A self-repair method for low-voltage power line carrier micro-power wireless communication network

Abstract

The present application discloses a self-repair method for a low-voltage power line carrier micro-power wireless communication network, which collects business requirements input by users, and introduces data processing and semantic understanding algorithms based on natural language processing and deep learning technology at the back end to perform semantic understanding of the business requirements, so as to capture the contextual semantic association features based on word granularity in the business requirements, so as to recommend a suitable bit error rate and set it as the bit error rate of the low-voltage power line carrier micro-power wireless communication network. In this way, the corresponding bit error rate requirement factor can be automatically set according to different business requirements, so as to achieve balanced distribution of network load while ensuring the quality of the communication link, which can make up for the defects of traditional repair solutions, enhance the self-repair ability of the network, and effectively improve the performance and reliability of the low-voltage power line carrier micro-power wireless communication network.

Description

Technical Field

[0001] The present application relates to the field of intelligent repair, and more specifically, to a self-repair method for a low-voltage power line carrier micro-power wireless communication network. Background Art

[0002] The low-voltage power line carrier micro-power wireless communication network is a network system that uses power lines as transmission media and superimposes micro-power carrier signals on low-voltage power lines for communication. This communication method has the advantages of wide coverage and low cost, and is widely used in smart grids, smart homes and other fields.

[0003] In low-voltage power line carrier micro-power wireless communication networks, signal transmission is affected by the characteristics of the power line itself, environmental interference and other factors, which will cause signal quality to deteriorate, thereby affecting the reliability and stability of communication. Therefore, how to effectively improve communication quality and reduce bit error rate has become one of the important issues in this network.

[0004] The bit error rate in carrier communication systems is usually affected by many factors, including channel fading, noise interference, multipath effects, etc. According to the specific requirements of different services for bit error rate, it is necessary to design corresponding bit error rate adjustment schemes to ensure that the communication quality can meet the business needs.

[0005] In the communication network of the power system, low-voltage power line carrier technology is a technology that uses power lines as transmission media to transmit data and signals on low-voltage power lines. Micro-power wireless communication technology refers to the use of low-power wireless signals for communication transmission. The combination of these two technologies can effectively expand the coverage of the communication network and improve the reliability of communication. In other words, the combination of low-voltage power line carrier technology and micro-power wireless communication technology makes it possible to build a more powerful communication infrastructure. Especially at the level of information perception of the power Internet of Things, this fusion technology can achieve wider equipment coverage and more reliable data transmission, providing important support for the realization of applications in smart grids, smart homes and other fields.

[0006] In practical applications, low-voltage power line carrier micro-power wireless communication networks may be affected by various interferences and faults, resulting in reduced communication quality or communication interruption. In order to ensure the stability and reliability of the communication network, these problems need to be discovered and repaired in a timely manner to reduce the impact of communication failures on services and users.

[0007] However, traditional repair solutions usually require manual intervention to identify and solve problems, resulting in slow response speeds for fault detection and repair. This makes it impossible to detect and handle communication failures in a timely manner, resulting in long communication interruptions and affecting network reliability and stability. In addition, traditional repair solutions usually set repair strategies uniformly for the entire network and cannot be personalized according to the particularities of different business needs. As a result, the repair solution can only ensure the quality of the communication link, but cannot achieve balanced distribution of network load.

[0008] Therefore, a self-healing solution for low-voltage power line carrier micro-power wireless communication network is desired. Summary of the invention

[0009] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a self-repair method for a low-voltage power line carrier micro-power wireless communication network, which collects the business requirements input by the user, and introduces data processing and semantic understanding algorithms based on natural language processing and deep learning technology in the back end to perform semantic understanding of the business requirements, so as to capture the contextual semantic association features based on word granularity in the business requirements, so as to recommend a suitable bit error rate and set it as the bit error rate of the low-voltage power line carrier micro-power wireless communication network. In this way, the corresponding bit error rate requirement factor can be automatically set according to different business requirements, so as to achieve balanced distribution of network load while ensuring the quality of the communication link, which can make up for the defects of traditional repair solutions and enhance the self-repair ability of the network, thereby effectively improving the performance and reliability of the low-voltage power line carrier micro-power wireless communication network.

[0010] According to one aspect of the present application, a low-voltage power line carrier micro-power wireless communication network self-repair method is provided, which includes:

[0011] Obtain business requirements input by users;

[0012] Performing word embedding encoding based on word granularity on the business requirements to obtain a sequence of business requirement word granularity embedding encoding vectors;

[0013] Passing the sequence of the business requirement word granularity embedding coding vectors through a dual-stream semantic encoder including a Bi-LSTM model and a GRU model to obtain a first-granularity business requirement semantic coding feature vector and a second-granularity business requirement semantic coding feature vector;

[0014] Inputting the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector into a feature element level dynamic integration module to obtain a multi-granularity business requirement semantic coding feature vector;

[0015] Based on the multi-granularity service demand semantic encoding feature vector, a recommended bit error rate is determined, and the recommended bit error rate is set as a bit error rate of a low-voltage power line carrier micro-power wireless communication network.

