described from three aspects.

[0032] First, a one-to-one correspondence is maintained between the physical object and the system at all stages, meaning that the system and the physical object maintain consistency at any point in time (material information includes information about itself, its relationships, location, status, and control). For example... Figure 1 As shown, the system maintains consistency with the physical object in any actual location and in any process.

[0033] Second, the entire movement of the coiled material is accompanied by continuous changes in mechanical position, sensors, and signals. For example... Figure 2 As shown, the entire process is divided into three parts: mechanical position (i.e., the actual position of the coiled material), lower-level machine (i.e., information collected by sensors and sent to the lower-level machine), and signal flow (i.e., all signals flow in a predetermined order and are controllable in a closed loop). In the mechanical position part, when the coil is at position 1, the lower-level machine obtains a signal through sensor 1. This signal is directly generated because there is material at position 1, meaning there is an absolute one-to-one correspondence between position 1 and sensor 1. Similarly, position N corresponds to sensor N, and position N+1 corresponds to sensor N+1, thus ensuring a one-to-one correspondence between the position and the sensor, i.e., a one-to-one correspondence between the mechanical position and the lower-level machine signal. The lower-level machine obtains the signal and processes it internally in an absolutely sequential manner, ensuring that the signal is not disordered. Furthermore, for convenient closed-loop processing, a large and small closed-loop approach is used for signal flow processing. (Continue to refer to...) Figure 2The path from X1 to X2 is G1 and -G1, where G1 is the forward signal flow and -G1 is the reverse signal flow. G1 and -G1 form a small closed loop between X1 and X2, ensuring the sequential and closed-loop transmission of the signal between X1 and X2, thus preventing signal loss. Similarly, the transmission between X2 and XN is similar. Ultimately, there are both small and large closed loops between X1 and XN+1, guaranteeing the sequential and closed-loop transmission of the entire process. This achieves the effect of sequential and closed-loop signal transmission. Since there is a one-to-one correspondence between signal and position, and also a one-to-one correspondence between material and position, there is a one-to-one correspondence between signal and physical object. That is, during the storage of electronic reel material, relevant detection sensors are constantly monitoring all the movement positions of the electronic reel material, ensuring that the corresponding signal is provided upon arrival at the desired position.

[0034] Third, the correspondence between the host computer, the slave computer, and the mechanical positions, such as... Figure 3 As shown. The host computer records information about the material tray (including the material itself, its relationships, location, status, and control information). This information is stored using sensors on the slave computer as trigger points. When the tray is in position 1, the slave computer receives a signal from sensor 1 and uploads it to the host computer as a trigger. The host computer then records the tray information and generates record 1, which is stored in memory. When the tray is in mechanical position N, the slave computer receives a signal from sensor N and uploads it to the host computer. The host computer then modifies record 1, generating a new record N, which is stored in memory. The material position, the slave computer, and the host computer maintain a one-to-one correspondence. Finally, when the host computer receives signal N+1 from the slave computer, it generates record N+1 and archives it. This completes the closed-loop control of one material tray.

[0035] Accordingly, the technical solution of this application proposes a method for storing electronic reel material, such as... Figure 4 As shown, it includes: S510, when the electronic reel material is transferred to the receiving station, the sensor detects the presence of the electronic reel material based on the arrival signal, and the identification device obtains the reel information; S520, when the gripping device moves the electronic reel material to the storage location, the sensor detects the completion of the action based on the action signal, and temporarily stores the reel information in the storage system; S530, when the electronic reel material is placed in the designated storage location, in response to the sensor detecting the completion of the action based on the action signal, the storage system archives the reel information and binds it to the designated storage location.

[0036] During the storage process of the electronic reel stock, the sensor is constantly monitoring all operational positions of the electronic reel stock, ensuring that a corresponding signal is provided when the electronic reel stock reaches its designated position. Simultaneously, signal feedback is provided at any storage state of the electronic reel stock to report its status. The status of the electronic reel stock is assigned to the reel information, thus managing the physical electronic reel stock, signals, and reel information as a fixed correspondence. In this way, by assigning the status of the electronic reel stock to the reel information, the physical electronic reel stock, signals, and reel information can be managed as a fixed correspondence, thereby achieving a more intelligent electronic reel stock storage process.

