RFID-based dumb resource device tag automatic classification method and system

Abstract

This invention proposes an automatic classification method and system for RFID-based dumb resource device tags, involving big data analysis, and more specifically belonging to the field of clustering and classification. The method includes: using a convolutional neural network model corresponding to a designated dumb resource device, based on various basic data including extended RFID data of the designated dumb resource device and its associated dumb resource devices, to intelligently analyze the future fault types and timing of the designated dumb resource; and performing time-sharing fault type classification on the designated dumb resource based on the intelligent analysis results. This invention addresses the technical problem of information silos due to limited dumb resource interaction capabilities in existing technologies. By extending the RFID data of dumb resource devices and employing a convolutional neural network model designed for designated dumb resource devices, it completes the recognition and classification of future time-sharing fault types of dumb resource devices based on the RFID data of the dumb resource device and its associated dumb resource devices, thereby solving the aforementioned technical problem.

Description

Technical Field

[0001] The present invention relates to big data analysis, more specifically to the field of clustering and classification, and particularly to an automatic classification method and system for dumb resource equipment tags based on RFID. Background Technology

[0002] Dumb resources refer to physical facilities and equipment in communication networks that lack intelligent processing capabilities and cannot directly interact with management systems. While they do not generate direct revenue or data processing capabilities, they play a crucial role in network connectivity, signal transmission, and infrastructure support. However, the limited communication capabilities of dumb resources mean that other devices and communication users can only gain a limited understanding of them through conventional RFID data, including static identification information. This makes it difficult to flexibly and reliably utilize the limited number of dumb resource devices and fully leverage their key roles in the communication network. Therefore, this study aims to employ big data analytics to cluster and classify various dumb resources from different perspectives, helping other devices and communication users improve the sufficiency, comprehensiveness, and predictability of their understanding of dumb resources.

[0003] For example, Chinese invention patent publication CN118799609A proposes a method, apparatus, and device for identifying dumb resource devices. The method includes: obtaining the target location of a dumb resource device in an image based on a target detection model; extracting type features of the target region image corresponding to the target location based on a target classification model; and determining the category to which the dumb resource device belongs based on the type features and a pre-built Faiss feature library for dumb resource devices. The implementation of this application uses a target classification model as a feature extractor to extract the type features of the corresponding dumb resource device, constructs a Faiss feature library for the device, and achieves a millisecond-level search response.

[0004] For example, Chinese invention patent publication CN120278177A proposes an RFID-driven method and apparatus for identifying dumb resources, relating to the field of resource identification technology. The method includes: identifying dumb resource components carrying RFID tags; performing high-density area identification based on the distribution location of the dumb resource components to obtain identified and non-identified areas; and constructing a dual-channel RFID reader, which includes a first channel and a second channel. The first channel includes a read / write transmission enhancement module, wherein the first channel is used to transmit signals from dumb resource components belonging to identified areas, and the second channel is used to transmit signals from dumb resource components in non-identified areas. This application addresses the technical problems of low RFID identification accuracy and insufficient signal transmission capability in scenarios where dumb resources are small and densely arranged, achieving the technical effects of improving identification accuracy, enhancing RFID signal transmission capability in dense areas, and realizing efficient management of dumb resources.

[0005] Therefore, the aforementioned existing technologies are either limited to classifying dumb resource devices based on their location, or, although they use RFID data, they only focus on the physical collection of RFID data from dumb resource devices. On the one hand, the types and content of information in the obtained RFID data are limited, making it impossible to expand other devices and communication users' full and comprehensive understanding of various dumb resources. On the other hand, it is impossible to specifically identify the types and timing of future failures of dumb resource devices, resulting in limited predictability of other devices' and communication users' understanding of various dumb resources. Consequently, it is difficult to use the limited number of dumb resource devices flexibly and reliably, and the key role of various dumb resources in communication networks cannot be fully utilized. Summary of the Invention

[0006] To address the technical problems in existing technologies, this invention provides an automatic classification method and system for RFID-based dumb resource devices. Targeting various dumb resource devices with limited interactive capabilities in communication networks, it expands upon their conventional RFID data to obtain richer, more diverse extended RFID data. Crucially, it employs a convolutional neural network model specifically designed for dumb resource devices. Based on a comprehensive and thoroughly screened database including extended RFID data, it performs big data analysis on the future fault types and timing of faults in the designated dumb resources. This allows for corresponding time-sharing fault type classification and storage, helping other devices and communication users to fully, comprehensively, and predictably understand various dumb resource devices, thus improving the flexibility and reliability of utilizing limited dumb resource devices.

[0007] According to a first aspect of the present invention, an automatic classification method for dumb resource equipment tags based on RFID is provided, the method comprising: By sequentially connecting the same convolutional neural networks, a convolutional neural network model corresponding to the specified dummy resource device is formed. Generate extended RFID data for the designated dumb resource device at the current moment. The extended RFID data for the designated dumb resource device at the current moment includes not only the static identification information of the designated dumb resource device, but also the dynamic interaction information, location change information and fault association information of the designated dumb resource device in each past time segment before the current moment. The generated extended RFID data of the designated dumb resource device at the current moment is stored in the blockchain storage address corresponding to the designated dumb resource device; A convolutional neural network model corresponding to the designated dummy resource device is adopted. Based on the occupation duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dummy resource device at the current moment, the extended RFID data of each device that has an information input-output relationship with the designated dummy resource in each previous time segment before the current moment, and the extended RFID data of each device that has a structural connection relationship with the designated dummy resource in each previous time segment before the current moment, the model intelligently analyzes the fault type and fault occurrence time corresponding to each fault that occurs in the designated dummy resource within the current time interval. The current time interval starts from the current moment. Based on the intelligent analysis results, time-division fault type classification is performed on the extended RFID data of the set dummy resources within the current time interval.

[0008] According to a second aspect of the present invention, an automatic classification system for dumb resource equipment tags based on RFID is provided, the system comprising a memory and a plurality of processors, the memory storing a computer program configured to be executed by the plurality of processors to perform the following steps: By sequentially connecting the same convolutional neural networks, a convolutional neural network model corresponding to the specified dummy resource device is formed. Generate extended RFID data for the designated dumb resource device at the current moment. The extended RFID data for the designated dumb resource device at the current moment includes not only the static identification information of the designated dumb resource device, but also the dynamic interaction information, location change information and fault association information of the designated dumb resource device in each past time segment before the current moment. The generated extended RFID data of the designated dumb resource device at the current moment is stored in the blockchain storage address corresponding to the designated dumb resource device; A convolutional neural network model corresponding to the designated dummy resource device is adopted. Based on the occupation duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dummy resource device at the current moment, the extended RFID data of each device that has an information input-output relationship with the designated dummy resource in each previous time segment before the current moment, and the extended RFID data of each device that has a structural connection relationship with the designated dummy resource in each previous time segment before the current moment, the model intelligently analyzes the fault type and fault occurrence time corresponding to each fault that occurs in the designated dummy resource within the current time interval. The current time interval starts from the current moment. Based on the intelligent analysis results, time-division fault type classification is performed on the extended RFID data of the set dummy resources within the current time interval.

[0009] According to a third aspect of the present invention, an automatic classification system for dumb resource equipment tags based on RFID is provided, the system comprising: An object parsing device is used to connect various identical convolutional neural networks in sequence to form a convolutional neural network model corresponding to a given dummy resource device. The data generation device is used to generate extended RFID data of the designated dumb resource device at the current moment. The extended RFID data of the designated dumb resource device at the current moment includes not only the static identification information of the designated dumb resource device, but also the dynamic interaction information, location change information and fault association information of the designated dumb resource device in each past time segment before the current moment. Content storage device, used to store the generated extended RFID data of the designated dumb resource device at the current moment to the blockchain storage address corresponding to the designated dumb resource device; The fault analysis device, connected to the object parsing device and the content storage device respectively, is used to intelligently analyze the fault type and fault occurrence time of each fault occurring in the set dummy resource device within the current time interval, based on the occupation duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the set dummy resource device at the current moment, the extended RFID data of each device that has an information input-output relationship with the set dummy resource in each previous time segment before the current moment, and the extended RFID data of each device that has a structural connection relationship with the set dummy resource in each previous time segment before the current moment. The current time interval starts from the current moment. The tag classification device, connected to the fault analysis device, is used to perform time-division fault type classification on the extended RFID data of a set dummy resource within the current time interval based on intelligent analysis results.