[0016] Compared with the prior art, the present application provides a self-repair method for a low-voltage power line carrier micro-power wireless communication network, which collects the business requirements input by the user, and introduces data processing and semantic understanding algorithms based on natural language processing and deep learning technology in the back end to perform semantic understanding of the business requirements, so as to capture the contextual semantic association features based on word granularity in the business requirements, so as to recommend a suitable bit error rate and set it as the bit error rate of the low-voltage power line carrier micro-power wireless communication network. In this way, the corresponding bit error rate requirement factor can be automatically set according to different business requirements, so as to achieve balanced distribution of network load while ensuring the quality of the communication link, which can make up for the defects of traditional repair solutions, enhance the self-repair ability of the network, and effectively improve the performance and reliability of the low-voltage power line carrier micro-power wireless communication network. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0018] Figure 1 It is a flow chart of a low-voltage power line carrier micro-power wireless communication network self-repair method according to an embodiment of the present application;

[0019] Figure 2 A system architecture diagram of a low-voltage power line carrier micro-power wireless communication network self-repair method according to an embodiment of the present application;

[0020] Figure 3 It is a flow chart of the training phase of a low-voltage power line carrier micro-power wireless communication network self-repair method according to an embodiment of the present application;

[0021] Figure 4 This is a flowchart of sub-step S4 of a low-voltage power line carrier micro-power wireless communication network self-repair method according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0023] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0024] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0025] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0026] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0027] Traditional repair solutions usually require manual intervention to identify and solve problems, resulting in slow response speeds for fault detection and repair. This makes it impossible to detect and handle communication faults in a timely manner, resulting in long communication interruptions and affecting network reliability and stability. In addition, traditional repair solutions usually set repair strategies uniformly for the entire network and cannot be personalized according to the particularities of different business needs. As a result, the repair solution can only ensure the quality of the communication link, but cannot achieve balanced distribution of network load.

[0028] It should be understood that the bit error rate refers to the ratio of the number of erroneous bits to the total number of transmitted bits during data transmission. When designing and maintaining communication networks, the bit error rate is a key performance indicator, which directly affects the quality and reliability of communication. Accordingly, considering that when repairing the low-voltage power line carrier micro-power wireless communication network, different services have different requirements for bit error rates. For example, voice communication may have a higher tolerance for bit error rates, while video communication or data transmission requires a lower bit error rate. Therefore, it is necessary to set the corresponding bit error rate requirement factor according to different business needs, so as to achieve balanced distribution of network load while ensuring the quality of the communication link, and ensure the efficiency and reliability of the network.

[0029] In the technical solution of the present application, a self-repair method for a low-voltage power line carrier micro-power wireless communication network is proposed. Figure 1 The present invention is a flowchart of a low-voltage power line carrier micro-power wireless communication network self-repair method according to an embodiment of the present application. Figure 2 1 is a system architecture diagram of a low-voltage power line carrier micro-power wireless communication network self-repair method according to an embodiment of the present application. Figure 1 and Figure 2 As shown, according to the embodiment of the present application, the self-repair method of the low-voltage power line carrier micro-power wireless communication network includes the following steps: S1, obtaining the business requirements input by the user; S2, performing word embedding encoding based on word granularity on the business requirements to obtain a sequence of business requirement word granularity embedding encoding vectors; S3, passing the sequence of business requirement word granularity embedding encoding vectors through a dual-stream semantic encoder including a Bi-LSTM model and a GRU model to obtain a first-granularity business requirement semantic encoding feature vector and a second-granularity business requirement semantic encoding feature vector; S4, inputting the first-granularity business requirement semantic encoding feature vector and the second-granularity business requirement semantic encoding feature vector into a feature element-level dynamic integration module to obtain a multi-granularity business requirement semantic encoding feature vector; S5, determining the recommended setting bit error rate based on the multi-granularity business requirement semantic encoding feature vector, and setting the recommended setting bit error rate as the bit error rate of the low-voltage power line carrier micro-power wireless communication network.

[0030] In particular, the S1 and S2 obtain the business requirements input by the user; and, the business requirements are word-embedded based on word granularity to obtain a sequence of business requirement word granularity embedding encoding vectors. Considering that there is a large amount of semantic information about the business requirements in the business requirements input by the user, the most basic semantic unit of these semantic information is a word or phrase. Based on this, in order to be able to more fully perform semantic analysis and understanding of the business requirements, so as to recommend setting a bit error rate suitable for the business requirements, in the technical solution of the present application, it is necessary to perform word segmentation on the business requirements and then pass through a word embedding layer based on the Word2Vec model to obtain a sequence of business requirement word granularity embedding encoding vectors. By performing word segmentation on the business requirement text, the long text can be divided into independent words or phrases, which helps to capture the semantic information and correlation relationship of each word in the business requirements in the subsequent process, so as to set the bit error rate. In particular, the Word2Vec model is a commonly used word embedding model, which is used to convert words into dense vector representations and capture the semantic features of each word in the business requirements, so as to better understand the meaning of the text.

