User Delay Prediction Based On Deep Learning

With the rapid development of mobile communication technology in recent years,the fifth generation of mobile communication technology(5G)is maturing.In 2019,the Ministry of Industry and Information Technology of China issued 5G to China Mobile,China Telecom,China Unicom,and China Radio and Television.For commercial licenses,the four major operators are making great efforts to accelerate the construction of 5G mobile communication networks.An important part of 5G network planning is base station site selection.Reasonable base station site selection can greatly save costs under the premise of satisfying coverage and network performance.The traditional fixed-point drive test and quantitative coverage model are difficult to meet the requirements of rapid,comprehensive and accurate evaluation of network performance.Although network performance simulation can achieve accurate analysis,the huge amount of calculation and time overhead make it impossible to apply it in actual network planning.As an important branch of artificial intelligence technology,deep learning has received great attention from researchers in various fields in recent years.At present,it is an inevitable trend to apply artificial intelligence technology to various fields of society.Artificial intelligence technology has achieved certain research results in the fields of medical treatment,security,and autonomous driving.In view of the shortcomings of traditional network planning methods,this paper introduces a deep neural network model,and proposes a user delay prediction method that can be used for 5G base station site selection to meet the needs of actual network planning.The main research contents and innovations are as follows:1.This paper proposes a method that can be used for user delay prediction.This method combines system simulation and deep neural model.First,a ray tracing model was combined with 5G wireless simulation platform to build a delay simulation model and obtain a large number of user delays data.The deep neural network model was used to learn the characteristics of the delay data to train the neural network model.The trained neural network model could use the output of the ray tracing model as input data to quickly and accurately predict the user's delay without going through the simulation platform to save a lot of time and computing resources.2.A three-view feature model was proposed for input feature extraction of neural network model.A key task in the process of training the neural network model is the extraction of input features.Reasonable input features can greatly improve the prediction accuracy of the neural network model.Aiming at the characteristics of the ray tracing model,a three-view feature model was proposed based on wireless communication theory.A large number of experimental results proved the validity and accuracy of the three-view feature model.