Research On Key Technologies Of 6G Network Based On Intelligent Grading

The mass data that will be generated during the intelligent construction and operation of 6G networks.It contains a large amount of redundant information.This poses a huge challenge to network data processing and intelligent grading.In recent years,researchers have explored artificial intelligence in future network operation and maintenance.Therefore,more and more data processing and network intelligent grading technologies have emerged.First of all,Thesis analyzes the characteristics of future network data such as high dimensionality,huge data volume and information redundancy.Through the principles of clustering analysis algorithm,multidimensional scaling transformation,principal component analysis method and linear discriminant dimensionality reduction algorithm,it is found that traditional data processing is the projection of data from high-dimensional space to low-dimensional space,which has limitations.In addition,the existing network intelligence classification is mainly based on the hierarchical network intelligence classification.It focuses on the classification evaluation of the network intelligence degree.The intelligent classification and autonomy of the network has not been realized.In summary,the reduction of data volume and the extraction of effective data features have a multiplier effect on the intelligent autonomy of the network.Secondly,thesis proposes network data processing method based on the factor analysis method in view of the possible high-dimensional and huge data characteristics of the 6G network and how to reduce the amount of data processing in 6G network.After studying the relevant factors affecting the 6G network,the network is divided into three models: the environment,the network element and the user according to the influencing factors of the network.According to specific situations such as the HataOkumura environment model,componentization of network element equipment,hierarchical network element capabilities and correlation of user behavior preferences,network factors are defined and designed.The factor analysis method is used to analyze and reorganize the data correlation,and the network ecological factors with low processing data volume are extracted from the network factors with high processing data volume.Simulation experiments are carried out on the reduction effect of data dimension and the degree of extraction of effective feature information.Studies have shown that factor analysis can effectively reduce data dimensions.Through the comparative analysis of the covariance error before and after data recovery and the neural network data recovery experiment,it proves that the data preprocessing method retains a large amount of characteristic information while reducing the amount of data.Finally,based on the premise of network element equipment componentization and network element capability classification,thesis proposes a general block diagram of network design based on intelligent classification from the perspective of service request,network resource matching,task construction,intelligent processing and target output.The intelligent path selection experiment is used to prove the possibility of the method.The research shows that: when multiple users request resources from the network,the network element equipment will automatically match the service demand capability level according to the capability level to form a network element path set.According to user requirements and different target performance requirements,such as the priority of network element capabilities,low resource consumption,small delay and short paths,etc.,the network intelligently selects intelligent decision-making in the path centralization,outputs the final path selection intelligent decision-making results,and provides reference for customized services for users.The verification results of the decision selection experiment from the network intelligent path prove the feasibility of the proposed method,which lays a foundation for the further comprehensive realization of intelligent grading of the network.