Research On Intelligent Channel Perception And Prediction Technology For B5G Wireless Communication System

The 5G communication system serves a wide variety of equipments.In addition to traditional 4G users,5G technology also improves high-quality communication services for platforms such as the Industrial Internet and the Internet of Vehicles.Accurate acqui-sition of Channel State Information(CSI)is a prerequisite for 5G communication systems to provide high-reliability and low-latency communication services.Especially in mobile scenarios,the channel state information obtained through traditional channel estimation methods is easily out of date.Therefore,designing an efficient and accurate channel estimation method is very important for 5G communication systems.In response to this problem,this paper proposes a channel sensing technology based on intelligent information fusion.This technology cleverly combines the date-driven neu-ral network-based channel prediction method with the model-driven Kalman filter-based channel prediction method to achieve an intelligent channel prediction technology with high prediction accuracy and strong robustness.This method has a significant improve-ment in channel prediction accuracy compared with existing algorithms.Compared with the existing Kalman filter prediction algorithm and neural network-based channel predic-tion algorithm,the error rate performance of the predicted channel for communication transmission is improved by 7d B and 12 d B,respectively.The main contributions of this article are as follows:· An intelligent channel sensing technology based on Kalman filter + deep neural net-work is proposed in this paper.This technology utilizes a deep neural network as a feature extractor to extract the features of the historical channel state information on the device side,and then combines the historical channel features with the current inaccurate channel pre-estimation value combined with the Kalman filter to com-plete the estimation of the channel parameters at the current moment.The proposed method solves the problems of low prediction accuracy of the existing model-based channel prediction algorithm and difficulty in training and poor robustness of the channel prediction algorithm based on neural network.· Combining mathematical analysis and simulation experiments,the reason why the channel prediction technology based on intelligent information fusion proposed in this paper is superior to the existing channel prediction technology based on neural network and the channel prediction technology based on Kalman filter is explained.It also derives in detail how the algorithm proposed in this paper solves the problems of poor robustness of the channel prediction algorithm based on neural network and poor prediction effect of traditional Kalman filter.The proposed method provides theoretical support for the subsequent use of the channel prediction algorithm pro-posed in this paper to improve the prediction performance,reduce the number of pilots used in communication,and improve transmission efficiency.· A ray tracing model is used to build an indoor electromagnetic wave transmission system.The system can simulate indoor communication environment.At the same time,a set of 26 GHz millimeter wave channel measurement system was completed,and millimeter wave line of sight propagation(Line of Sight,LOS)and non-line of sight propagation(Non-Line of Sight,NLOS)scenarios were collected in the campus environment.Measured data set.And the effectiveness of the proposed channel sensing technology based on intelligent information fusion is verified on the simulation and measured data sets.