| The simplified LMMSE channel estimation algorithm for OFDM system is studied in thispaper. Firstly, the complex components of LMMSE channel estimation algorithm have beenanalysed. Secondly, the complexity of channel correlation matrix and the inversion matrix hasbeen reduced respectively. Besides this, a simulation linkage has been established based onthe software of Matlab, by using which, the proposed simplified LMMSE channel estimationalgorithm’s performance has been verified. The main contents include:(1) A novel simulation linkage based on the protocol of IEEE802.16a has been built,under the Matlab environment. It mainly includes3models. The first model can realize thefunctions of generating transmitted data, modulation, inserting pilots and CP. The secondmodel is used to build the ITU and3GPP channel models, which including the characteristicsof multi-path delay and rayleigh fading. The third model can take the operation ofdemodulation and classifying the received data.(2) A novel simplified LMMSE channel estimation algorithm has been proposed. Firstly,the periodic characteristic of the complex exponential function has been taken advatage tosimplify the solution of the channel autocorrelation matrix. Secondly, the simplification ofinverse matrix in LMMSE algorithm has been realized since the matrix satisfied the feature ofconjugate symmetry toeplitz, which is more simple than the conventionial choleskydecomposition method. It is proved by simulation that compared with the contrast algorithms,the computing times of complex multiplication and complex addition in solvingautocorrelation matrix are reduced by31.78%and16.49%respectively. When solving theinverse matrix, both the computing times of complex multiplication and complex addition arereduced from o(N3) to o(N2). |
PDF Full Text Request