[0031] In particular, the S3, the sequence of the business demand word granularity embedding coding vectors is passed through a dual-stream semantic encoder including a Bi-LSTM model and a GRU model to obtain a first granularity business demand semantic coding feature vector and a second granularity business demand semantic coding feature vector. It should be understood that since the word embedding layer based on the Word2Vec model can well capture the semantic understanding features of each word in the business demand, but the understanding and perception ability of the correlation relationship and associated semantics between each word is weak, but since the semantics of each word in the business demand does not exist independently, the contextual semantics between each word reflects the semantics and meaning of the business demand, which is conducive to the subsequent setting of the bit error rate. Based on this, in order to extract higher-level semantic features and associations between the semantic coding features of each word in the business demand, so as to better represent and understand the meaning of the business demand text, in the technical solution of the present application, the sequence of the business demand word granularity embedding coding vectors is further passed through a dual-stream semantic encoder including a Bi-LSTM model and a GRU model to obtain a first granularity business demand semantic coding feature vector and a second granularity business demand semantic coding feature vector. It should be understood that the Bi-LSTM model (bidirectional long short-term memory network) and the GRU model (gated recurrent unit) are commonly used recurrent neural network (RNN) variants for processing sequence data. The Bi-LSTM model can simultaneously consider the contextual information between the semantics of each word in the business demand, while the GRU model simplifies the structure of the LSTM to a certain extent and improves the computational efficiency. At the same time, the GRU model can pay more attention to the semantic features based on the medium and short distance dependency association between the embedding features of each word in the sequence of the word embedding encoding vector. Therefore, by combining Bi-LSTM and GRU to build a two-stream semantic encoder, their respective advantages can be fully utilized, such as the modeling ability of Bi-LSTM for long-distance dependencies and the computational efficiency and modeling ability of GRU for medium and short distance dependencies, thereby improving the semantic understanding and representation ability of the sequence data of word granularity embedding features. Therefore, through the structure of this two-stream semantic encoder, the semantic features of the business demand text can be extracted at different granularities, so that the final feature vector can more comprehensively and accurately express the semantic information of the text, providing more powerful support for subsequent text classification, information retrieval or other natural language processing tasks.

[0032] In particular, the S4 inputs the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector into the feature element level dynamic integration module to obtain a multi-granularity business requirement semantic coding feature vector. It should be understood that since the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector respectively capture the semantic association feature information based on different granularities and different levels between the semantic features of each word embedding in the business requirement, in order to be able to effectively integrate the two different granularity business requirement word granularity semantic association features, so as to more comprehensively express and understand the meaning of the business requirement text, in the technical solution of the present application, the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector are further input into the feature element level dynamic integration module to obtain a multi-granularity business requirement semantic coding feature vector. Through the processing of the feature element level dynamic integration module, the business requirement semantic coding feature vectors of the two different granularities can be dynamically combined at the element level based on word granularity, so that the model can flexibly learn how to best integrate these features to better express the textual semantic information of the business requirement. That is to say, by integrating business requirement context semantic information of different granularities, the semantic information in the text can be captured more comprehensively, including semantic relations at the word level, semantic structures at the sentence level, and the semantic meaning of the entire text, thereby improving the semantic expression ability of the model. This dynamic integration method can better adapt to different types of business requirement texts and contexts, which helps to improve the performance and generalization ability of the model in business requirement semantic understanding and bit error rate recommendation setting tasks. In particular, in a specific example of the present application, such as Figure 4 As shown, S4 includes: S41, cascading the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector to obtain a multi-granularity business requirement semantic association feature vector; S42, using a transformation matrix and a bias vector to linearly transform the multi-granularity business requirement semantic association feature vector to obtain a linear transformation multi-granularity business requirement semantic association feature vector; S43, using a sigmoid activation function to activate the linear transformation multi-granularity business requirement semantic association feature vector to obtain an activated linear transformation multi-granularity business requirement semantic association feature vector; S44, using the characteristic values ​​of each position in the activated linear transformation multi-granularity business requirement semantic association feature vector as a threshold value, calculating the weighted sum of the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector to obtain the multi-granularity business requirement semantic coding feature vector.

