CN-121980290-A - Model construction method and system for communication interference cognitive system

CN121980290ACN 121980290 ACN121980290 ACN 121980290ACN-121980290-A

Abstract

The invention provides a model construction method and a system for a communication interference cognitive system, wherein the model construction method and the system are used for constructing a communication interference cognitive reference set comprising interference scene association features and interference response reference features, then carrying out interference feature association modeling on the basis of the communication interference cognitive reference set to form an association map, carrying out cognitive model structure adaptation according to the association map to generate an initial interference cognitive model, carrying out interference scene adaptation iteration optimization on the initial model to form an optimized target interference cognitive model, and finally outputting the target model for interference recognition and response processing of the communication interference cognitive system. According to the method, the accuracy and the adaptability of the communication interference cognitive system in a complex interference environment are effectively improved by comprehensively capturing the interference characteristics and deeply mining the association relation and continuously optimizing the model structure.

Inventors

XU SHENGFA
XING JINGYI
FU XIONGJUN
CHEN HONGLIANG
GUO HUIPING
LI SHUANGYU
LI TIANTIAN
LI QIANHUI

Assignees

北京东方计量测试研究所

Dates

Publication Date: 20260505
Application Date: 20260122

Claims (10)

1. A method of model construction for a communication interference cognitive system, the method comprising: Constructing a communication interference cognition reference set, wherein the communication interference cognition reference set comprises interference scene association features and interference response reference features; Performing interference feature association modeling based on the communication interference cognitive benchmark set to form an interference feature association map; performing cognitive model structure adaptation according to the interference characteristic association map to generate an initial interference cognitive model; Performing interference scene adaptation iterative optimization on the initial interference cognitive model to form an optimized target interference cognitive model; and outputting the target interference cognitive model.
2. The method for modeling a communication interference cognitive system of claim 1, wherein the modeling a communication interference cognitive reference set, the communication interference cognitive reference set comprising an interference scenario correlation feature and an interference response reference feature, comprises: collecting scene description information and response behavior record information related to communication interference; Extracting the interference scene elements from the scene description information to separate out the core elements of the interference scene; Carrying out association relation analysis based on the interference scene core elements to generate interference scene association characteristics contained in the communication interference cognitive benchmark set; Performing response pattern refinement on the response behavior record information to separate an interference response core pattern; developing reference parameter setting based on the interference response core mode, and generating interference response reference characteristics contained in the communication interference cognitive reference set; and fusing the interference scene association features and the interference response reference features to form the communication interference cognitive reference set.
3. The method for modeling a cognitive system of communication interference according to claim 1, wherein the performing interference feature association modeling based on the cognitive reference set of communication interference to form an interference feature association map comprises: extracting key attribute information of interference scene association features in the communication interference cognitive reference set; extracting key attribute information of interference response reference features in the communication interference cognitive reference set; Performing association rule mining on the key attribute information of the interference scene association feature and the key attribute information of the interference response reference feature; Constructing a characteristic association edge set based on association rules obtained by mining, wherein each association edge in the characteristic association edge set corresponds to a group of association relations of the key attribute information; establishing an interference feature association map by taking the interference scene association features in the communication interference cognition reference set and the interference response reference features in the communication interference cognition reference set as nodes and taking the feature association edge set as a connection relation; And carding the node connection states in the interference characteristic association map to strengthen the signal transmission of the core association edges in the characteristic association edge set.
4. The method for modeling a cognitive system in communication with interference as claimed in claim 3, wherein said developing association rule mining on the key attribute information of the association feature of the interference scenario and the key attribute information of the reference feature of the interference response comprises: Inputting the key attribute information of the interference scene association feature and the key attribute information of the interference response reference feature into an association analysis module; carrying out co-occurrence frequency statistics on the key attribute information of the related features of the interference scene and the key attribute information of the reference features of the interference response to form attribute co-occurrence frequency distribution; setting association screening conditions based on the attribute co-occurrence frequency distribution, and screening co-occurrence attribute combinations meeting the association screening conditions; performing causal relationship verification on the co-occurrence attribute combination, and determining causal association directions among attributes in the co-occurrence attribute combination; generating a correlation intensity quantization result by combining the causal correlation direction and the attribute co-occurrence frequency distribution; Integrating the co-occurrence attribute combination, the causal association direction and the association intensity quantization result to form an association rule set; and redundancy removal is implemented on the association rule set, and core association rules in the association rule set are reserved.
5. The method for modeling a communication interference cognitive system according to claim 1, wherein the performing cognitive model structure adaptation according to the interference feature correlation map generates an initial interference cognitive model, and the method comprises: Analyzing node distribution characteristics and edge connection density characteristics of the interference characteristic association map; Determining a feature input layer dimension configuration scheme of a model based on the node distribution features; determining a hidden layer network topology structure of the model based on the edge connection density characteristics; building a model input layer according to the feature input layer dimension configuration scheme, wherein the model input layer receives feature vectors corresponding to the interference feature association patterns; constructing a model hiding layer according to the hiding layer network topology structure, wherein the model hiding layer comprises a characteristic association strengthening sub-layer and a characteristic conversion sub-layer; building a model output layer, wherein the model output layer outputs an interference cognition result; Connecting the model input layer, the model hiding layer and the model output layer to form an initial interference cognitive model; and optimizing interlayer connection of the initial disturbance cognitive model, and improving feature transfer efficiency among the model input layer, the model hiding layer and the model output layer.