[0010] Compared with the prior art, the present invention has at least the following outstanding substantive features: The first approach targets various dummy resource devices with limited interactive capabilities in communication networks. By expanding their conventional RFID data, richer extended RFID data can be obtained. Crucially, by utilizing the extended RFID data of the dummy resource devices themselves, as well as the extended RFID data of other associated dummy resource devices with which there have been past information input / output relationships and / or structural connections, a big data-based intelligent analysis model can be adopted to obtain the future fault types and fault occurrence times of the dummy resources themselves. This allows for corresponding time-sharing fault type classification and storage, thereby helping other devices and communication users to improve the sufficiency, comprehensiveness, and predictability of their understanding of dummy resource devices, so as to flexibly and reliably use the limited number of various dummy resource devices. The second aspect involves constructing convolutional neural network (CNN) models with different structures for different dummy resource devices to perform intelligent analysis of the future fault types and occurrence times of the dummy resources. Specifically, by sequentially connecting identical CNNs, a CNN model corresponding to each dummy resource device is constructed. The number of sequentially connected CNNs is positively correlated with the total number of device types of dummy resources. Furthermore, the number of times the CNN model corresponding to each dummy resource device undergoes learning follows the numerical trend of the device's manufacturing time. Each CNN includes sequentially connected input, convolutional, pooling, and fully connected layers. The convolutional layers use the TanH function as the activation function, and the pooling layers use the ReLU function. This customized approach to CNN models with different structures for different dummy resource devices ensures the stability and effectiveness of the intelligent analysis results of future dummy resource fault data. Thirdly: When performing each learning operation on the convolutional neural network model corresponding to the designated dummy resource device, the fault type and occurrence time of each fault that occurred in the designated dummy resource within a certain past time interval are used as the output content of the convolutional neural network model corresponding to the designated dummy resource device. The duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dummy resource device at the start time of the certain past time interval, the extended RFID data of each device that has an information input-output relationship with the designated dummy resource in each previous time segment at the start time of the certain past time interval, and the extended RFID data of each device that has a structural connection relationship with the designated dummy resource in each previous time segment at the start time of the certain past time interval are used as the input content of the convolutional neural network model corresponding to the designated dummy resource device. This ensures the learning effect of each learning operation. Fourthly, to perform intelligent analysis of the future fault types and timing of the set dummy resources themselves, various basic data, including extended RFID data, were used. These included the duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the set dummy resource device at the current moment, the extended RFID data of each device that had an information input / output relationship with the set dummy resource in previous time segments before the current moment, and the extended RFID data of each device that had a structural connection relationship with the set dummy resource in previous time segments before the current moment. The thorough and comprehensive screening of the above basic data further ensured the stability and effectiveness of the intelligent analysis results of the future fault data of the set dummy resources. Fifthly: For each dumb resource device, the extended RFID data at any given moment includes not only the static identification information of the dumb resource device, but also the dynamic interaction information, location change information, and fault association information of the dumb resource device in each past time segment before the stated time. The dynamic interaction information of the dumb resource device in each time segment consists of a set of information-associated device identifiers and a set of structural-associated device identifiers corresponding to each time segment at evenly spaced intervals within the time segment. The location change information of the dumb resource device in each time segment consists of the real-time location information of each time segment at evenly spaced intervals within the time segment. The fault association information of the dumb resource device in each time segment consists of the fault presence identifier and fault type code value of the dumb resource device at evenly spaced intervals within the time segment. This completes the customized data structure design for various basic data, including extended RFID data. Attached Figure Description

[0011] The various embodiments of the present invention will now be described in detail with reference to the accompanying drawings: Figure 1 The invention demonstrates the overall working scenario of the RFID-based automatic classification method and system for dumb resource equipment tags.

[0012] Figures 2 to 5 The steps of the automatic classification method for dumb resource equipment tags based on RFID, as described in Implementation Scheme 1 to Implementation Scheme 4, are presented in sequence.

[0013] Figure 6 and Figure 7 The internal structure configurations of the RFID-based dumb resource equipment tag automatic classification system described in Implementation Scheme 5 and Implementation Scheme 6 are shown respectively. Detailed Implementation

[0014] like Figure 1 The diagram illustrates the overall working scenario of the RFID-based automatic classification method and system for dumb resource equipment tags described in this invention. This invention relates to big data analysis, more specifically to the field of clustering and classification.

[0015] The specific technical process of this invention is as follows: Technical Process A: To perform intelligent analysis of the future failure types and failure times of the dummy resources themselves, convolutional neural network models with different structures are built for different dummy resource devices; In detail, by sequentially connecting various identical convolutional neural networks, a convolutional neural network model corresponding to a given dummy resource device is constructed, such as... Figure 1As shown, the main aspects of setting up the convolutional neural network model corresponding to the dumb resource device and customizing its structure are as follows: Firstly, the number of sequentially connected convolutional neural networks is positively correlated with the total number of device types where dumb resource devices exist; For example, when the total number of dummy resource devices used by the current system is 50, the number of convolutional neural networks connected sequentially is 2; when the total number of dummy resource devices used by the current system is 75, the number of convolutional neural networks connected sequentially is 3; when the total number of dummy resource devices used by the current system is 100, the number of convolutional neural networks connected sequentially is 4; when the total number of dummy resource devices used by the current system is 125, the number of convolutional neural networks connected sequentially is 5, and so on. The second aspect is to set the number of times the convolutional neural network model corresponding to the dumb resource device has been trained, following the numerical trend of the dumb resource device's manufacturing time. For example, when the manufacturing life of a dumb resource device is set to 10 months, the number of times the corresponding convolutional neural network model is trained is set to 500; when the manufacturing life of a dumb resource device is set to 20 months, the number of times the corresponding convolutional neural network model is trained is set to 1000; when the manufacturing life of a dumb resource device is set to 30 months, the number of times the corresponding convolutional neural network model is trained is set to 1500; when the manufacturing life of a dumb resource device is set to 40 months, the number of times the corresponding convolutional neural network model is trained is set to 2000, and so on. Thirdly, each convolutional neural network used includes an input layer, a convolutional layer, a pooling layer, and a fully connected layer connected in sequence. The convolutional layer uses the TanH function as the activation function, and the pooling layer uses the ReLU function as the activation function. Fourthly: When performing each learning operation on the convolutional neural network model corresponding to the designated dummy resource device, the fault type and occurrence time of each fault that occurred in the designated dummy resource within a certain past time interval are used as the output content of the convolutional neural network model corresponding to the designated dummy resource device. The occupancy duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dummy resource device at the start time of the certain past time interval, the extended RFID data of each device that has an information input-output relationship with the designated dummy resource in each previous past time segment at the start time of the certain past time interval, and the extended RFID data of each device that has a structural connection relationship with the designated dummy resource in each previous past time segment at the start time of the certain past time interval are used as the input content of the convolutional neural network model corresponding to the designated dummy resource device. This ensures the learning effect of each learning operation. In this way, through the customized structural design of the above aspects, the stability and effectiveness of the intelligent analysis results of future fault data of the dummy resources are guaranteed; Technical Process B: To perform intelligent analysis that sets the future failure type and timing of the dummy resource itself, various basic data, including extended RFID data, are used. In detail, such as Figure 1 As shown, the various basic data include the duration of each time segment, the number of evenly spaced moments, the extended RFID data of the designated dumb resource device itself, the extended RFID data of other dumb resource devices with input / output associations, and the extended RFID data of other dumb resource devices with structural associations. These five basic data are respectively the duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dumb resource device at the current moment, the extended RFID data of each device that has an information input / output relationship with the designated dumb resource in previous time segments before the current moment at the current moment, and the extended RFID data of each device that has a structural connection relationship with the designated dumb resource in previous time segments before the current moment at the current moment. In more detail, for each dumb resource device, the extended RFID data at any given moment includes not only the static identification information of the dumb resource device, but also the dynamic interaction information, location change information, and fault association information of the dumb resource device in each past time segment before the stated time. The dynamic interaction information of the dumb resource device in each time segment consists of a set of information-associated device identifiers and a set of structural-associated device identifiers corresponding to each time segment at even intervals within the time segment. The location change information of the dumb resource device in each time segment consists of the real-time location information of each time segment at even intervals within the time segment. The fault association information of the dumb resource device in each time segment consists of the fault presence identifier and fault type code value of the dumb resource device at even intervals within the time segment. Thus, a customized data structure design for various basic data, including extended RFID data, is completed. Here, since the current time interval starts at the current moment, the current time interval is a type of future time interval; In this way, through the thorough and comprehensive screening of the above-mentioned basic data, the stability and effectiveness of the intelligent analysis results of the future failure data of the dummy resources are further guaranteed. Technical Process C: Using Technical Process A, a convolutional neural network model specifically designed for setting up dumb resource equipment is adopted. Based on various basic data that have been fully and comprehensively screened by Technical Process B, intelligent analysis of future failure data of the set dumb resources is completed. In detail, such as Figure 1 The intelligent analysis obtains the future fault data of the set dummy resource as the future fault type and the future fault occurrence time, that is, the fault type and occurrence time corresponding to each fault that occurs in the set dummy resource within the current time interval. Technical Process D: Based on the intelligent analysis results of Technical Process C, perform time-division fault type classification on the extended RFID data of the set dummy resources within the current time interval; In detail, based on the fault type and time of occurrence of each fault that occurs in the designated dummy resource within the current time interval obtained by intelligent analysis, the extended RFID data of the designated dummy resource within the current time interval is classified into time-division fault types. In more detail, at the time of the fault obtained by intelligent analysis, the extended RFID data of the set dummy resource is classified as the fault type obtained by intelligent analysis, and at other times, the extended RFID data of the set dummy resource is classified as the fault-free type. Technical Process E: Based on the time-division fault type classification results of the extended RFID data of the set dummy resource within the current time interval, perform classified storage of the extended RFID data of the set dummy resource within the current time interval; In detail, different fault types correspond to different storage intervals of extended RFID data, thereby completing the classification and management of the set dumb resources at various future moments.