[0037] The material tray information includes information about the electronic coil itself, information about the relationships between the electronic coils, information about the location of the electronic coil, information about the material status attributes of the electronic coil, and information about the management of the electronic coil.

[0038] Furthermore, the lower-level computer controls multiple sensors to collect signals from the sensors indicating the location of the electronic reel material at the mechanical position corresponding to the sensor, and processes the signals into a sequential signal stream, which is a large-small closed-loop stream. The upper-level computer generates a record corresponding to the mechanical position of the electronic reel based on signals triggered by the sensors from the lower-level computer, and stores this record in the storage system.

[0039] During the aforementioned storage process, the mechanical position of the electronic reel is collected and updated by different sensors. However, during the transfer and storage of the electronic reel material, noise in the sensor signals or other factors may interfere with the updating of the reel information, leading to inaccurate storage of the electronic reel material information and making it difficult to ensure the consistency between the physical material and the reel information.

[0040] In particular, considering that the difference between the second tray information and the first tray information is determined based on the sensor signals collected by the second sensor, it is possible to determine whether there is an anomaly in the tray information update based on the semantic difference features between the second tray information and the first tray information, as well as the alignment between the semantic features of the second sensor signals.

[0041] Based on this, the technical concept of this application is to collect information from a first and a second material tray, and obtain sensor signals collected by a second sensor. Then, in the backend, data processing and semantic understanding algorithms based on artificial intelligence and deep learning are introduced to capture the semantic differences between the first and second material tray information, and the sensor signals are analyzed. In this way, the alignment between the semantic differences in the material tray information and the sensor signal features can be used to detect anomalies in the material tray information update process, and an alert for an anomaly in the status update can be issued when an anomaly is detected, ensuring the consistency between the physical material and the material tray information, thereby helping to improve the accuracy and reliability of electronic reel material storage.

[0042] Accordingly, such as Figure 5 As shown, the electronic reel material storage method further includes: S610, acquiring first reel information and second reel information; S620, acquiring sensor signals collected by a second sensor; S630, performing semantic encoding on the first reel information and the second reel information to obtain semantic encoding feature vectors of the first reel information and the second reel information; S640, calculating a semantic difference feature vector between the semantic encoding feature vectors of the first reel information and the second reel information; S650, inputting the sensor signal into a signal feature extractor based on a convolutional neural network model to obtain a sensor signal waveform semantic feature vector; S660, performing backpropagation expression compensation of metric information on the semantic difference feature vector based on the sensor signal waveform semantic feature vector to obtain an optimized semantic difference feature vector; S670, calculating the cosine similarity between the optimized semantic difference feature vector and the semantic difference feature vector; S680, generating a state update anomaly warning prompt in response to the cosine similarity being greater than or equal to a preset threshold.

[0043] Specifically, in the technical solution of this application, firstly, information on the first and second trays is acquired, along with sensor signals collected by the second sensor. Next, considering that the tray information includes information about the electronic coil tray itself, information about the relationships between the electronic coil trays, the position information of the electronic coil tray, the material state attribute information of the electronic coil tray, and the management information of the electronic coil tray, in order to better understand the semantic features in the first and second tray information, the technical solution of this application further performs semantic encoding on the first and second tray information to obtain semantic encoding feature vectors for the first and second tray information, and calculates the semantic difference feature vector between the first and second tray information semantic encoding feature vectors. By calculating the semantic difference feature vector between the two tray information semantic encoding features, the semantic differences and changes in the tray information at different points in time can be more accurately measured during the tray information update process. This measurement and quantification helps to identify whether there are significant changes in the tray information during the update, thereby using this semantic difference characterization for subsequent state update anomaly detection tasks.

[0044] Then, the sensor signal is input into a signal feature extractor based on a convolutional neural network model for feature mining to extract the implicit waveform semantic feature information in the sensor signal, thereby obtaining the sensor signal waveform semantic feature vector.

