Study On Channel Estimation Algorithm Based On Pilot In OFDM System

Modern life more and more depend on communication technology, and the requirement of communication system is more and more complex, but the transmission rate of communication system and the system error rate is always the most concerned. For its high transmission capacity, effective against frequency selective fading and high utilization rate of spectrum, Orthogonal frequency division multiplexing (OFDM) system has been widely applied in various communication systems. OFDM system generally adopts the coherent demodulation, the channel estimation techniques attract more and more attention.This paper introduce the various classical channel estimation algorithm based on pilot, and through the MATLAB simulation to realize bit error rate (SER) and mean square error (MSE) performance curve of least-squares (LS), linear minimum mean square error (LMMSE), LMMSE algorithm based on singular value decomposition (SVD). LS algorithm is the most basic channel estimation algorithm for its low computational complexity, simple implementation, since estimation accuracy of the LS algorithm is not high. For the use of channel two-dimensional statistics prior information, LMMSE algorithm improves estimation precision, but its computation complexity is high, and it is difficult to get channel two-dimensional statistical prior information, LMMSE algorithm is limited in practical application. LMMSE algorithm based on SVD reduced the computational complexity compared to LMMSE algorithm, but it still need to get channel two-dimensional statistical prior information.In connection with the LMMSE algorithm requires channel two-dimensional statistical prior information, this paper proposes an improved LMMSE algorithm based on window function method. This algorithm can get the required parameters by sliding window and avoid to calculate each radial signal delay and power calculation. It is more practical in the actual system. In the condition of require of estimation accuracy is not too high, this algorithm has more advantages. The simulation results show that the performance of SER algorithm is close to the ideal condition of equivalent linear minimum mean square error (ELMMSE) algorithm.