[0016] Therefore, through the coordinated operation of the above-mentioned technical processes, this invention addresses various types of dumb resource devices with limited interactive capabilities in communication networks. By expanding their conventional RFID data, it obtains richer extended RFID data. Crucially, by utilizing the extended RFID data of the dumb resource device itself, as well as the extended RFID data of other associated dumb resource devices with which it has a past information input / output relationship and / or structural connection, a big data-based intelligent analysis model is adopted to obtain the future fault type and fault occurrence time of the set dumb resource. This allows for corresponding time-sharing fault type classification and storage, thereby helping other devices and communication users to improve the sufficiency, comprehensiveness, and predictability of their understanding of dumb resource devices, enabling flexible and reliable use of a limited number of dumb resource devices.

[0017] The key points of this invention are: RFID data extension including past dynamic interaction information, location change information and fault association information of dumb resource devices; directional design of convolutional neural network models with different structures for different dumb resource devices; intelligent fault analysis based on the extended RFID data of the dumb resource device itself and various related dumb resource devices in the past; and customized data structure design including various basic data of extended RFID data.

[0018] The following will describe in detail the RFID-based automatic classification method and system for dumb resource equipment tags of the present invention through an implementation scheme.

[0019] <Implementation Plan 1> Figure 2 The flowchart illustrates the steps of an automatic classification method for dumb resource equipment tags based on RFID according to embodiment 1 of the present invention.

[0020] like Figure 2 As shown, the RFID-based automatic classification method for dumb resource equipment tags includes the following specific steps: Step S1: Connect the identical convolutional neural networks in sequence to form the convolutional neural network model corresponding to the specified dumb resource device; In detail, connecting the same convolutional neural networks in sequence means that the same convolutional neural networks are connected in series, that is, the output of one convolutional neural network is the input of another convolutional neural network, and the structures of the convolutional neural networks are the same. Step S2: Generate extended RFID data for the designated dumb resource device at the current moment. The extended RFID data for the designated dumb resource device at the current moment includes not only the static identification information of the designated dumb resource device, but also the dynamic interaction information, location change information and fault association information of the designated dumb resource device in each past time segment before the current moment. In detail, each dumb resource device may have different extended RFID data at different times, and the extended RFID data of the dumb resource device at each time includes the static identification information of the dumb resource device as well as the dynamic interaction information, location change information and fault association information of each past time segment before the current time. In other words, the static identification information of the dumb resource device can be the same at all times, while other data can change at any time; This will help other devices and communication users to improve their understanding of dumb resource devices, so as to make flexible and reliable use of the limited number of dumb resource devices. Step S3: Store the generated extended RFID data of the designated dumb resource device at the current moment into the blockchain storage address corresponding to the designated dumb resource device; In detail, one can choose to use blockchain storage nodes to deploy each blockchain storage address to complete the storage of extended RFID data corresponding to different dumb resource devices. Of course, other blockchain storage modes can also be used. Step S4: Using the convolutional neural network model corresponding to the designated dummy resource device, based on the occupation duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dummy resource device at the current moment, the extended RFID data of each device that has an information input-output relationship with the designated dummy resource in each previous time segment before the current moment, and the extended RFID data of each device that has a structural connection relationship with the designated dummy resource in each previous time segment before the current moment, the model intelligently analyzes the fault type and fault occurrence time corresponding to each fault that occurs in the designated dummy resource within the current time interval. The current time interval starts from the current moment. For example, if the current time is 11:00 AM, the previous time segments before the current time are 10:50 AM to 11:00 AM, 10:40 AM to 10:50 AM, 10:30 AM to 10:40 AM, 10:20 AM to 10:30 AM, 10:10 AM to 10:20 AM, 10:00 AM to 10:10 AM, 9:50 AM to 10:00 AM, 9:40 AM to 9:50 AM, 9:30 AM to 9:40 AM, and 9:20 AM to 9:30 AM, for a total of 10 previous time segments. Obviously, the number of time segments in each past time segment before the current moment is also customized. For example, it is positively correlated with the total number of device types with dumb resource devices, so that different precision and different scale of associated device past data are selected for device types of different sizes. Step S5: Based on the intelligent analysis results, perform time-division fault type classification on the extended RFID data of the designated dummy resource within the current time interval; Among them, the time-division fault type classification of extended RFID data of set dummy resources in the current time interval based on intelligent analysis results includes: classifying the extended RFID data of set dummy resources in the current time interval based on the fault type and fault occurrence time corresponding to each fault of set dummy resources in the current time interval obtained by intelligent analysis. Of course, there are also cases where the set dummy resource obtained by intelligent analysis has not experienced any type of failure within the current time interval. In this case, the values ​​of the failure type and the time of failure obtained by intelligent analysis can both be empty characters, for example, both are ASCII code values ​​of empty characters. Among them, the time-division fault type classification of the extended RFID data of the set dummy resource in the current time interval based on the fault type and fault occurrence time of each fault obtained by intelligent analysis includes: classifying the extended RFID data of the set dummy resource into the fault type obtained by intelligent analysis at the fault occurrence time, and classifying the extended RFID data of the set dummy resource into the fault type at other times. This helps other devices and communication users improve their predictability of understanding dumb resource devices, so as to make flexible and reliable use of the limited number of dumb resource devices. Among them, the dynamic interaction information of the dumb resource device in each time segment is the set of information associated device identifiers corresponding to each time interval of the dumb resource device in the time segment and the set of structure associated device identifiers corresponding to each time interval of the dumb resource device in the time segment. In detail, the number of each time interval evenly spaced within the time segment is also variable, for example, it can be positively correlated with the factory time of the set dumb resource, thereby further improving the accuracy of intelligent analysis; Among them, the position change information of the dumb resource device in each time segment is the positioning information of each real-time position of the dumb resource device at each time interval within the time segment, and the fault association information of the dumb resource device in each time segment is the fault existence identifier and fault type code value of the dumb resource device at each time interval within the time segment. For example, if the duration of each time segment is 10 minutes, the time segment can be divided into 20 equal parts to obtain moments with even intervals of 30 seconds. The process of sequentially connecting the same convolutional neural networks to form a convolutional neural network model corresponding to a set dumb resource device includes: the number of sequentially connected convolutional neural networks is positively correlated with the total number of device types of dumb resource devices, and the number of times the convolutional neural network model corresponding to the set dumb resource device has been learned follows the numerical trend of the manufacturing time of the set dumb resource device. For example, the number of sequentially connected convolutional neural networks is positively correlated with the total number of device types of dummy resource devices, including: when the total number of device types of dummy resource devices used by the current system is 50, the number of sequentially connected convolutional neural networks is 2; when the total number of device types of dummy resource devices used by the current system is 75, the number of sequentially connected convolutional neural networks is 3; when the total number of device types of dummy resource devices used by the current system is 100, the number of sequentially connected convolutional neural networks is 4; when the total number of device types of dummy resource devices used by the current system is 125, the number of sequentially connected convolutional neural networks is 5, and so on. For example, the trend of the number of times the convolutional neural network model corresponding to the dumb resource device has been trained in accordance with the numerical change of the dumb resource device's manufacturing time includes: when the dumb resource device has been manufactured for 10 months, the number of times the convolutional neural network model corresponding to the dumb resource device has been trained is 500; when the dumb resource device has been manufactured for 20 months, the number of times the convolutional neural network model corresponding to the dumb resource device has been trained is 1000; when the dumb resource device has been manufactured for 30 months, the number of times the convolutional neural network model corresponding to the dumb resource device has been trained is 1500; when the dumb resource device has been manufactured for 40 months, the number of times the convolutional neural network model corresponding to the dumb resource device has been trained is 2000, and so on. Specifically, during each learning iteration of the convolutional neural network model corresponding to the designated dummy resource device, the fault type and occurrence time of each fault that occurred in the designated dummy resource within a certain past time interval are used as the output content of the convolutional neural network model corresponding to the designated dummy resource device. The occupancy duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dummy resource device at the start time of the certain past time interval, the extended RFID data of each device that has an information input-output relationship with the designated dummy resource in each previous past time segment at the start time of the certain past time interval, and the extended RFID data of each device that has a structural connection relationship with the designated dummy resource in each previous past time segment at the start time of the certain past time interval are used as the input content of the convolutional neural network model corresponding to the designated dummy resource device for this learning iteration. In detail, you can choose to use the MATLAB toolbox to complete the testing and simulation of each learning process of the convolutional neural network model corresponding to the specified dummy resource device; Among them, the set of information associated device identifiers corresponding to the dumb resource device at each moment is a set of identifier information composed of the first and last static identifier information of each device that has an information input and output relationship with the dumb resource device at that moment; Among them, the set of structurally associated device identifiers corresponding to the dumb resource device at each moment is a set of identifier information composed of the first and last static identifier information of each device that has a structural connection relationship with the dumb resource device at that moment. For example, each set of static identification information adopts a numerically normalized representation, such as a numerical representation after binary numerical conversion. In this way, the set of information associated device identifiers corresponding to the dumb resource device at each moment and the set of structural associated device identifiers corresponding to the dumb resource device at each moment are both binary code streams.