[0033] Specifically, the S41 cascades the semantic encoding feature vector of the business requirement at the first granularity and the semantic encoding feature vector of the business requirement at the second granularity to obtain a multi-granularity business requirement semantic association feature vector. The purpose of cascading the semantic encoding feature vector of the business requirement at the first granularity and the semantic encoding feature vector of the business requirement at the second granularity is to integrate the semantic information at different granularities to enrich and enhance the feature representation. This cascading operation can help the model consider the semantic features at different granularities at the same time, thereby improving the model's understanding and modeling capabilities of business requirements.

[0034] Specifically, the S42 uses a transformation matrix and a bias vector to perform a linear transformation on the multi-granularity business demand semantic association feature vector to obtain a linear transformation multi-granularity business demand semantic association feature vector. Among them, the use of a transformation matrix and a bias vector to perform a linear transformation on the multi-granularity business demand semantic association feature vector is to introduce weights and bias terms, thereby achieving linear combination and translation of features, so that the model can learn the linear relationship of input features. In a specific example of the present application, after calculating the matrix multiplication of the multi-granularity business demand semantic association feature vector and the transformation matrix, the obtained feature vector is then added to the bias vector by position to obtain the linear transformation multi-granularity business demand semantic association feature vector.

[0035] Specifically, in S43, the linear transformation multi-granularity business demand semantic association feature vector is activated using a sigmoid activation function to obtain an activated linear transformation multi-granularity business demand semantic association feature vector. The purpose of using an activation function to activate the linearly transformed multi-granularity business demand semantic association feature vector is to introduce nonlinear features, increase the expressive power of the model, and enable the model to learn complex nonlinear relationships. It is worth mentioning that the sigmoid activation function is a commonly used nonlinear activation function that maps a real number to a continuous value ranging from 0 to 1.

[0036] Specifically, the S44 uses the feature values ​​of each position in the activated linear transformation multi-granularity business requirement semantic association feature vector as the threshold value, calculates the weighted sum of the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector to obtain the multi-granularity business requirement semantic coding feature vector. In a specific example of the present application, first, an exponential function of the first-granularity business requirement semantic coding feature vector with a natural constant as the base is calculated to obtain a first-granularity business requirement semantic class support feature vector; then, an exponential function of the second-granularity business requirement semantic coding feature vector with a natural constant as the base is calculated to obtain a second-granularity business requirement semantic class support feature vector; then, the product of the eigenvalue of each position in the first-granularity business requirement semantic class support feature vector and a threshold value is calculated to obtain a weighted item first-granularity business requirement semantic class support feature vector; the product of the eigenvalue of each position in the second-granularity business requirement semantic class support feature vector and one minus the threshold value is calculated to obtain a weighted item second-granularity business requirement semantic class support feature vector; and then the weighted item first-granularity business requirement semantic class support feature vector and the second-granularity business requirement semantic class support feature vector are added according to the position points to obtain the multi-granularity business requirement semantic coding feature vector.

[0037] In summary, in the above embodiment, the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector are input into the feature element level dynamic integration module to obtain the multi-granularity business requirement semantic coding feature vector, including: inputting the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector into the feature element level dynamic integration module and processing them with the following feature element level dynamic integration formula to obtain the multi-granularity business requirement semantic coding feature vector; wherein the feature element level dynamic integration formula is:

[0038]

[0039] Among them, v 1 and v 2 are the feature values ​​of each position in the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector, v is the feature value of each position in the multi-granularity business requirement semantic coding feature vector, V 1 and V 2are respectively the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector, concat(·,·) represents the cascade of vectors, W is the transformation matrix, b is the bias vector, T is the activated linear transformation multi-granularity business requirement semantic association feature vector, t is the threshold value of each position in the activated linear transformation multi-granularity business requirement semantic association feature vector, t∈[0,1], sigmoid(·) represents the sigmoid activation function.

[0040] It is worth mentioning that in other specific examples of the present application, the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector can also be input into the feature element level dynamic integration module in other ways to obtain a multi-granularity business requirement semantic coding feature vector, for example: input the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector; input the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector into the feature element level dynamic integration module; integrate the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector at the feature element level, and various methods can be used, such as element-by-element addition, element-by-element multiplication, element-by-element splicing, etc., to obtain the fused feature representation; to obtain the multi-granularity business requirement semantic coding feature vector.