[0045] Furthermore, in order to detect anomalies in the material tray information update process by utilizing the alignment between the semantic difference features of the material tray information and the sensor signal features, it is necessary to calculate the feature difference between the two, i.e., the differential feature distribution. However, in the technical solution of this application, when calculating the feature difference between the semantic difference feature vector and the sensor signal waveform semantic feature vector to represent the differential feature distribution between the two, considering the inconsistency in the fine-grained feature distribution between the semantic difference feature vector and the sensor signal waveform semantic feature vector, this will lead to abnormal stacking of the differential feature representations of the two at the boundary of the feature manifold. Based on this, in the technical solution of this application, when calculating the differential feature distribution between the semantic difference feature vector and the sensor signal waveform semantic feature vector, the semantic difference feature vector is compensated for by backpropagation expression based on metric information based on the sensor signal waveform semantic feature vector.

[0046] To address the issue of fine-grained inconsistency in feature distribution that arises during the computation of the semantic difference feature vector and the semantic feature vector of the sensor signal waveform, a backpropagation expression compensation method based on metric information is proposed. This method aims to reduce the abnormal stacking of the difference feature distribution on the boundary of the feature manifold, thereby improving the robustness and detection performance of the model.

[0047] Specifically, the method first obtains an element-wise difference representation vector of the material tray information-sensor signal semantic shift by calculating the element-wise difference between the semantic difference feature vector and the sensor signal waveform semantic feature vector, thereby capturing the difference between the two feature vectors. Next, the semantic difference feature vector and the sensor signal waveform semantic feature vector are concatenated into a joint feature vector and input into a one-dimensional convolutional layer to generate a convolutional encoded representation vector of the material tray information-sensor signal semantic shift. That is, one-dimensional convolutional encoding is used to capture the deep-level shift implicit information between the semantic difference feature vector and the sensor signal waveform semantic feature vector. Simultaneously, the semantic difference feature vector is linearly embedded and modulated using the first embedding matrix and bias vector to obtain a linear modulation vector of the semantic difference feature vector, and the element-wise difference between the linear modulation vector of the semantic difference feature vector and the sensor signal waveform semantic feature vector is calculated to obtain the material tray information-sensor signal semantic shift modulation differential representation vector. This material tray information-sensor signal semantic shift modulation differential representation vector is used to represent the positional representation of the shift information between the semantic difference feature vector and the sensor signal waveform semantic feature vector in linear space. That is, in the technical solution of this application, the position difference vector of the material tray information-sensor signal semantic shift, the convolutional coding representation vector of the material tray information-sensor signal semantic shift, and the modulation difference representation vector of the material tray information-sensor signal semantic shift respectively represent the explicit expression of the fine-grained distribution difference shift information between the semantic difference feature vector and the sensor signal waveform semantic feature vector in different feature spaces.

[0048] Next, the element-wise difference representation vector of the material tray information-sensor signal semantic shift is used as the value vector, and the convolutional encoding representation vector and modulation difference representation vector of the material tray information-sensor signal semantic shift are used as the query vector and key vector, respectively. These vectors are input into a shift corrector based on a pseudo-converter structure to generate a multi-scale hierarchical fusion shift modulation weight vector. In this way, an attention mechanism is used to evaluate the importance of features at different scales; that is, the attention mechanism is used to simulate the hierarchical sedimentation representation of the fine-grained difference features between the semantic difference feature vector and the sensor signal waveform semantic feature vector in different feature spaces. Finally, the element-wise product of the multi-scale hierarchical fusion shift modulation weight vector of the material tray information-sensor signal semantic difference feature vector is calculated to obtain the optimized semantic difference feature vector. That is, weights are assigned to each element in the semantic difference feature vector through a weighting operation, reducing the influence of local perturbations and improving the quality of the feature vector.