[0021] <Implementation Plan 2> Figure 3 The flowchart illustrates the steps of an automatic classification method for dumb resource equipment tags based on RFID according to embodiment 2 of the present invention.

[0022] like Figure 3 As shown, with Figure 2 The implementation scheme differs from that in the previous one. After performing time-division fault type classification on the extended RFID data of the designated dummy resource within the current time interval based on the intelligent analysis results, that is, after step S5, the method further includes: Step S6: Based on the time-division fault type classification results of the extended RFID data of the set dummy resources in the current time interval, perform classified storage of the extended RFID data of the set dummy resources in the current time interval; For example, the types of failures that may exist in dumb resources include, but are not limited to, water ingress, material aging, communication interruption, incompatibility, shell damage, poor contact, connection wear, configuration errors, etc. Different types of failures correspond to different failure type code values, and different failure type code values ​​can all be represented in binary form. Specifically, based on the time-division fault type classification results of the extended RFID data of the set dummy resources in the current time interval, the classification and storage of the extended RFID data of the set dummy resources in the current time interval includes: different fault types correspond to different storage intervals of the extended RFID data.

[0023] <Implementation Plan 3> Figure 4 The flowchart illustrates the steps of an automatic classification method for dumb resource equipment tags based on RFID according to embodiment 3 of the present invention.

[0024] like Figure 4 As shown, with Figure 2 The implementation scheme differs from the previous one. In this method, a convolutional neural network model corresponding to the designated dummy resource device is used. Based on the occupation duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dummy resource device at the current moment, the extended RFID data of each device that has an information input-output relationship with the designated dummy resource in previous time segments before the current moment, and the extended RFID data of each device that has a structural connection relationship with the designated dummy resource in previous time segments before the current moment, the method intelligently analyzes the fault type and fault occurrence time corresponding to each type of fault occurring in the designated dummy resource within the current time interval. The current time interval begins at the current moment, i.e., after step S4. The method further includes: Step S7: Receive the fault type and fault occurrence time corresponding to each fault that occurs in the current time interval of the set dummy resource, and synchronously wirelessly transmit the fault type and fault occurrence time corresponding to each fault that occurs in the current time interval of the set dummy resource to the remote big data storage node. Alternatively, a remote cloud computing storage node can be used to replace the remote big data storage node, and the fault type and time of occurrence of each fault that occurs in the current time interval of the set dummy resource can be received wirelessly. The process of receiving the fault type and time of occurrence of each fault occurring in the current time interval of the set dummy resource, and synchronously wirelessly transmitting the fault type and time of occurrence of each fault occurring in the current time interval of the set dummy resource to the remote big data storage node includes: using a time-division duplex communication mechanism to complete the synchronous wireless transmission of the fault type and time of occurrence of each fault occurring in the current time interval of the set dummy resource to the remote big data storage node.

[0025] <Implementation Plan 4> Figure 5 The flowchart illustrates the steps of an RFID-based automatic classification method for dumb resource equipment tags according to embodiment 4 of the present invention.

[0026] like Figure 5 As shown, with Figure 2 The implementation scheme differs from that in the previous one. After sequentially connecting the various identical convolutional neural networks to form a convolutional neural network model corresponding to the designated dummy resource device, i.e. after step S1, the method further includes: Step S8: Receive the convolutional neural network model corresponding to the designated dummy resource device, and perform model storage of the convolutional neural network model corresponding to the designated dummy resource device; The process of receiving and storing the convolutional neural network model corresponding to the set dumb resource device includes: storing various model parameters of the convolutional neural network model corresponding to the set dumb resource device to complete the model storage of the convolutional neural network model corresponding to the set dumb resource device. In detail, the different model parameters of the convolutional neural network model are stored in different storage units.

[0027] Next, the various method embodiments of the present invention will be described in detail.

[0028] In the RFID-based automatic classification method for dumb resource equipment tags according to various embodiments of the present invention: The number of time segments in each past time segment before the current moment is positively correlated with the total number of device types of dumb resource devices; For example, the number of time segments in each past time segment before the current moment is positively correlated with the total number of device types of dummy resource devices, including: when the total number of device types of dummy resource devices is 50, the number of time segments in each past time segment before the current moment is 10; when the total number of device types of dummy resource devices is 75, the number of time segments in each past time segment before the current moment is 15; when the total number of device types of dummy resource devices is 100, the number of time segments in each past time segment before the current moment is 20; when the total number of device types of dummy resource devices is 125, the number of time segments in each past time segment before the current moment is 25, and so on. Among them, the positive correlation between the number of time segments in each past time segment before the current moment and the total number of device types of dummy resource devices includes: using a numerical mapping formula to express the numerical mapping relationship between the number of time segments in each past time segment before the current moment and the total number of device types of dummy resource devices; Among them, the positive correlation between the number of time segments in each past time segment before the current moment and the total number of device types of dummy resource devices also includes: in the numerical mapping formula, the total number of device types of dummy resource devices is used as the input parameter of the numerical mapping formula, and the number of time segments in each past time segment before the current moment, which is positively correlated with the total number of device types of dummy resource devices, is used as the output parameter of the numerical mapping formula. For example, a numerical simulation mode can be used to test and simulate numerical mapping formulas; Among them, the various types of dumb resource equipment include cables, antennas, sensors, optical distribution boxes, fiber optic connectors, optical splitters, junction boxes, conduits, and patch panels; In detail, different device types correspond to different device type code values, and the different device type code values ​​can all be represented in binary form.