[0041] In particular, the S5 determines the recommended bit error rate based on the multi-granularity business demand semantic coding feature vector, and sets the recommended bit error rate as the bit error rate of the low-voltage power line carrier micro-power wireless communication network. The multi-granularity business demand semantic coding feature vector is passed through a decoder-based bit error rate recommender to obtain a decoded value, and the decoded value represents the recommended bit error rate; the decoded value is set to the bit error rate of the low-voltage power line carrier micro-power wireless communication network. In other words, the contextual semantic association features based on word granularity at different granularities in the business demand are used for decoding regression to recommend a suitable bit error rate, and then the decoded value is set to the bit error rate of the low-voltage power line carrier micro-power wireless communication network. In this way, the corresponding bit error rate requirement factor can be automatically set according to different business needs, so as to achieve balanced distribution of network load while ensuring the quality of the communication link, which can make up for the defects of traditional repair solutions and enhance the self-repair ability of the network, thereby effectively improving the performance and reliability of the low-voltage power line carrier micro-power wireless communication network. Specifically, the multi-granularity business requirement semantic encoding feature vector is passed through a decoder-based bit error rate recommender to obtain a decoded value, wherein the decoded value represents the bit error rate of the recommended setting, including: using the decoder-based bit error rate recommender to decode and regress the multi-granularity business requirement semantic encoding feature vector according to the following formula to obtain a bit error rate for representing the recommended setting; wherein the formula is: Where X represents the semantic encoding feature vector of the multi-granularity business requirement, Y is the decoded value, and W is the weight matrix. Represents matrix multiplication.

[0042] Specifically, considering the misalignment of the semantic feature element-level distribution caused by the contextual semantic association encoding at different granularities of the word segmentation embedded semantic representation based on business needs, this causes local feature overflow when the first-granularity business need semantic encoding feature vector and the second-granularity business need semantic encoding feature vector are input into the feature element-level dynamic integration module for feature element-level dynamic integration, affecting the accuracy of the decoded values ​​obtained by the multi-granularity business need semantic encoding feature vectors through the decoder-based bit error rate recommender.

[0043] Preferably, the multi-granularity business demand semantic coding feature vector is passed through a decoder-based bit error rate recommender to obtain a decoding value, comprising the following steps: calculating the multi-granularity business demand semantic coding and representation matrix and the multi-granularity business demand semantic coding difference representation matrix of the multi-granularity business demand semantic coding feature vector, wherein the value of each position of the multi-granularity business demand semantic coding and representation matrix is ​​the mean of a pair of eigenvalues ​​of two positions of the multi-granularity business demand semantic coding feature vector respectively corresponding to the position coordinates, and the value of each position of the multi-granularity business demand semantic coding difference representation matrix is ​​the absolute value of the difference between a pair of eigenvalues ​​of two positions of the multi-granularity business demand semantic coding feature vector respectively corresponding to the position coordinates; performing matrix multiplication of the transposed vector of the multi-granularity business demand semantic coding feature vector and the multi-granularity business demand semantic coding and representation matrix to obtain the multi-granularity business demand semantic coding and representation vector and transposing the multi-granularity business demand semantic coding difference representation matrix with the multi-granularity business demand semantic coding The feature vectors are matrix multiplied to obtain a multi-granularity business requirement semantic coding difference representation vector, wherein the multi-granularity business requirement semantic coding feature vector is a column vector; the point sum of the multi-granularity business requirement semantic coding and representation vector and the transposed vector of the multi-granularity business requirement semantic coding difference representation vector is calculated to obtain a first multi-granularity business requirement semantic coding and difference representation vector; the matrix product of the multi-granularity business requirement semantic coding and representation matrix and the multi-granularity business requirement semantic coding difference representation matrix is ​​calculated, and the transposed vector of the multi-granularity business requirement semantic coding feature vector is matrix multiplied with the matrix product to obtain a second multi-granularity business requirement semantic coding and difference representation vector; the point sum of the first multi-granularity business requirement semantic coding and difference representation vector and the second multi-granularity business requirement semantic coding and difference representation vector is calculated to obtain a corrected multi-granularity business requirement semantic coding feature vector; the corrected multi-granularity business requirement semantic coding feature vector is passed through the decoder-based bit error rate recommender to obtain the decoded value.

[0044] That is, by using the sum representation matrix and difference representation matrix of the group aggregation statistical evaluation of the local distribution of the eigenvalue granularity of the multi-granularity business demand semantic coding feature vector as the retrieval and response distribution enhancement of the sequence of the multi-granularity business demand semantic coding feature vector, and constructing a reference-free distribution retrieval response framework based on the group aggregation sum-difference feature distribution open domain of the multi-granularity business demand semantic coding feature vector, the distribution response redundancy caused by the local overflow characteristics of the multi-granularity business demand semantic coding feature vector is avoided through response superposition, so as to realize the fidelity constraint of the self-aggregation statistical correlation of the response of the multi-granularity business demand semantic coding feature vector to the target decoding regression domain, and improve the accuracy of the decoding value obtained by the decoder-based bit error rate recommender of the multi-granularity business demand semantic coding feature vector.

[0045] It should be understood that before using the above-mentioned neural network model for inference, it is necessary to train the word embedding layer based on the Word2Vec model, the dual-stream semantic encoder including the Bi-LSTM model and the GRU model, the feature element level dynamic integration module, and the decoder-based bit error rate recommender. That is to say, in the low-voltage power line carrier micro-power wireless communication network self-repair method of the present application, it also includes a training phase for training the word embedding layer based on the Word2Vec model, the dual-stream semantic encoder including the Bi-LSTM model and the GRU model, the feature element level dynamic integration module, and the decoder-based bit error rate recommender.

