The Research Of Semi-Blind Channel Estimation Algorithm For OFDM System

The target and requirement of wireless communication technology is to transmit data fast and reliably. Orthogonal frequency division multiplexing(OFDM) technology can offer high speed transmittion and advantage of anti-fading and anti-timedelay which has been cared a lot in the field of wireless communicate application. OFDM is respected to be the key technology of mobile communication beyond 3G.This paper focuses on semi-blind estimation algorithm of wireless OFDM system. First, it introduces basic theory of wireless OFDM system and channel estimation theory relatively. Insight analyst has been done on subspace-based semi-blind channel eatimation and Expectation-maximization (EM) channel estimation. Those two methods have been simulated to test their performance. By adding a little pilot frequency, semi-blind estimation algorithm improves the convergence of blind channel algorithm. Compared with the results of direct judge method and channel estimation based on pilot frequency, under the reasonable algorithm difficulty circumstance, the longer the symbol frame the better the semi-blind estimation algorithm performance will be. The semi-blind channel estimation algorithm has been studied on the basis of Expectation-maximization (EM). It is the max-similar function iterated twice by E and M. The new channel rectify formula can be obtained by continuous refreshing E, M function with special pilot frequency insert method. The new SNR will be advanced 2dB than the traditional training sequence channel under the same BER. Then we give the Cramer-Rao Bound (CRB) of semi-blind channel estimation, and last judging its channel estimation efficiency.