Research On The Channel Estimation Algorithms In MIMO-OFDM System

MIMO-OFDM is a technology that uses multiple antennas at both transmitter and receiver with OFDM modulation on each antenna. MIMO can increase the channel capacity linearly, i.e., any arbitrary high channel capacity can be achieved by increasing the number of antennas, which thus breaks through the Shannon capacity limits of SISO (single input single output) channel. OFDM can greatly simplify the baseband processing units at receiver and improve the data rate. Thus, MIMO-OFDM has much higher system capacity than single OFDM or MIMO systems, thereby can satisfy the high data-rate and large system capacity requirements of the forth generation mobile communication.This thesis firstly analyses how to increase the system capacity through MIMO technique and basic principles of improving data rate by OFDM technique briefly, whereafter the MIMO-OFDM wireless communication system model is introduced and the theoretical upper bound of error-correction performance for space-time trellis coded MIMO-OFDM system is derived primarily. Subsequently, we investigate the training sequence based channel estimation methods. Based on the characteristic that the frequency-domain channel of MIMO-OFDM system varies slowly, we use only part of the received data in frequency domain to simplify the MMSE channel estimation method, and furthermore design the optimal training sequence and analyze the MSE bound of simplified channel estimation method proposed.Finally, we focus to investigate the blind channel estimation method of MIMO-OFDM system, i.e., subspace method. To adopt this method, the information source must be assured to be non-stationary or cyclically stationary. Through the demonstration of several theorems, we establish the blind identifiable conditions of MIMO-OFDM channel, and thereby propose a blind channel estimation method based on noise subspace. Simulation result shows that the proposed noise subspace based blind channel...