6. The method for modeling a cognitive system in accordance with claim 5, wherein said modeling a hidden layer according to the hidden layer network topology comprises a feature association enhancement sub-layer and a feature conversion sub-layer, comprising: Determining the number of neuron connections of the feature association enhancement sub-layer based on the edge connection density features of the interference feature association map; Building an internal connection framework of a feature association strengthening sub-layer based on the neuron connection quantity, wherein the internal connection framework is matched with an edge connection mode of the interference feature association map; Setting an activation function of a feature association strengthening sub-layer, wherein the activation function strengthens signal transmission of effective association features in the interference feature association map; Determining the feature dimension conversion proportion of a feature conversion sub-layer based on the node distribution features of the interference feature association map; building a weight matrix structure of a feature conversion sublayer based on the feature dimension conversion proportion, wherein the weight matrix structure is matched with the input feature dimension of the feature vector received by the model input layer and the output feature dimension of the model output layer for outputting an interference cognition result; setting a regularization processing mode of a feature conversion sublayer, wherein the regularization processing mode inhibits the overfitting of the initial interference cognitive model; connecting the characteristic association strengthening sub-layer and the characteristic conversion sub-layer according to a preset sequence to form a complete model hiding layer; And debugging the sub-layer connection nodes of the model hiding layer, and optimizing the feature conversion effect between the feature association strengthening sub-layer and the feature conversion sub-layer.
7. The method for constructing a model of a communication interference cognitive system according to claim 1, wherein the performing iterative optimization of interference scene adaptation for the initial interference cognitive model to form an optimized target interference cognitive model includes: acquiring an interference scene adaptation verification data set, wherein the interference scene adaptation verification data set comprises feature data and cognitive result labels corresponding to various interference scenes; inputting the interference scene adaptation verification data set into the initial interference cognitive model to obtain a model initial cognitive output result output by the initial interference cognitive model; comparing the model initial cognitive output result with a cognitive result label in the interference scene adaptation verification data set to form cognitive deviation quantized data; positioning a module with insufficient suitability in the initial interference cognitive model structure based on the cognitive deviation quantification data; parameter adjustment and structure optimization are carried out on the module with insufficient suitability, and an adjusted middle interference cognitive model is obtained; inputting the interference scene adaptation verification data set into the intermediate interference cognitive model to obtain a model intermediate cognitive output result output by the intermediate interference cognitive model; Comparing the model intermediate cognitive output result with a cognitive result label in the interference scene adaptation verification data set to form new cognitive deviation quantification data; Repeating the steps of parameter adjustment and structure optimization for the module with insufficient suitability based on the new cognitive deviation quantification data until an intermediate interference cognitive model meeting the matching requirements is generated; Determining an intermediate interference cognitive model meeting proper matching requirements as an optimized target interference cognitive model; And integrating the overall structure of the target interference cognitive model, so as to improve the operation efficiency of the target interference cognitive model.
8. The method for modeling a communication interference cognitive system of claim 7, wherein the means for locating the insufficient suitability in the initial interference cognitive model structure based on the cognitive bias quantization data comprises: Performing hierarchical decomposition on the cognitive deviation quantized data to obtain an input layer deviation component of the initial interference cognitive model, a hidden layer deviation component of the initial interference cognitive model and an output layer deviation component of the initial interference cognitive model, wherein the hidden layer of the initial interference cognitive model comprises a characteristic association strengthening sub-layer and a characteristic conversion sub-layer; comparing the deviation threshold values of the input layer deviation component, the hidden layer deviation component and the output layer deviation component with the corresponding layer, and determining a target layer with exceeding deviation; If the target layer is the input layer of the initial interference cognitive model, analyzing the feature dimension matching degree of the input layer dimension configuration scheme of the initial interference cognitive model and the interference scene adaptation verification data set, and positioning the problem points with insufficient adaptation of the input layer dimension configuration; If the target layer is a hidden layer of the initial interference cognitive model, further decomposing the deviation component of the hidden layer into a deviation part corresponding to a characteristic association strengthening sub-layer in the hidden layer and a deviation part corresponding to a characteristic conversion sub-layer in the hidden layer; Comparing the deviation part corresponding to the characteristic association strengthening sub-layer in the hidden layer with the deviation threshold value of the deviation part corresponding to the characteristic conversion sub-layer in the hidden layer and the corresponding sub-layer, and determining a target sub-layer with the deviation exceeding the standard; Analyzing the adaptation degree of the internal connection architecture or the weight matrix structure of the target sub-layer and the characteristic association mode of the interference scene adaptation verification data set; If the target layer is an output layer of the initial interference cognitive model, analyzing the dimension matching degree of the output dimension of the output layer and the cognitive result label in the interference scene adaptation verification data set and the suitability of the output activation function of the output layer; Integrating the problem analysis results of each target layer or target sub-layer, and positioning a module with insufficient suitability in the initial interference cognitive model structure; and classifying the problem types of the modules with insufficient suitability.
9. The method for constructing a model of a cognitive system in communication with claim 5, wherein the constructing a model input layer according to the dimension configuration scheme of the feature input layer, the model input layer receiving feature vectors corresponding to the interference feature association map, comprises: Determining the number of input layer neurons of the model input layer based on the feature input layer dimension configuration scheme; building an infrastructure of the model input layer according to the number of neurons of the input layer, wherein the infrastructure comprises a neuron arrangement mode and a connection port; Setting a characteristic receiving protocol of the model input layer, wherein the characteristic receiving protocol is matched with a transmission format of a characteristic vector corresponding to the interference characteristic association map; Building a feature preprocessing sub-module of the model input layer, wherein the feature preprocessing sub-module performs feature alignment on feature vectors received by the model input layer; Connecting the basic framework, the characteristic receiving protocol and the characteristic preprocessing submodule to form a complete model input layer; Testing the characteristic receiving efficiency of the model input layer, and optimizing the response speed of the model input layer; and adjusting the neuron activation threshold of the model input layer, and improving the accuracy of the model input layer in receiving the feature vector.
10. A model building system for a communication interference recognition system, characterized in that the model building system for a communication interference recognition system comprises a processor and a memory, wherein the memory is connected with the processor, the memory is used for storing programs, instructions or codes, and the processor is used for executing the programs, instructions or codes in the memory so as to realize the model building method for a communication interference recognition system according to any one of claims 1-9.