[0029] And in the RFID-based automatic classification method for dumb resource equipment tags according to various embodiments of the present invention: The set of information associated device identifiers corresponding to the dumb resource device at each moment is an identifier information set composed of the first and last static identifier information corresponding to each device that has an information input and output relationship with the dumb resource device at that moment. It includes: the static identifier information corresponding to each device is represented by binary value, and the first and last static identifier information corresponding to each device that has an information input and output relationship with the dumb resource device is concatenated in ascending order of the size of the binary value to obtain the corresponding identifier information set. Thus, the numerical representation of the corresponding set of identifier information obtained is a binary numerical code stream; The set of structurally associated device identifiers corresponding to the dumb resource device at each moment is a set of identifier information composed of the first and last static identifier information corresponding to each device that has a structural connection relationship with the dumb resource device at that moment. It includes: the static identifier information corresponding to each device is represented by binary value, and the first and last static identifier information corresponding to each device that has a structural connection relationship with the dumb resource device is concatenated in ascending order of the size of the binary value to obtain the corresponding identifier information set. The set of information associated device identifiers corresponding to the dumb resource device at each moment is a set of identifier information composed of the first and last static identifier information of each device that has an information input-output relationship with the dumb resource device at that moment. It also includes: when a certain device inputs its output signal to the dumb resource device at that moment, and / or a certain device receives the output signal of the dumb resource device at that moment, it is determined that the certain device has an information input-output relationship with the dumb resource device at that moment. Furthermore, the set of structurally associated device identifiers corresponding to the dumb resource device at each moment is a set of identifier information composed of the first and last static identifier information of each device that has a structural connection relationship with the dumb resource device at that moment. It also includes: when a certain device has a structural connection relationship with the dumb resource device at that moment, it is determined that the certain device has a structural connection relationship with the dumb resource device at that moment.

[0030] <Implementation Plan 5> Figure 6 This is a schematic diagram of the structure of an RFID-based dumb resource equipment tag automatic classification system according to embodiment 5 of the present invention.

[0031] like Figure 6 As shown, the RFID-based dumb resource equipment tag automatic classification system includes a memory and multiple processors. The memory stores a computer program configured to be executed by the multiple processors to complete the following steps: Step S1: Connect the identical convolutional neural networks in sequence to form the convolutional neural network model corresponding to the specified dumb resource device; In detail, connecting the same convolutional neural networks in sequence means that the same convolutional neural networks are connected in series, that is, the output of one convolutional neural network is the input of another convolutional neural network, and the structures of the convolutional neural networks are the same. Step S2: Generate extended RFID data for the designated dumb resource device at the current moment. The extended RFID data for the designated dumb resource device at the current moment includes not only the static identification information of the designated dumb resource device, but also the dynamic interaction information, location change information and fault association information of the designated dumb resource device in each past time segment before the current moment. In detail, each dumb resource device may have different extended RFID data at different times, and the extended RFID data of the dumb resource device at each time includes the static identification information of the dumb resource device as well as the dynamic interaction information, location change information and fault association information of each past time segment before the current time. In other words, the static identification information of the dumb resource device can be the same at all times, while other data can change at any time; This will help other devices and communication users to improve their understanding of dumb resource devices, so as to make flexible and reliable use of the limited number of dumb resource devices. Step S3: Store the generated extended RFID data of the designated dumb resource device at the current moment into the blockchain storage address corresponding to the designated dumb resource device; In detail, one can choose to use blockchain storage nodes to deploy each blockchain storage address to complete the storage of extended RFID data corresponding to different dumb resource devices. Of course, other blockchain storage modes can also be used. Step S4: Using the convolutional neural network model corresponding to the designated dummy resource device, based on the occupation duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dummy resource device at the current moment, the extended RFID data of each device that has an information input-output relationship with the designated dummy resource in each previous time segment before the current moment, and the extended RFID data of each device that has a structural connection relationship with the designated dummy resource in each previous time segment before the current moment, the model intelligently analyzes the fault type and fault occurrence time corresponding to each fault that occurs in the designated dummy resource within the current time interval. The current time interval starts from the current moment. For example, if the current time is 11:00 AM, the previous time segments before the current time are 10:50 AM to 11:00 AM, 10:40 AM to 10:50 AM, 10:30 AM to 10:40 AM, 10:20 AM to 10:30 AM, 10:10 AM to 10:20 AM, 10:00 AM to 10:10 AM, 9:50 AM to 10:00 AM, 9:40 AM to 9:50 AM, 9:30 AM to 9:40 AM, and 9:20 AM to 9:30 AM, for a total of 10 previous time segments. Obviously, the number of time segments in each past time segment before the current moment is also customized. For example, it is positively correlated with the total number of device types with dumb resource devices, so that different precision and different scale of associated device past data are selected for device types of different sizes. Step S5: Based on the intelligent analysis results, perform time-division fault type classification on the extended RFID data of the designated dummy resource within the current time interval; Among them, the time-division fault type classification of extended RFID data of set dummy resources in the current time interval based on intelligent analysis results includes: classifying the extended RFID data of set dummy resources in the current time interval based on the fault type and fault occurrence time corresponding to each fault of set dummy resources in the current time interval obtained by intelligent analysis. Of course, there are also cases where the set dummy resource obtained by intelligent analysis has not experienced any type of failure within the current time interval. In this case, the values ​​of the failure type and the time of failure obtained by intelligent analysis can both be empty characters, for example, both are ASCII code values ​​of empty characters. Among them, the time-division fault type classification of the extended RFID data of the set dummy resource in the current time interval based on the fault type and fault occurrence time of each fault obtained by intelligent analysis includes: classifying the extended RFID data of the set dummy resource into the fault type obtained by intelligent analysis at the fault occurrence time, and classifying the extended RFID data of the set dummy resource into the fault type at other times. This helps other devices and communication users improve their predictability of understanding dumb resource devices, so as to make flexible and reliable use of the limited number of dumb resource devices. Among them, the dynamic interaction information of the dumb resource device in each time segment is the set of information associated device identifiers corresponding to each time interval of the dumb resource device in the time segment and the set of structure associated device identifiers corresponding to each time interval of the dumb resource device in the time segment. In detail, the number of each time interval evenly spaced within the time segment is also variable, for example, it can be positively correlated with the factory time of the set dumb resource, thereby further improving the accuracy of intelligent analysis; Among them, the position change information of the dumb resource device in each time segment is the positioning information of each real-time position of the dumb resource device at each time interval within the time segment, and the fault association information of the dumb resource device in each time segment is the fault existence identifier and fault type code value of the dumb resource device at each time interval within the time segment. For example, if the duration of each time segment is 10 minutes, the time segment can be divided into 20 equal parts to obtain moments with even intervals of 30 seconds. The process of sequentially connecting the same convolutional neural networks to form a convolutional neural network model corresponding to a set dumb resource device includes: the number of sequentially connected convolutional neural networks is positively correlated with the total number of device types of dumb resource devices, and the number of times the convolutional neural network model corresponding to the set dumb resource device has been learned follows the numerical trend of the manufacturing time of the set dumb resource device. For example, the number of sequentially connected convolutional neural networks is positively correlated with the total number of device types of dummy resource devices, including: when the total number of device types of dummy resource devices used by the current system is 50, the number of sequentially connected convolutional neural networks is 2; when the total number of device types of dummy resource devices used by the current system is 75, the number of sequentially connected convolutional neural networks is 3; when the total number of device types of dummy resource devices used by the current system is 100, the number of sequentially connected convolutional neural networks is 4; when the total number of device types of dummy resource devices used by the current system is 125, the number of sequentially connected convolutional neural networks is 5, and so on. For example, the trend of the number of times the convolutional neural network model corresponding to the dumb resource device has been trained in accordance with the numerical change of the dumb resource device's manufacturing time includes: when the dumb resource device has been manufactured for 10 months, the number of times the convolutional neural network model corresponding to the dumb resource device has been trained is 500; when the dumb resource device has been manufactured for 20 months, the number of times the convolutional neural network model corresponding to the dumb resource device has been trained is 1000; when the dumb resource device has been manufactured for 30 months, the number of times the convolutional neural network model corresponding to the dumb resource device has been trained is 1500; when the dumb resource device has been manufactured for 40 months, the number of times the convolutional neural network model corresponding to the dumb resource device has been trained is 2000, and so on. Specifically, during each learning iteration of the convolutional neural network model corresponding to the designated dummy resource device, the fault type and occurrence time of each fault that occurred in the designated dummy resource within a certain past time interval are used as the output content of the convolutional neural network model corresponding to the designated dummy resource device. The occupancy duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dummy resource device at the start time of the certain past time interval, the extended RFID data of each device that has an information input-output relationship with the designated dummy resource in each previous past time segment at the start time of the certain past time interval, and the extended RFID data of each device that has a structural connection relationship with the designated dummy resource in each previous past time segment at the start time of the certain past time interval are used as the input content of the convolutional neural network model corresponding to the designated dummy resource device for this learning iteration. In detail, you can choose to use the MATLAB toolbox to complete the testing and simulation of each learning process of the convolutional neural network model corresponding to the specified dummy resource device; Among them, the set of information associated device identifiers corresponding to the dumb resource device at each moment is a set of identifier information composed of the first and last static identifier information of each device that has an information input and output relationship with the dumb resource device at that moment; Among them, the set of structurally associated device identifiers corresponding to the dumb resource device at each moment is a set of identifier information composed of the first and last static identifier information of each device that has a structural connection relationship with the dumb resource device at that moment. For example, each set of static identification information adopts a numerical representation form with normalized values, such as the numerical representation form after binary value conversion. In this way, the set of information associated device identifiers corresponding to the dumb resource device at each moment and the set of structural associated device identifiers corresponding to the dumb resource device at each moment are both binary code streams. like Figure 6 As shown, for example, N processors are given, where N is a natural number greater than or equal to 1.