[0049] Accordingly, in step S660, based on the semantic feature vector of the sensor signal waveform, the semantic difference feature vector is compensated for by backpropagation of metric information to obtain an optimized semantic difference feature vector, including: calculating the positional difference between the semantic difference feature vector and the semantic feature vector of the sensor signal waveform to obtain a positional difference representation vector of tray information-sensor signal semantic shift; concatenating the semantic difference feature vector and the semantic feature vector of the sensor signal waveform into a joint feature vector, and then inputting the joint feature vector into a one-dimensional convolutional layer to obtain a convolutional encoding representation vector of tray information-sensor signal semantic shift; using an embedding matrix and a bias vector to perform linear embedding modulation encoding on the semantic difference feature vector to obtain a linear modulation vector of the semantic difference feature vector, and then calculating the positional difference between the linear modulation vector of the semantic difference feature vector and the semantic feature vector of the sensor signal waveform. The material tray information-sensor signal semantic shift modulation differential representation vector is obtained. Using the material tray information-sensor signal semantic shift differential representation vector as the value vector, the material tray information-sensor signal semantic shift convolutional coding representation vector, and the material tray information-sensor signal semantic shift modulation differential representation vector as the query vector and key vector, respectively, the material tray information-sensor signal semantic shift differential representation vector, the material tray information-sensor signal semantic shift convolutional coding representation vector, and the material tray information-sensor signal semantic shift modulation differential representation vector are input into a shift corrector based on a pseudo-converter structure to obtain a material tray information-sensor signal semantic multi-scale hierarchical fusion shift modulation weight vector. The position-based dot product between the material tray information-sensor signal semantic multi-scale hierarchical fusion shift modulation weight vector and the semantic differential feature vector is calculated to obtain the optimized semantic differential feature vector.

[0050] Specifically, using the position difference representation vector of the material tray information-sensor signal semantic shift as the value vector, the convolutional encoding representation vector of the material tray information-sensor signal semantic shift, and the modulation difference representation vector of the material tray information-sensor signal semantic shift as the query vector and key vector, respectively, the material tray information-sensor signal semantic shift representation vector, the convolutional encoding representation vector of the material tray information-sensor signal semantic shift, and the modulation difference representation vector of the material tray information-sensor signal semantic shift are input into a shift corrector based on a transducer-like structure to obtain a multi-scale hierarchical fusion shift modulation weight vector of the material tray information-sensor signal semantics. This includes: inputting the convolutional encoding representation vector of the material tray information-sensor signal semantic shift and the modulation difference representation vector of the material tray information-sensor signal semantics into a shift corrector based on a transducer-like structure. After multiplying the transpose of the vector, the result is divided by the square root of the length of the material tray information-sensor signal semantic shift modulation differential representation vector by the position point to obtain the material tray information-sensor signal semantic shift multidimensional information interaction representation matrix. The material tray information-sensor signal semantic shift multidimensional information interaction representation matrix is ​​then subjected to soft-maximum normalization using the Softmax function to obtain the material tray information-sensor signal semantic shift multidimensional information interaction probabilistic matrix. The matrix multiplication between the material tray information-sensor signal semantic shift position differential representation vector and the material tray information-sensor signal semantic shift multidimensional information interaction probabilistic matrix is ​​then processed using the ReLU activation function to obtain the material tray information-sensor signal semantic multiscale hierarchical fusion shift modulation weight vector.

[0051] In a specific example, based on the semantic feature vector of the sensor signal waveform, backpropagation expression compensation of metric information is performed on the semantic difference feature vector to obtain an optimized semantic difference feature vector. This includes: based on the semantic feature vector of the sensor signal waveform, backpropagation expression compensation of metric information is performed on the semantic difference feature vector using the following expression compensation formula to obtain the optimized semantic difference feature vector; wherein, the expression compensation formula is: in, The semantic difference feature vector, This is the semantic feature vector of the sensor signal waveform. To classify by position difference, The semantic shift representation of the material tray information-sensor signal is expressed as a vector based on positional difference. This indicates vector concatenation processing. It is a one-dimensional convolutional layer. The data is represented by a semantic shift convolutional encoding vector of the material tray information and sensor signal. and These are the embedding matrix and the bias vector, respectively. This is a semantic shift modulation differential representation vector of the material tray information and sensor signal. For vector multiplication, The length of the material tray information-sensor signal semantic shift modulation differential representation vector. For the Softmax function, For ReLU activation functions, The shift modulation weight vector is a multi-scale hierarchical fusion of tray information and sensor signal semantics. For positional product, This is the optimized semantic difference feature vector.