Description

Model construction method and system for communication interference cognitive system Technical Field The invention relates to the field of artificial intelligence, in particular to a model construction method and system for a communication interference cognitive system. Background At the moment of rapid development of communication technology, the problem of communication interference is more complex and diversified, and the stability and reliability of a communication system are seriously threatened. Traditional communication interference recognition methods mainly rely on manually set rules or simple statistical models to identify the type of interference and respond. However, these methods have a number of limitations. On one hand, the manual setting rule is difficult to cover various complex interference scenes comprehensively, the new interference forms cannot be timely and accurately identified, on the other hand, the simple statistical model has limited processing capacity on the interference features, and the inherent association between the interference scenes and the response cannot be deeply mined, so that the cognitive accuracy and the adaptability in the complex interference environment are poor. Therefore, a model building method capable of comprehensively, accurately and adaptively recognizing communication interference is needed to improve the performance of the communication system in the interference environment. Disclosure of Invention In view of the above-mentioned problems, in combination with the first aspect of the present invention, an embodiment of the present invention provides a method for constructing a model for a cognitive system of communication interference, the method including: Constructing a communication interference cognition reference set, wherein the communication interference cognition reference set comprises interference scene association features and interference response reference features; Performing interference feature association modeling based on the communication interference cognitive benchmark set to form an interference feature association map; performing cognitive model structure adaptation according to the interference characteristic association map to generate an initial interference cognitive model; Performing interference scene adaptation iterative optimization on the initial interference cognitive model to form an optimized target interference cognitive model; and outputting the target interference cognitive model. In yet another aspect, an embodiment of the present invention further provides a model building system for a cognitive system of communication interference, including a processor, and a machine-readable storage medium, where the machine-readable storage medium is connected to the processor, and the machine-readable storage medium is used to store a program, an instruction, or a code, and the processor is used to execute the program, the instruction, or the code in the machine-readable storage medium, so as to implement the method described above. Based on the above aspects, in the embodiment of the invention, the communication interference cognition reference set comprising the interference scene association features and the interference response reference features is constructed, the interference feature association modeling is carried out based on the communication interference cognition reference set to form the association map, the complex relationship between the interference scene and the response is clearly presented, and the cognition depth of the model to the interference features is enhanced. And implementing an initial interference cognitive model generated by structural adaptation of the cognitive model according to the association graph, so that the initial interference cognitive model can be preliminarily adapted to interference characteristics. Through the iterative optimization of the interference scene adaptation aiming at the initial model, the model is continuously adjusted and perfected, and the finally formed target interference cognitive model has high accuracy and adaptability, can accurately identify various communication interferences and make reasonable responses, and effectively improves the performance and stability of the communication interference cognitive system in a complex interference environment. Drawings Fig. 1 is a schematic execution flow diagram of a model building method for a cognitive system of communication interference according to an embodiment of the present invention. Fig. 2 is a schematic diagram of exemplary hardware and software components of a model building system for a communication-interference awareness system provided by an embodiment of the present invention. Detailed Description The present invention is specifically described below with reference to the accompanying drawings, and fig. 1 is a schematic flow chart of a model construction method for a cognitive system with communicati