Research On Channel Estimation Algorithm For LTE Based On Basis Expansion Model

Adopting orthogonal frequency division multiplexing(OFDM)and multiple-input multiple-output(MIMO)antennas as the key technologies,the Long Term Evolution(LTE)communication system has lots of advantages such as high spectral efficiency,large channel capacity,and hence has a broad application prospect.However,the requirement of providing high speed data transmissions to high speed mobile terminals,which is put forward by the high speed railway communication system,is a new challenge for LTE.Existing research results have shown that channel estimation is an effective method to enhance system performance and guarantee normal communication.In this thesis,several classical channel estimation algorithms are discussed,and the channel estimation algorithm utilizing the basis expansion model(BEM),which is suitable for high speed mobile scenario,is investigated.Moreover,the performances of the above algorithms are analyzed by extensive simulations.(1)Slow time-varying(TV)channel estimation algorithms are investigated as the basis of fast TV channel estimation.First,the design schemes,the structure features,and the application scenarios of block-type pilot,comb-type pilot,and bulk-type pilot are discussed in detail.Next,the bit error rate(BER)and the mean square error(MSE)of least squares(LS)algorithm,minimum mean square error(MMSE)algorithm,and linear minimum mean square error(LMMSE)algorithm are analyzed by using the COST-207 rural multipath channel model.Result shows that compared with the algorithms of MMSE and LMMSE,LS algorithm has lager estimation error because the channel statistical characteristic is not used and the influence of the noise is ignored.However,it has also been widely applied on account of the dominant advantages of small computational load and low computational complexity.(2)Fast TV channel estimation algorithm based on the BEM is investigated.Firstly,the architecture of high speed railway communication system is established based on Rician fading channel model,and the reason of the infeasibility of using classical channel estimation algorithms for fast TV channel estimation is analyzed.Secondly,the principle of approximating fast TV channel using the BEM is studied,and the MSE of the complex-exponential BEM(CE-BEM),the generalized CE-BEM(GCE-BEM),the polynomial BEM(P-BEM),and the discrete cosine transformation BEM(DCT-BEM)is evaluated via simulation.Furthermore,the influences of the number of the basis functions and the Rician factor are discussed.Results show that the P-BEM susceptible to the Doppler effect has lager modeling error,and the GCE-BEM possesses more excellent performance.Furthermore,the estimation error could be reduced effectively via increasing the number of the basis functions,and the algorithm still has satisfactory performance in severe fading channel.(3)Estimating the coefficients of the basis functions using equally spaced comb-type pilot cluster and LS algorithm is investigated.Firstly,the matrix of the coefficients of the basis functions is isolated using matrix transformation.Secondly,equally spaced comb-type pilot cluster is designed according to the fast TV characteristics of the channel.Then,pilot cluster and LS algorithm are employed to solve the coefficients of the basis functions.Finally,the CE-BEM and the GCE-BEM are taken as research objects to evaluate the performance of the MSE through system simulation.Results show that since the inherent drawbacks of LS algorithm,some estimation error exists while using LS algorithm to estimate the coefficients of the basis functions.Even so,it also has the feasibility.Besides,the performance of using LS algorithm to estimate the coefficients can be improved through increasing the number of the basis functions.