[0052] In summary, in the method described above, the positional difference vector of the material tray information-sensor signal semantic shift, the convolutional encoding representation vector of the material tray information-sensor signal semantic shift, and the modulation difference representation vector of the material tray information-sensor signal semantic shift represent the explicit expression of the fine-grained distributional difference shift information between the semantic difference feature vector and the sensor signal waveform semantic feature vector in different feature spaces, respectively. The generation of the multi-scale hierarchical fusion shift modulation weight vector of the material tray information-sensor signal semantic shift utilizes an attention mechanism to evaluate the importance of features at different scales, providing weight information for feature shift correction. Ultimately, the optimized generation of the semantic difference feature vector reduces the impact of perturbations, improves the model's robustness to local perturbations in the input data, enhances the model's adaptability to noisy or variable datasets, and improves the alignment calculation quality between the semantic difference feature vector and the sensor signal waveform semantic feature vector.

[0053] Next, the cosine similarity between the optimized semantic difference feature vector and the original semantic difference feature vector is calculated. It should be understood that the optimized semantic difference feature vector includes feature information compensated by backpropagation of the measurement information between the sensor signal waveform semantic features and the material tray information semantic difference features, while the original semantic difference feature vector includes the original semantic difference features between the first and second material tray information. By calculating the cosine similarity between these two, their similarity and difference can be quantified. If the calculated cosine similarity is greater than or equal to a preset threshold, it means that the optimized feature vector is very similar to the original feature vector. This indicates a high similarity between the optimized material tray information semantic difference features compensated by the sensor signal waveform semantic feature vector and the original material tray information semantic difference features, suggesting an anomaly in the state update information. Furthermore, in response to the cosine similarity being greater than or equal to the preset threshold, a state update anomaly warning is generated. In this way, the alignment between the semantic differences in the material tray information and the sensor signal features can be used to detect anomalies in the material tray information update process, and an alert for an anomaly in the status update can be issued when an anomaly is detected, so as to ensure the consistency between the physical material and the material tray information, thereby helping to improve the accuracy and reliability of electronic reel material storage.

[0054] Based on the above embodiments, see Figure 6 The diagram shown is a structural schematic of an electronic reel material storage system 800 according to an embodiment of this application. The electronic reel material storage system 800 includes: a reel information acquisition module 810 for acquiring first reel information and second reel information; a sensor signal acquisition module 820 for acquiring sensor signals acquired by a second sensor; a semantic encoding module 830 for semantically encoding the first reel information and the second reel information to obtain semantic encoding feature vectors for the first and second reel information; a semantic difference feature vector calculation module 840 for calculating the semantic difference feature vector between the semantic encoding feature vectors for the first and second reel information; and signal feature extraction. Module 850 is used to input the sensor signal into a signal feature extractor based on a convolutional neural network model to obtain a semantic feature vector of the sensor signal waveform; backpropagation expression compensation module 860 is used to perform backpropagation expression compensation of the semantic difference feature vector based on the semantic feature vector of the sensor signal waveform to obtain an optimized semantic difference feature vector; cosine similarity calculation module 870 is used to calculate the cosine similarity between the optimized semantic difference feature vector and the semantic difference feature vector; an anomaly warning module 880 is used to generate a state update anomaly warning prompt in response to the cosine similarity being greater than or equal to a preset threshold.

[0055] Here, those skilled in the art will understand that the specific functions and operations of each module in the above-described electronic reel material storage system 800 have been referenced above. Figure 5 The method for storing electronic reel materials is described in detail in the text, and therefore, its repeated description will be omitted.

[0056] Figure 7 This is an application scenario diagram of the electronic reel material storage method according to an embodiment of this application. For example... Figure 7 As shown, in this application scenario, firstly, the information of the first tray and the information of the second tray are obtained (e.g., Figure 7 As shown in the diagram, D1) and the sensor signal acquired by the second sensor (e.g., Figure 7 As shown in D2), the first reel information, the second reel information, and the sensor signal are then input to a server (e.g., a server with an electronic reel material storage algorithm deployed on it). Figure 7 In the S shown, the server is able to use the storage algorithm of the electronic reel material to process the first reel information, the second reel information and the sensor signal to determine whether to generate a status update anomaly warning.