[0046] Figure 3 FIG. 1 is a flow chart of the training phase of the low-voltage power line carrier micro-power wireless communication network self-repair method according to an embodiment of the present application. Figure 3As shown, according to the embodiment of the present application, the self-repair method of the low-voltage power line carrier micro-power wireless communication network includes: a training phase, including: S110, obtaining training data, wherein the training data includes training business requirements input by a user; S120, after word segmentation processing of the training business requirements, passing through a word embedding layer based on a Word2Vec model to obtain a sequence of word granularity embedding coding vectors of the training business requirements; S130, passing the sequence of word granularity embedding coding vectors of the training business requirements through a dual-stream semantic encoder including a Bi-LSTM model and a GRU model to obtain a training first-granularity business requirement semantic coding feature vector and a training second-granularity business requirement semantic coding feature vector; S140, cascading the training first-granularity business requirement semantic coding feature vector and the training second-granularity business requirement semantic coding feature vector to obtain a training multi-granularity business requirement semantic association feature vector; S150, using a transformation matrix and a bias vector to linearly transform the training multi-granularity business requirement semantic association feature vector to obtain a training linear transformation multi-granularity business requirement semantic association feature vector; S160, using sig The moid activation function activates the training linear transformation multi-granularity business demand semantic association feature vector to obtain the training activated linear transformation multi-granularity business demand semantic association feature vector; S170, using the feature values ​​of each position in the training activated linear transformation multi-granularity business demand semantic association feature vector as the threshold value, calculates the weighted sum of the training first-granularity business demand semantic coding feature vector and the training second-granularity business demand semantic coding feature vector to obtain the training multi-granularity business demand semantic coding feature vector; S180, performs feature optimization on the training multi-granularity business demand semantic coding feature vector to obtain an optimized training multi-granularity business demand semantic coding feature vector; S190, passes the optimized training multi-granularity business demand semantic coding feature vector through the decoder-based bit error rate recommender to obtain a decoding loss function value; S200, trains the word embedding layer based on the Word2Vec model, the dual-stream semantic encoder including the Bi-LSTM model and the GRU model, the feature element-level dynamic integration module and the decoder-based bit error rate recommender based on the decoding loss function value.

[0047] Preferably, in an example embodiment, the feature optimization of the training multi-granularity business demand semantic encoding feature vector is performed to obtain an optimized training multi-granularity business demand semantic encoding feature vector, including the following steps: performing a dot addition operation on the training multi-granularity business demand semantic encoding feature vector and the square root of the length of the multi-granularity business demand semantic encoding feature vector and the inverse of the square root of the two norm of the training multi-granularity business demand semantic encoding feature vector to obtain a training first multi-granularity business demand semantic encoding intermediate feature vector; calculating the exponential function of the training first multi-granularity business demand semantic encoding intermediate feature vector with a natural constant as the base to obtain a training second multi-granularity business demand semantic encoding intermediate feature vector; performing a dot multiplication operation on the training multi-granularity business demand semantic encoding feature vector and the one norm and weight hyperparameter of the training multi-granularity business demand semantic encoding feature vector to obtain a training third multi-granularity business demand semantic encoding intermediate feature vector; performing a dot addition operation on the training second multi-granularity business demand semantic encoding intermediate feature vector and the training third multi-granularity business demand semantic encoding intermediate feature vector to obtain an optimized training multi-granularity business demand semantic encoding feature vector.

[0048] Specifically, in the above example embodiment, the structured norm representation of the training multi-granularity business demand semantic coding feature vector is used as the local canonical coordinates for each eigenvalue of the training multi-granularity business demand semantic coding feature vector to determine the rotational offset of the vector overall distribution representation of the training multi-granularity business demand semantic coding feature vector relative to the eigenvalue for the offset prediction direction of each eigenvalue of the training multi-granularity business demand semantic coding feature vector as the center, and the bounding box of the vector eigenvalue distribution of the training multi-granularity business demand semantic coding feature vector is used for eigenvalue constraint to improve the constraint of the training multi-granularity business demand semantic coding feature vector under the overall decoding regression distribution, thereby improving the training speed of the model and the accuracy of the decoding value obtained by the decoder-based bit error rate recommender of the training multi-granularity business demand semantic coding feature vector. In this way, the corresponding bit error rate requirement factor can be automatically set according to different business needs, so as to achieve balanced distribution of network load while ensuring the quality of the communication link, which can make up for the defects of traditional repair schemes and enhance the self-repair ability of the network, thereby effectively improving the performance and reliability of low-voltage power line carrier micro-power wireless communication networks.

