Research On Bayesian Network Based Mobile Prediction For Beyond 5G Networks

Compared with the fourth-generation mobile communication system(4G),the fifth-generation mobile communication system(5G)and beyond 5G put forward stricter service performance requirements,but also provides more service scenarios.If the service performance of the 5G network does not meet the corresponding standards,it may cause some services to be unusable,which will have a great impact on the services of end users.Among them,reliability is an important indicator of network service quality,and the significance of its research is particularly important.With the advent of the era of intelligent information,artificial intelligence will play a major role in promoting the development of cellular network systems.Next-generation cellular networks can use machine-learning algorithms with massive data processing capabilities to perform predictive reasoning,thereby better guaranteeing the quality of service for end users.Considering that Bayesian networks can perform efficient causal reasoning on uncertain problems,this thesis uses this machine learning method to learn the data generated by the simulated cellular network,thereby obtaining a service reliability index with throughput Prediction model.The thesis mainly considers the modeling and prediction of service reliability in the two scenarios of current serving base station and considering base station handover.Aiming at the current network reliability prediction problem of serving base stations,the thesis first constructs a cellular network model for parameter generation,and secondly uses the Bayesian network method to express,parameter learn and reasoning prediction of cellular network data.In the parameter design process,the load of the base station,the distance between the user and the base station,and the propagation parameters between the user and the base station all have certain effects on the user's received signal and interference plus noise ratio.Simulation results show that when the load of the base station is low,the distance from the user is close,and there is a small amount of shadow in the propagation environment,the state probability of the network service of reliability prediction being high can reach 75%.Aiming at the problem of network reliability prediction when considering base station handover,the thesis does the following work.First,the switching index used in the paper is SINR,and the SINR value of the current sampling point be smoothed to prevent the sharp fluctuation of the continuous sampling point value.Secondly,the thesis uses the SINR value of multiple consecutive points to judge the switching point of the base station.Finally,the thesis adopts a GTT(Greedy Thick Thinning)structure-learning algorithm in the learning stage of the Bayesian network,which can obtain the Bayesian network structure that best fits the relationship between the parameters of the cellular network.After learning the Bayesian network structure,a service reliability prediction-model can obtained.The simulation results show that,whether it is forward predictive reasoning or reverse diagnostic reasoning,the model's reasoning results for network service reliability are relatively ideal.Therefore,the use of Bayesian networks for reliability reasoning of cellular networks is of great significance to the improvement of network service performance and the promotion of network automation.