[0057] Based on the above embodiments, this application also provides an electronic device with another exemplary implementation. In some possible implementations, the electronic device in this application may include a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, can implement the steps of the electronic reel material storage method in the above embodiments.

[0058] Embodiments of this application also provide a computer-readable storage medium storing computer-executable instructions. When the computer-executable instructions are executed by a processor, a method for storing electronic reel material according to embodiments of this application, as described with reference to the above figures, can be performed. The computer-readable storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0059] Embodiments of this application also provide a computer program product or computer program including computer-executable instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer-executable instructions from the computer-readable storage medium and executes the computer-executable instructions, causing the computer device to perform a method for storing electronic reel stock according to embodiments of this application.

[0060] Those skilled in the art will understand that the content disclosed in this application can be varied and modified in many ways. For example, the various devices or components described above can be implemented by hardware, or by software, firmware, or a combination of some or all of the three.

[0061] Furthermore, while this application makes various references to certain units in the system according to embodiments of this application, any number of different units can be used and run on the client and / or server. The units described are merely illustrative, and different aspects of the system and method may use different units.

[0062] Those skilled in the art will understand that all or part of the steps in the above methods can be implemented by a program instructing related hardware, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk. Optionally, all or part of the steps in the above embodiments can also be implemented using one or more integrated circuits. Accordingly, each module / unit in the above embodiments can be implemented in hardware or as a software functional module. This application is not limited to any particular combination of hardware and software.

[0063] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It should also be understood that terms such as those defined in a common dictionary shall be interpreted as having a meaning consistent with their meaning in the context of the relevant art, and not as having an idealized or highly formalized meaning, unless expressly defined herein.

[0064] The above is a description of this application and should not be considered as a limitation thereof. Although several exemplary embodiments of this application have been described, those skilled in the art will readily understand that many modifications can be made to the exemplary embodiments without departing from the novel teachings and advantages of this application.

Claims

1. A method for storing electronic reel stock, characterized in that, include: When the electronic reel material is transferred to the receiving station, the sensor detects the presence of the electronic reel material based on the arrival signal, and the identification device obtains the reel information. When the gripping device moves the electronic reel material to the storage location, the sensor detects the completion of the action based on the action signal and temporarily stores the reel information in the storage system. When the electronic reel material is placed in the designated storage location, in response to the sensor sensing the completion of the action based on the action signal, the storage system archives the reel information and binds it to the designated storage location.

2. The method for storing electronic reel material according to claim 1, characterized in that, During the storage process of the electronic reel material, the sensor is constantly monitoring all the operating positions of the electronic reel material to ensure that the electronic reel material receives the corresponding signal when it reaches the designated position.

3. The method for storing electronic reel material according to claim 2, characterized in that, There is signal feedback in any storage state of the electronic reel material to report the status of the electronic reel material, and the status of the electronic reel material is assigned to the reel information to manage the physical electronic reel material, signal and reel information as a fixed correspondence.

4. The method for storing electronic reel material according to claim 1, characterized in that, The material tray information includes information about the electronic coil itself, information about the relationships between the electronic coils, information about the location of the electronic coil, information about the material status attributes of the electronic coil, and information about the management of the electronic coil.

5. The method for storing electronic reel material according to claim 1, characterized in that, The lower-level computer controls multiple sensors to collect signals from the sensors indicating that the electronic reel material is located at the mechanical position corresponding to the sensor, and processes the signals into a sequential signal stream, which is a large-small closed-loop method.

6. The method for storing electronic reel material according to claim 1, characterized in that, The host computer generates a record corresponding to the mechanical position of the electronic reel based on the signal triggered by the sensor of the slave computer, and stores it in the storage system.