[0049] In summary, according to the embodiment of the present application, the self-repair method of the low-voltage power line carrier micro-power wireless communication network is explained, which collects the business requirements input by the user, and introduces data processing and semantic understanding algorithms based on natural language processing and deep learning technology in the back end to perform semantic understanding of the business requirements, so as to capture the contextual semantic association features based on word granularity in the business requirements, so as to recommend a suitable bit error rate and set it as the bit error rate of the low-voltage power line carrier micro-power wireless communication network. In this way, the corresponding bit error rate requirement factor can be automatically set according to different business requirements, so as to achieve balanced distribution of network load while ensuring the quality of the communication link, which can make up for the defects of traditional repair solutions and enhance the self-repair ability of the network, thereby effectively improving the performance and reliability of the low-voltage power line carrier micro-power wireless communication network.

[0050] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A low voltage power line carrier micro-power wireless communication network self-repair method, characterized in that: include: Obtain business requirements input by users; Performing word embedding encoding based on word granularity on the business requirements to obtain a sequence of business requirement word granularity embedding encoding vectors; Passing the sequence of the business requirement word granularity embedding coding vectors through a dual-stream semantic encoder including a Bi-LSTM model and a GRU model to obtain a first-granularity business requirement semantic coding feature vector and a second-granularity business requirement semantic coding feature vector; Inputting the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector into a feature element level dynamic integration module to obtain a multi-granularity business requirement semantic coding feature vector; Determine a bit error rate of a recommended setting based on the multi-granularity service demand semantic encoding feature vector, and set the bit error rate of the recommended setting as a bit error rate of a low-voltage power line carrier micro-power wireless communication network; Determining a recommended bit error rate based on the multi-granularity service demand semantic encoding feature vector, and setting the recommended bit error rate as a bit error rate of a low-voltage power line carrier micro-power wireless communication network, including: The multi-granularity business requirement semantic encoding feature vector is passed through a decoder-based bit error rate recommender to obtain a decoded value, wherein the decoded value represents a bit error rate of a recommended setting; Setting the decoded value to a bit error rate of the low voltage power line carrier micropower wireless communication network; The multi-granularity business demand semantic coding feature vector is passed through a decoder-based bit error rate recommender to obtain a decoding value, comprising the following steps: calculating a multi-granularity business demand semantic coding and representation matrix and a multi-granularity business demand semantic coding difference representation matrix of the multi-granularity business demand semantic coding feature vector, wherein the value of each position of the multi-granularity business demand semantic coding and representation matrix is ​​the mean of a pair of eigenvalues ​​of two positions of the multi-granularity business demand semantic coding feature vector respectively corresponding to the position coordinates, and the value of each position of the multi-granularity business demand semantic coding difference representation matrix is ​​the absolute value of the difference between a pair of eigenvalues ​​of two positions of the multi-granularity business demand semantic coding feature vector respectively corresponding to the position coordinates; performing matrix multiplication of the transposed vector of the multi-granularity business demand semantic coding feature vector and the multi-granularity business demand semantic coding and representation matrix to obtain a multi-granularity business demand semantic coding and representation vector and multiplying the multi-granularity business demand semantic coding difference representation matrix with the multi-granularity business demand semantic coding feature vector. The vectors are matrix multiplied to obtain a multi-granularity business requirement semantic coding difference representation vector, wherein the multi-granularity business requirement semantic coding feature vector is a column vector; the point sum of the multi-granularity business requirement semantic coding and representation vector and the transpose vector of the multi-granularity business requirement semantic coding difference representation vector is calculated to obtain a first multi-granularity business requirement semantic coding and difference representation vector; the matrix product of the multi-granularity business requirement semantic coding and representation matrix and the multi-granularity business requirement semantic coding difference representation matrix is ​​calculated, and the transpose vector of the multi-granularity business requirement semantic coding feature vector is matrix multiplied with the matrix product to obtain a second multi-granularity business requirement semantic coding and difference representation vector; the point sum of the first multi-granularity business requirement semantic coding and difference representation vector and the second multi-granularity business requirement semantic coding and difference representation vector is calculated to obtain a corrected multi-granularity business requirement semantic coding feature vector; the corrected multi-granularity business requirement semantic coding feature vector is passed through the decoder-based bit error rate recommender to obtain the decoded value.

2. A low voltage power line carrier micro-power wireless communication network self-repair method according to claim 1, characterized in that: The business requirements are encoded based on word granularity to obtain a sequence of business requirement word granularity embedded coding vectors, including: after word segmentation processing, the business requirements are passed through a word embedding layer based on a Word2Vec model to obtain a sequence of business requirement word granularity embedded coding vectors.