[0032] <Implementation Plan 6> Figure 7 This is a schematic diagram of the structure of an RFID-based dumb resource equipment tag automatic classification system according to embodiment 6 of the present invention.

[0033] like Figure 7 As shown, the RFID-based dumb resource equipment tag automatic classification system includes the following components: An object parsing device is used to connect various identical convolutional neural networks in sequence to form a convolutional neural network model corresponding to a given dummy resource device. In detail, connecting the same convolutional neural networks in sequence means that the same convolutional neural networks are connected in series, that is, the output of one convolutional neural network is the input of another convolutional neural network, and the structures of the convolutional neural networks are the same. The data generation device is used to generate extended RFID data of the designated dumb resource device at the current moment. The extended RFID data of the designated dumb resource device at the current moment includes not only the static identification information of the designated dumb resource device, but also the dynamic interaction information, location change information and fault association information of the designated dumb resource device in each past time segment before the current moment. In detail, each dumb resource device may have different extended RFID data at different times, and the extended RFID data of the dumb resource device at each time includes the static identification information of the dumb resource device as well as the dynamic interaction information, location change information and fault association information of each past time segment before the current time. In other words, the static identification information of the dumb resource device can be the same at all times, while other data can change at any time; This will help other devices and communication users to improve their understanding of dumb resource devices, so as to make flexible and reliable use of the limited number of dumb resource devices. Content storage device, used to store the generated extended RFID data of the designated dumb resource device at the current moment to the blockchain storage address corresponding to the designated dumb resource device; For example, content storage devices can be blockchain storage nodes or other types of blockchain storage elements; In detail, one can choose to use blockchain storage nodes to deploy each blockchain storage address to complete the storage of extended RFID data corresponding to different dumb resource devices. Of course, other blockchain storage modes can also be used. The fault analysis device, connected to the object parsing device and the content storage device respectively, is used to intelligently analyze the fault type and fault occurrence time of each fault occurring in the set dummy resource device within the current time interval, based on the occupation duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the set dummy resource device at the current moment, the extended RFID data of each device that has an information input-output relationship with the set dummy resource in each previous time segment before the current moment, and the extended RFID data of each device that has a structural connection relationship with the set dummy resource in each previous time segment before the current moment. The current time interval starts from the current moment. For example, if the current time is 11:00 AM, the previous time segments before the current time are 10:50 AM to 11:00 AM, 10:40 AM to 10:50 AM, 10:30 AM to 10:40 AM, 10:20 AM to 10:30 AM, 10:10 AM to 10:20 AM, 10:00 AM to 10:10 AM, 9:50 AM to 10:00 AM, 9:40 AM to 9:50 AM, 9:30 AM to 9:40 AM, and 9:20 AM to 9:30 AM, for a total of 10 previous time segments. Obviously, the number of time segments in each past time segment before the current moment is also customized. For example, it is positively correlated with the total number of device types with dumb resource devices, so that different precision and different scale of associated device past data are selected for device types of different sizes. A tag classification device, connected to a fault analysis device, is used to perform time-division fault type classification on extended RFID data of a set dummy resource within the current time interval based on intelligent analysis results; Among them, the time-division fault type classification of extended RFID data of set dummy resources in the current time interval based on intelligent analysis results includes: classifying the extended RFID data of set dummy resources in the current time interval based on the fault type and fault occurrence time corresponding to each fault of set dummy resources in the current time interval obtained by intelligent analysis. Of course, there are also cases where the set dummy resource obtained by intelligent analysis has not experienced any type of failure within the current time interval. In this case, the values ​​of the failure type and the time of failure obtained by intelligent analysis can both be empty characters, for example, both are ASCII code values ​​of empty characters. Among them, the time-division fault type classification of the extended RFID data of the set dummy resource in the current time interval based on the fault type and fault occurrence time of each fault obtained by intelligent analysis includes: classifying the extended RFID data of the set dummy resource into the fault type obtained by intelligent analysis at the fault occurrence time, and classifying the extended RFID data of the set dummy resource into the fault type at other times. This helps other devices and communication users improve their predictability of understanding dumb resource devices, so as to make flexible and reliable use of the limited number of dumb resource devices. Among them, the dynamic interaction information of the dumb resource device in each time segment is the set of information associated device identifiers corresponding to each time interval of the dumb resource device in the time segment and the set of structure associated device identifiers corresponding to each time interval of the dumb resource device in the time segment. In detail, the number of each time interval evenly spaced within the time segment is also variable, for example, it can be positively correlated with the factory time of the set dumb resource, thereby further improving the accuracy of intelligent analysis; Among them, the position change information of the dumb resource device in each time segment is the positioning information of each real-time position of the dumb resource device at each time interval within the time segment, and the fault association information of the dumb resource device in each time segment is the fault existence identifier and fault type code value of the dumb resource device at each time interval within the time segment. For example, if the duration of each time segment is 10 minutes, the time segment can be divided into 20 equal parts to obtain moments with even intervals of 30 seconds. The process of sequentially connecting the same convolutional neural networks to form a convolutional neural network model corresponding to a set dumb resource device includes: the number of sequentially connected convolutional neural networks is positively correlated with the total number of device types of dumb resource devices, and the number of times the convolutional neural network model corresponding to the set dumb resource device has been learned follows the numerical trend of the manufacturing time of the set dumb resource device. For example, the number of sequentially connected convolutional neural networks is positively correlated with the total number of device types of dummy resource devices, including: when the total number of device types of dummy resource devices used by the current system is 50, the number of sequentially connected convolutional neural networks is 2; when the total number of device types of dummy resource devices used by the current system is 75, the number of sequentially connected convolutional neural networks is 3; when the total number of device types of dummy resource devices used by the current system is 100, the number of sequentially connected convolutional neural networks is 4; when the total number of device types of dummy resource devices used by the current system is 125, the number of sequentially connected convolutional neural networks is 5, and so on. For example, the trend of the number of times the convolutional neural network model corresponding to the dumb resource device has been trained in accordance with the numerical change of the dumb resource device's manufacturing time includes: when the dumb resource device has been manufactured for 10 months, the number of times the convolutional neural network model corresponding to the dumb resource device has been trained is 500; when the dumb resource device has been manufactured for 20 months, the number of times the convolutional neural network model corresponding to the dumb resource device has been trained is 1000; when the dumb resource device has been manufactured for 30 months, the number of times the convolutional neural network model corresponding to the dumb resource device has been trained is 1500; when the dumb resource device has been manufactured for 40 months, the number of times the convolutional neural network model corresponding to the dumb resource device has been trained is 2000, and so on. Specifically, during each learning iteration of the convolutional neural network model corresponding to the designated dummy resource device, the fault type and occurrence time of each fault that occurred in the designated dummy resource within a certain past time interval are used as the output content of the convolutional neural network model corresponding to the designated dummy resource device. The occupancy duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dummy resource device at the start time of the certain past time interval, the extended RFID data of each device that has an information input-output relationship with the designated dummy resource in each previous past time segment at the start time of the certain past time interval, and the extended RFID data of each device that has a structural connection relationship with the designated dummy resource in each previous past time segment at the start time of the certain past time interval are used as the input content of the convolutional neural network model corresponding to the designated dummy resource device for this learning iteration. In detail, you can choose to use the MATLAB toolbox to complete the testing and simulation of each learning process of the convolutional neural network model corresponding to the specified dummy resource device; Among them, the set of information associated device identifiers corresponding to the dumb resource device at each moment is a set of identifier information composed of the first and last static identifier information of each device that has an information input and output relationship with the dumb resource device at that moment; Among them, the set of structurally associated device identifiers corresponding to the dumb resource device at each moment is a set of identifier information composed of the first and last static identifier information of each device that has a structural connection relationship with the dumb resource device at that moment. For example, each set of static identification information adopts a numerically normalized representation, such as a numerical representation after binary numerical conversion. In this way, the set of information associated device identifiers corresponding to the dumb resource device at each moment and the set of structural associated device identifiers corresponding to the dumb resource device at each moment are both binary code streams.

