Research Of Channel Impulseresponse Estimation Based On Basic Expansion Model In The Scene Of High-speed Railway

China high-speed railwayhas been rapidly developed in recent years and the existing mobile communication system for railway has been unable to meet the high-capacity and high-speed demands of scheduling control, operation monitoring and passenger telecommunications. So the International Union of Railways has made it clear that, to realize the leaping development of railway wireless communication, the communication system for railway will make a direct transition to Long Term Evolution for Railway (LTE-R) from Global System for Mobile communications-Railway (GSM-R).LTE-R communication systemis facing many challenges, one of which is how to obtain the channel statement information (CSI) in complex and mutable high-speed railway scene, and then to complete the channel estimation. In high-speed railway scene, radio access is switching frequently among the base stations, proposing a fast and efficient requirement for channel estimation. On the other hand, the high-speed movement causes the obvious Doppler frequency shift, leading to serious inter carrier interference(ICI) between OFDM symbols, requiring for high accuracy and stability for channel estimation. Aiming at the above problems, it is necessary to deepening the study of channel estimation in high speed railway scene and put forward a kind ofchannel estimation strategy which can make a balance between the estimation accuracy and estimation efficiency. Only in this way the foundation for communication security and quality of service in high-speed railwaycommunications can be laid.This paper sets around two core issues, one is to research the channel characteristics in high-speed railway environment, and the other is to find a proper way to promote the efficiency and accuracy of channel estimation. The paper starts from the principles and criterions of classical channel estimation based on pilot, and discussing several commonly used principles such as Least Square (LS) and Minimum Mean Square Error (MMSE). Then it discusses the small scale fading characteristics, analysis the time selective fading and frequency selective fading and their key parameterssuch as Doppler Spread, Coherence Time, Multipath Delay Spread and Coherence Bandwidth based on the measured data. This part makes it clear the difficulties the channel estimation in high-speed railway scene is facing and the parameters that must be considered during theenvironment setting and simulation.Finally the paper researches the channel estimation based on Basic Expansion Model (BEM), studies the method of several different kinds of BEM basic function models such as Discrete Karhuen-Loeve BEM (DKL-BEM), Discrete Prolate-Sphere BEM (DPS-BEM)and Over-sampled ComplexExponentials BEM (OCE-BEM), and analyses their performance gap and application scopein high-speed railway scene. Then the strategy of LS/MMSE hybrid BEM channel estimation is promoted. The strategy firstly use the LS principles to fulfil the pre-estimation of BEM coefficients, then the correlation between adjacent OFDM symbols is used to combine the Finite Impulse Response (FIR) Filter and the MMSE principle to re-optimize the pre-estimation values of the BEM coefficients. Through the simulation, it is proved that this method can use a low BEM order (Q>2) and low FIR filter order(K>5)to achieve a relativelygoodMean Square Error(MSE) performance of channel estimation (MSE<10-3)and it is able to work out a compromisebetween the accuracy and efficiency of channel estimation by adjusting the order of BEM and the order of filter under MSEperformance threshold of channel estimation.