| In the past two decades,China's high-speed rail has developed rapidly,and has achieved global attention.In addition,the implementation of the “Belt and Road” strategy has also brought opportunities for the development of high-speed rail,making high-speed rail shoulder the mission of the era of interconnection.As the scale of high-speed rail construction continues to expand and the speed of high-speed rail continues to increase,the corresponding demand for high-speed rail wireless communication services is also increasing.In order to meet the ever-changing demands of high-speed rail mobile communications for data transmission rates,high-speed rail communication systems have transitioned from the original GSM-R to LTE-R.In addition,5G-R has also been proposed and included in key research categories.With the successful application of deep learning in the fields of computer vision,natural language processing,speech recognition,etc.,machine learning has also been refocused.Researchers are actively trying to extend these technologies to the field of wireless communication,and then generate intelligent communication systems.Intelligent communication is considered to be one of the mainstream directions of wireless communication development after 5G.The basic idea is to introduce deep learning into all aspects of the wireless communication system,to achieve the organic integration of wireless communication and artificial intelligence technology,and to greatly improve the effectiveness of wireless communication systems.Inspired by the above ideas,this thesis is devoted to the research of channel prediction and signal detection technology of high-speed rail mobile communication based on deep learning,and exploring the intelligent development of highspeed rail mobile communication.This thesis first characterizes the path loss in wireless communication scenarios from two largescale and small-scale fading characteristics.Next,it describes some basic models of deep learning,and introduces the current mainstream framework of deep learning.In addition,this article focuses on the research of channel prediction based on LSTM-NN and signal detection based on deep learning in high-speed rail mobile communication scenarios,including the following innovative work.(1)First of all,this thesis proposes a channel prediction method based on LSTM-NN long-term and short-term memory network in the context of high-speed rail for the rapid failure of channel state information caused by high moving speed.This is a completely data-driven channel prediction method,which can effectively capture and extract the channel characteristics of the previous time to accurately predict the channel state information at the future time,greatly reducing the pilot overhead required by the traditional channel prediction method To provide technical support for the high-speed rail communication system.Simulation results show that the performance of high-speed rail channel prediction based on LSTM-NN is more stable and the prediction accuracy is higher.(2)Secondly,this thesis innovatively proposes a MIMO signal detection scheme based on deep learning for the particularity and complexity of high-speed rail scenes.The scheme divides the highspeed rail driving section into regions,thereby designing a neural network model that satisfies the spatial region compatibility,so that the entire system can directly detect signals according to the location of the high-speed rail,greatly reducing the complexity of high-speed rail signal detection.In the end,we chose viaduct as the research scenario for simulation analysis,further divided the single scenario into multiple areas,and verified the effectiveness of the scheme.The purpose of this thesis is to study the channel prediction and signal detection technology based on deep learning in high-speed rail scenarios,further improve the performance of high-speed rail communication systems,and explore the intelligent development of high-speed rail communications. |
PDF Full Text Request