[0034] Furthermore, the present invention may also reference the following technical contents to further demonstrate the outstanding substantial progress of the present invention: The number of sequentially connected convolutional neural networks is positively correlated with the total number of device types of dumb resource devices. The number of times the convolutional neural network model corresponding to the dumb resource device has been trained follows the numerical trend of the dumb resource device's manufacturing time. Each convolutional neural network includes a sequentially connected input layer, convolutional layer, pooling layer, and fully connected layer. The convolutional layer uses the TanH function as the activation function, and the pooling layer uses the ReLU function as the activation function. The number of sequentially connected convolutional neural networks is positively correlated with the total number of device types of dumb resource devices. The number of times the convolutional neural network model corresponding to the dumb resource device has been trained also follows the numerical trend of the manufacturing time of the dumb resource device: a first numerical change curve is used to represent the numerical trend of the manufacturing time of the dumb resource device, and a second numerical change curve is used to represent the numerical trend of the number of times the convolutional neural network model corresponding to the dumb resource device has been trained. Among them, the number of convolutional neural networks connected in sequence is positively correlated with the total number of device types of dumb resource devices, and the number of times the convolutional neural network model corresponding to the dumb resource device has been learned in accordance with the numerical change trend of the manufacturing time of the dumb resource device also includes: the curvature of the first numerical change curve at uniform intervals is equal to the curvature of the second numerical change curve at uniform intervals. For example, the number of sequentially connected convolutional neural networks is positively correlated with the total number of device types of dumb resource devices, and the number of times the convolutional neural network model corresponding to the dumb resource device has been learned follows the numerical change trend of the manufacturing time of the dumb resource device. This also includes: selecting programmable logic devices to synchronously execute the simulation and testing of the first and second numerical change curves.

[0035] Although the invention has been described in conjunction with specific embodiments thereof, it will be apparent to those skilled in the art that various modifications and variations can be made to the invention without departing from its scope and spirit. Therefore, it should be understood that the above embodiments are not restrictive in any respect, but rather illustrative.

Claims

1. An automatic classification method for dumb resource equipment tags based on RFID, characterized in that, The method includes: By sequentially connecting the same convolutional neural networks, a convolutional neural network model corresponding to the specified dummy resource device is formed. Generate extended RFID data for the designated dumb resource device at the current moment. The extended RFID data for the designated dumb resource device at the current moment includes not only the static identification information of the designated dumb resource device, but also the dynamic interaction information, location change information and fault association information of the designated dumb resource device in each past time segment before the current moment. The generated extended RFID data of the designated dumb resource device at the current moment is stored in the blockchain storage address corresponding to the designated dumb resource device; A convolutional neural network model corresponding to the designated dummy resource device is adopted. Based on the occupation duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dummy resource device at the current moment, the extended RFID data of each device that has an information input-output relationship with the designated dummy resource in each previous time segment before the current moment, and the extended RFID data of each device that has a structural connection relationship with the designated dummy resource in each previous time segment before the current moment, the model intelligently analyzes the fault type and fault occurrence time corresponding to each fault that occurs in the designated dummy resource within the current time interval. The current time interval starts from the current moment. Based on the intelligent analysis results, time-division fault type classification is performed on the extended RFID data of the set dummy resources within the current time interval.

2. The automatic classification method for dumb resource equipment tags based on RFID as described in claim 1, characterized in that: Based on the intelligent analysis results, the extended RFID data of the designated dummy resource in the current time interval is classified into time-division fault types, including: based on the fault type and fault occurrence time of each fault that occurs in the designated dummy resource in the current time interval obtained by intelligent analysis, the extended RFID data of the designated dummy resource in the current time interval is classified into time-division fault types. Among them, the time-division fault type classification of the extended RFID data of the set dummy resource in the current time interval based on the fault type and fault occurrence time of each fault obtained by intelligent analysis includes: classifying the extended RFID data of the set dummy resource into the fault type obtained by intelligent analysis at the fault occurrence time, and classifying the extended RFID data of the set dummy resource into the fault type at other times. Among them, the dynamic interaction information of the dumb resource device in each time segment is the set of information associated device identifiers corresponding to each time interval of the dumb resource device in the time segment and the set of structure associated device identifiers corresponding to each time interval of the dumb resource device in the time segment. Among them, the position change information of the dumb resource device in each time segment is the positioning information of the dumb resource device at each time interval within the time segment, and the fault association information of the dumb resource device in each time segment is the fault existence identifier and fault type code value of the dumb resource device at each time interval within the time segment.

3. The automatic classification method for dumb resource equipment tags based on RFID as described in claim 2, characterized in that: The convolutional neural networks that are the same are connected in sequence to form a convolutional neural network model corresponding to the set dumb resource device. The number of convolutional neural networks connected in sequence is positively correlated with the total number of device types of dumb resource devices. The number of times the convolutional neural network model corresponding to the set dumb resource device has been learned follows the numerical trend of the manufacturing time of the set dumb resource device. Specifically, during each learning iteration of the convolutional neural network model corresponding to the designated dummy resource device, the fault type and occurrence time of each fault that occurred in the designated dummy resource within a certain past time interval are used as the output content of the convolutional neural network model corresponding to the designated dummy resource device. The occupancy duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dummy resource device at the start time of the certain past time interval, the extended RFID data of each device that has an information input-output relationship with the designated dummy resource in each previous past time segment at the start time of the certain past time interval, and the extended RFID data of each device that has a structural connection relationship with the designated dummy resource in each previous past time segment at the start time of the certain past time interval are used as the input content of the convolutional neural network model corresponding to the designated dummy resource device for this learning iteration. Among them, the set of information associated device identifiers corresponding to the dumb resource device at each moment is a set of identifier information composed of the first and last static identifier information of each device that has an information input and output relationship with the dumb resource device at that moment; Among them, the set of structurally associated device identifiers corresponding to the dumb resource device at each moment is a set of identifier information composed of the first and last static identifier information of each device that has a structural connection relationship with the dumb resource device at that moment.