3. A low voltage power line carrier micro-power wireless communication network self-repair method according to claim 2, characterized in that: Inputting the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector into a feature element level dynamic integration module to obtain a multi-granularity business requirement semantic coding feature vector, including: Cascading the first granularity business requirement semantic encoding feature vector and the second granularity business requirement semantic encoding feature vector to obtain a multi-granularity business requirement semantic association feature vector; Using a transformation matrix and a bias vector to linearly transform the multi-granularity business requirement semantic association feature vector to obtain a linearly transformed multi-granularity business requirement semantic association feature vector; Using a sigmoid activation function to activate the linear transformation multi-granularity business demand semantic association feature vector to obtain an activated linear transformation multi-granularity business demand semantic association feature vector; Taking each position feature value in the activated linear transformation multi-granularity business requirement semantic association feature vector as a threshold value, the weighted sum of the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector is calculated to obtain the multi-granularity business requirement semantic coding feature vector.

4. A low voltage power line carrier micro-power wireless communication network self-repair method according to claim 3, characterized in that: The multi-granularity business requirement semantic association feature vector is linearly transformed using a transformation matrix and a bias vector to obtain a linearly transformed multi-granularity business requirement semantic association feature vector, including: calculating the matrix multiplication of the multi-granularity business requirement semantic association feature vector and the transformation matrix, and then adding the obtained feature vector to the bias vector by position to obtain the linearly transformed multi-granularity business requirement semantic association feature vector.

5. A low voltage power line carrier micro-power wireless communication network self-repair method according to claim 4, characterized in that: Taking each position feature value in the activated linear transformation multi-granularity business requirement semantic association feature vector as a threshold value, calculating the weighted sum of the first granularity business requirement semantic coding feature vector and the second granularity business requirement semantic coding feature vector to obtain the multi-granularity business requirement semantic coding feature vector, including: Calculating an exponential function of the semantic encoding feature vector of the first-granularity business requirement with a natural constant as the base to obtain a first-granularity business requirement semantic class support feature vector; Calculating an exponential function of the second-granularity business requirement semantic encoding feature vector with a natural constant as the base to obtain a second-granularity business requirement semantic class support feature vector; Calculate the product of the feature value of each position in the first granularity business requirement semantic class support feature vector and the threshold value to obtain a weighted item first granularity business requirement semantic class support feature vector; Calculate the product of the feature value of each position in the second granularity business requirement semantic class support feature vector and one minus a threshold value to obtain a weighted item second granularity business requirement semantic class support feature vector; The weighted item first granularity business requirement semantic class support feature vector and the second granularity business requirement semantic class support feature vector are added according to the position points to obtain the multi-granularity business requirement semantic coding feature vector.

6. A low voltage power line carrier micro-power wireless communication network self-repair method according to claim 5, characterized in that: It also includes a training step: for training the word embedding layer based on the Word2Vec model, the dual-stream semantic encoder including the Bi-LSTM model and the GRU model, the feature element level dynamic integration module and the decoder-based bit error rate recommender.

7. A low voltage power line carrier micro-power wireless communication network self-repair method according to claim 6, characterized in that: The training step comprises: Acquiring training data, wherein the training data includes training business requirements input by a user; After word segmentation processing is performed on the training business requirements, a word embedding layer based on the Word2Vec model is used to obtain a sequence of word granularity embedding encoding vectors of the training business requirements; Passing the sequence of the training business requirement word granularity embedding coding vectors through a dual-stream semantic encoder including a Bi-LSTM model and a GRU model to obtain a training first granularity business requirement semantic coding feature vector and a training second granularity business requirement semantic coding feature vector; Cascading the training first-granularity business requirement semantic encoding feature vector and the training second-granularity business requirement semantic encoding feature vector to obtain a training multi-granularity business requirement semantic association feature vector; Using a transformation matrix and a bias vector to linearly transform the training multi-granularity business requirement semantic association feature vector to obtain a training linearly transformed multi-granularity business requirement semantic association feature vector; Using a sigmoid activation function to activate the training linear transformation multi-granularity business demand semantic association feature vector to obtain a training activated linear transformation multi-granularity business demand semantic association feature vector; Taking each position feature value in the training activated linear transformation multi-granularity business requirement semantic association feature vector as a threshold value, calculating the weighted sum of the training first-granularity business requirement semantic coding feature vector and the training second-granularity business requirement semantic coding feature vector to obtain a training multi-granularity business requirement semantic coding feature vector; Performing feature optimization on the training multi-granularity business requirement semantic encoding feature vector to obtain an optimized training multi-granularity business requirement semantic encoding feature vector; Passing the optimized training multi-granularity business demand semantic encoding feature vector through the decoder-based bit error rate recommender to obtain a decoding loss function value; Based on the decoding loss function value, the word embedding layer based on the Word2Vec model, the dual-stream semantic encoder including the Bi-LSTM model and the GRU model, the feature element level dynamic integration module and the decoder-based bit error rate recommender are trained.

Images