The Research And Implementation Of TD-LTE Uplink Channel Estimation&Equalization Method In High-speed Train Environment

With the development of high speed railway,the design of new generation communication system LTE(Long Term Evolution)is challenged by the larger time-varying doppler frequency shift,frequently bases changing and the demand of the high data transfer rate.Channel estimation algorithm for the high-speed train is an important research direction for the research of channel estimation for LTE.As an important part of the system,channel estimation was considered in the beginning of the design of the LTE.To reduce the complexity of receivers,pilot signals were introduced in the slot structure,by what can ensure the receiver can finish the non-blind channel estimation.On the basis of the pilot signals,how to improve the performance of channel estimation and equalization is one of the key issues to perfect the LTE uplink process.This article focuses on channel estimation algorithm for TD-LTE uplink in the high-speed train environment.By studying the existing LS(Least square)channel estimation method,the channel estimation based on the virtual channel impulse response(VCFR)expand and the channel estimation algorithm based on basis expansion model(BEM),an improved channel estimation method on the basis of the existing channel estimation algorithm based on the BEM were proposed and realized on the simulation platform and DSP platform.The main research works and innovations points are listed as follows:An improved channel estimation method based on BEM was proposed and verified.A program of designed and implemented for the channel estimation method was designed and implemented on the DSP platform.As a supplement of the TD-LTE uplink,the channel equalization method was also simulated on the simulation platform and implemented on the DSP platform with the same optimization.In this article puts forward the improved channel estimation method based on BEM,the main thought of the method is the subframe should be divided into three parts such as pilot symbols,data symbols between the pilot symbols and the data symbols out of the pilot symbols for different treatment.In order to suppress the noise effectively and improve the accuracy of channel estimation for pilot symbols,we changed the LS channel estimation for the pilot symbols in the original method based on BEM instead of the method based on virtual channel frequency response(VCFR)expanding.So it can provide an accurate basis result for the following basis expansion model.Then using the base expansion model simulate of the quick change of the channel and get the intermediate channel estimation for data symbols between the pilot symbols.Finally using the autoregressive(AR)model to modify the channel estimation of the data symbols which out of the pilot symbols.The new method can retain the advantage of existing method based on BEM that it can simulate the change of channel and can solve the problem that it can't get the accurate channel estimation of the symbols.The simulation results show the proposed method compared with the original method,can get more accurate channel estimation results.Combining with the characteristics of MSC8157 DSP platform,multi-core task allocation scheme were designed for the channel estimation.On this basis,fully considering the accuracy and computing speed a reliable fixed point solution for the method was designed.Finally through the optimization method,compressed the channel estimation method's time cost.The test on the hardware platform shows that the proposed improved method in this paper is a channel estimation method has engineering practical value.