7. The method for storing electronic reel material according to claim 4, characterized in that, Also includes: Obtain information about the first and second material trays; Acquire sensor signals collected by the second sensor; Semantic encoding is performed on the first tray information and the second tray information to obtain the semantic encoding feature vector of the first tray information and the semantic encoding feature vector of the second tray information. Calculate the semantic difference feature vector between the semantic encoding feature vector of the first material tray information and the semantic encoding feature vector of the second material tray information; The sensor signal is input into a signal feature extractor based on a convolutional neural network model to obtain a semantic feature vector of the sensor signal waveform. Based on the semantic feature vector of the sensor signal waveform, the semantic difference feature vector is compensated by backpropagation of metric information to obtain an optimized semantic difference feature vector; Calculate the cosine similarity between the optimized semantic difference feature vector and the semantic difference feature vector; In response to the cosine similarity being greater than or equal to a preset threshold, an abnormal state update warning is generated.

8. The method for storing electronic reel material according to claim 7, characterized in that, Based on the semantic feature vector of the sensor signal waveform, the semantic difference feature vector is compensated by backpropagation of metric information to obtain an optimized semantic difference feature vector, including: Calculate the position-based difference between the semantic difference feature vector and the sensor signal waveform semantic feature vector to obtain the tray information-sensor signal semantic shift position-based difference representation vector; After concatenating the semantic difference feature vector and the sensor signal waveform semantic feature vector into a joint feature vector, the joint feature vector is input into a one-dimensional convolutional layer to obtain the material tray information-sensor signal semantic shift convolutional encoding representation vector. After linearly embedding and modulating the semantic differential feature vector into a semantic differential feature vector linear modulation vector using an embedding matrix and a bias vector, the positional difference between the semantic differential feature vector linear modulation vector and the sensor signal waveform semantic feature vector is calculated to obtain the tray information-sensor signal semantic shift modulation differential representation vector. Using the position difference representation vector of the material tray information-sensor signal semantic shift as the value vector, the convolutional coding representation vector of the material tray information-sensor signal semantic shift, and the modulation differential representation vector of the material tray information-sensor signal semantic shift as the query vector and key vector, the material tray information-sensor signal semantic shift representation vector of position difference, the material tray information-sensor signal semantic shift convolutional coding representation vector, and the material tray information-sensor signal semantic shift modulation differential representation vector are input into a shift corrector based on a pseudo-converter structure to obtain the material tray information-sensor signal semantic multi-scale hierarchical fusion shift modulation weight vector. The optimized semantic differential feature vector is obtained by multiplying the position-based shift modulation weight vector of the material tray information-sensor signal semantic multi-scale hierarchical fusion with the semantic differential feature vector.

9. The method for storing electronic reel material according to claim 8, characterized in that, Using the position difference representation vector of the material tray information-sensor signal semantic shift as the value vector, the convolutional coding representation vector of the material tray information-sensor signal semantic shift, and the modulation differential representation vector of the material tray information-sensor signal semantic shift as the query vector and key vector, respectively, the material tray information-sensor signal semantic shift representation vector, the convolutional coding representation vector of the material tray information-sensor signal semantic shift, and the modulation differential representation vector of the material tray information-sensor signal semantic shift are input into a shift corrector based on a pseudo-converter structure to obtain a multi-scale hierarchical fusion shift modulation weight vector of the material tray information-sensor signal semantics, including: Multiply the transpose of the material tray information-sensor signal semantic shift convolutional encoding representation vector and the material tray information-sensor signal semantic shift modulation differential representation vector, and then divide the transpose by the square root of the length of the material tray information-sensor signal semantic shift modulation differential representation vector to obtain the material tray information-sensor signal semantic shift multidimensional information interaction representation matrix. The Softmax function is used to perform soft maximum value normalization on the material tray information-sensor signal semantic shift multidimensional information interaction representation matrix to obtain the material tray information-sensor signal semantic shift multidimensional information interaction probabilistic matrix. The matrix multiplication between the position difference representation vector of the material tray information-sensor signal semantic shift and the multidimensional information interaction probability matrix of the material tray information-sensor signal semantic shift is calculated and then processed by the ReLU activation function to obtain the multi-scale hierarchical fusion shift modulation weight vector of the material tray information-sensor signal semantics.

Images