4. The automatic classification method for dumb resource equipment tags based on RFID as described in claim 3, characterized in that, After classifying the extended RFID data of the designated dummy resource within the current time interval based on the intelligent analysis results into time-division fault types, the method further includes: Based on the time-division fault type classification results of the extended RFID data of the set dummy resources in the current time interval, the extended RFID data of the set dummy resources in the current time interval is classified and stored. Specifically, based on the time-division fault type classification results of the extended RFID data of the set dummy resources in the current time interval, the classification and storage of the extended RFID data of the set dummy resources in the current time interval includes: different fault types correspond to different storage intervals of the extended RFID data.

5. The automatic classification method for dumb resource equipment tags based on RFID as described in claim 3, characterized in that, By employing a convolutional neural network model corresponding to a designated dummy resource device, and based on the occupancy duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dummy resource device at the current moment, the extended RFID data of each device at the current moment that has an information input-output relationship with the designated dummy resource in previous time segments before the current moment, and the extended RFID data of each device at the current moment that has a structural connection relationship with the designated dummy resource in previous time segments before the current moment, the method intelligently analyzes the fault type and fault occurrence time corresponding to each type of fault occurring in the designated dummy resource within the current time interval. After the current time interval is taken as the starting moment, the method further includes: Receive the fault type and fault occurrence time corresponding to each fault that occurs in the current time interval of the set dummy resource, and synchronously wirelessly transmit the fault type and fault occurrence time corresponding to each fault that occurs in the current time interval of the set dummy resource to the remote big data storage node. The process of receiving the fault type and time of occurrence of each fault occurring in the current time interval of the set dummy resource, and synchronously wirelessly transmitting the fault type and time of occurrence of each fault occurring in the current time interval of the set dummy resource to the remote big data storage node includes: using a time-division duplex communication mechanism to complete the synchronous wireless transmission of the fault type and time of occurrence of each fault occurring in the current time interval of the set dummy resource to the remote big data storage node.

6. The automatic classification method for dumb resource equipment tags based on RFID as described in claim 3, characterized in that, After sequentially connecting various identical convolutional neural networks to form a convolutional neural network model corresponding to a given dummy resource device, the method further includes: Receive the convolutional neural network model corresponding to the specified dummy resource device, and execute the model storage of the convolutional neural network model corresponding to the specified dummy resource device; The process of receiving and storing the convolutional neural network model corresponding to the designated dummy resource device includes: storing various model parameters of the convolutional neural network model corresponding to the designated dummy resource device to complete the model storage of the convolutional neural network model corresponding to the designated dummy resource device.

7. The automatic classification method for dumb resource equipment tags based on RFID as described in any one of claims 3-6, characterized in that: The number of time segments in each past time segment before the current moment is positively correlated with the total number of device types of dumb resource devices; Among them, the positive correlation between the number of time segments in each past time segment before the current moment and the total number of device types of dummy resource devices includes: using a numerical mapping formula to express the numerical mapping relationship between the number of time segments in each past time segment before the current moment and the total number of device types of dummy resource devices; Among them, the positive correlation between the number of time segments in each past time segment before the current moment and the total number of device types of dummy resource devices also includes: in the numerical mapping formula, the total number of device types of dummy resource devices is used as the input parameter of the numerical mapping formula, and the number of time segments in each past time segment before the current moment, which is positively correlated with the total number of device types of dummy resource devices, is used as the output parameter of the numerical mapping formula. Among them, the various types of dumb resource equipment include cables, antennas, sensors, optical distribution boxes, fiber optic connectors, optical splitters, junction boxes, conduits, and patch panels.

8. The automatic classification method for dumb resource equipment tags based on RFID as described in any one of claims 3-6, characterized in that: The set of information associated device identifiers corresponding to the dumb resource device at each moment is an identifier information set composed of the first and last static identifier information corresponding to each device that has an information input and output relationship with the dumb resource device at that moment. It includes: the static identifier information corresponding to each device is represented by binary value, and the first and last static identifier information corresponding to each device that has an information input and output relationship with the dumb resource device is concatenated in ascending order of the size of the binary value to obtain the corresponding identifier information set. The set of structurally associated device identifiers corresponding to the dumb resource device at each moment is a set of identifier information composed of the first and last static identifier information corresponding to each device that has a structural connection relationship with the dumb resource device at that moment. It includes: the static identifier information corresponding to each device is represented by binary value, and the first and last static identifier information corresponding to each device that has a structural connection relationship with the dumb resource device is concatenated in ascending order of the size of the binary value to obtain the corresponding identifier information set. The set of information associated device identifiers corresponding to the dumb resource device at each moment is a set of identifier information composed of the first and last static identifier information of each device that has an information input-output relationship with the dumb resource device at that moment. It also includes: when a certain device inputs its output signal to the dumb resource device at that moment, and / or a certain device receives the output signal of the dumb resource device at that moment, it is determined that the certain device has an information input-output relationship with the dumb resource device at that moment. The set of structurally associated device identifiers corresponding to the dumb resource device at each moment is a set of identifier information composed of the first and last static identifier information of each device that has a structural connection relationship with the dumb resource device at that moment. It also includes: when a certain device has a structural connection relationship with the dumb resource device at that moment, it is determined that the certain device has a structural connection relationship with the dumb resource device at that moment.

9. An RFID-based automatic classification system for dumb resource equipment tags, characterized in that, The system includes a memory and multiple processors, the memory storing a computer program, characterized in that the computer program is configured to be executed by the multiple processors to complete the following steps: By sequentially connecting the same convolutional neural networks, a convolutional neural network model corresponding to the specified dummy resource device is formed. Generate extended RFID data for the designated dumb resource device at the current moment. The extended RFID data for the designated dumb resource device at the current moment includes not only the static identification information of the designated dumb resource device, but also the dynamic interaction information, location change information and fault association information of the designated dumb resource device in each past time segment before the current moment. The generated extended RFID data of the designated dumb resource device at the current moment is stored in the blockchain storage address corresponding to the designated dumb resource device; A convolutional neural network model corresponding to the designated dummy resource device is adopted. Based on the occupation duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the designated dummy resource device at the current moment, the extended RFID data of each device that has an information input-output relationship with the designated dummy resource in each previous time segment before the current moment, and the extended RFID data of each device that has a structural connection relationship with the designated dummy resource in each previous time segment before the current moment, the model intelligently analyzes the fault type and fault occurrence time corresponding to each fault that occurs in the designated dummy resource within the current time interval. The current time interval starts from the current moment. Based on the intelligent analysis results, time-division fault type classification is performed on the extended RFID data of the set dummy resources within the current time interval.

10. An RFID-based automatic classification system for dumb resource equipment tags, characterized in that, The system includes: An object parsing device is used to connect various identical convolutional neural networks in sequence to form a convolutional neural network model corresponding to a given dummy resource device. The data generation device is used to generate extended RFID data of the designated dumb resource device at the current moment. The extended RFID data of the designated dumb resource device at the current moment includes not only the static identification information of the designated dumb resource device, but also the dynamic interaction information, location change information and fault association information of the designated dumb resource device in each past time segment before the current moment. Content storage device, used to store the generated extended RFID data of the designated dumb resource device at the current moment to the blockchain storage address corresponding to the designated dumb resource device; The fault analysis device, connected to the object parsing device and the content storage device respectively, is used to intelligently analyze the fault type and fault occurrence time of each fault occurring in the set dummy resource device within the current time interval, based on the occupation duration of each time segment, the number of evenly spaced moments within each time segment, the extended RFID data of the set dummy resource device at the current moment, the extended RFID data of each device that has an information input-output relationship with the set dummy resource in each previous time segment before the current moment, and the extended RFID data of each device that has a structural connection relationship with the set dummy resource in each previous time segment before the current moment. The current time interval starts from the current moment. The tag classification device, connected to the fault analysis device, is used to perform time-division fault type classification on the extended RFID data of a set dummy resource within the current time interval based on intelligent analysis results.

Images