An Improved Subspace Based Semi-blind Channel Estimation For STBC-MIMO-OFDM System

1. IntroductionThe target of broad band wireless communication is to transmit data fast and reliably, while there are two tough problems in front of us, they are multipath fading channel and bandwidth efficiency. Orthogonal Frequency Division Multiplexing (OFDM) technique changed the frequency selective multipath fading channel into flat fading channel in frequency domain which effectively mitigates the effects of multipath propagation and, hence, increase data rate. In additional, Multiple Input Multiple Output (MIMO) technique can generate independently parallel multi-channel data stream in spatial at the same time, which directly and effectively increase the transmission efficiency. Therefore, combining MIMO and OFDM is believed to have the ability to achieve two purposes, the one is high transmission speed, the other is strongly reliability.MIMO-OFDM technique brings the increased complexity of system design with the improvement of system performance. In MIMO-OFDM system, each receive antenna receives the signal that is the superposition of mixed signals transmission through the MIMO channel. With the purpose of offsetting the loss in transmission and identify the transmitted signal of each transmitting antenna, it is necessary to holding the knowledge of time-varying channel transmission function between antennas. If utilizing the algorithms used in OFDM system to some certain pair of antenna, the interference signal will often result in the SNR below 0dB and directly lead great error. Consequently, channel estimation of MIMO-OFDM system is a research area of great challenging and significance.The channel estimation methods can be divided into three categories, they are non-blind channel estimation, blind channel estimation and semi-blind channel estimation. The non-blind channel estimation generally is applicable to the continuous transmission or burst transmission system. The algorithm based on pilot symbols or trained sequence, the receiver achieves the initial estimate. When the system sends the useful information, the algorithm will use the results of the initial decision to update and complete the real-time channel estimation. This kind of algorithm has good capability and is easy to realize, but it decreases bandwidth and transmission efficiency. The biggest different between non-blind and blind channelestimation is that the blind method is based on the limited characters of the transmitted information symbols and their statistical trait. So this kind of algorithm has the most frequency efficiency but slow converge, complicated computation, and can not be effectively tracking the channel. Using the information from blind channel estimation algorithm and known sampling symbols to finish channel estimation, semi-blind channel estimation solves the problems of spectrum waste from none-blind channel estimation and high complication of blind channel estimation methods. Hence the semi-blind channel estimation algorithm for MMO-OFDM system is regarded to be a promising way for channel estimation.2. An improved subspace based semi-blind channel estimation for STBC-MIMO-OFDM systemTo overcome the inadequacies of the original algorithm, this paper proposes an improved subspace based semi-blind channel estimation for STBC-MIMO-OFDM system in section 5.3.2. The main idea is that in the original method, the receiver used the statistics trait of the signal and replaced mathematical average by statistical average to calculate correlation matrix of received signal, then got the noise subspace by subspace method. While the proposed improved method uses the diagonalization of the overall covariance matrices of both received signal and noise signal, accordingly, the noise signal subspace is estimated. Then based on the knowledge that the noise subspace and the channel impulse response are orthogonal, all channel responses of a STBC-MIMO-OFDM system can be identified blindly subject to two ambiguity matrices. A method is then presented to resolve the two ambiguity matrices by using few pilot symbols in LS algorithm. In the end the channel impulse response is estimated. With the estimated channels, a frequency domain approach is presented to recover the transmitted symbols. The improved method increases accuracy, thereby directly improves the performance of entire algorithm.3. Simulation and ConclusionThe simulations of comparative experiments of the improved algorithm on computer with MATLAB are performed. The first experiment under the Jake model channel and a multiuser STBC-MIMO-OFDM system is considered here. The original method, the proposed improved method and the optimal method are compared in different aspects. Another experiment is a single-userSTBC-MIMO-OFDM system with"bad"channels (not coprime channels). The results of the two experiment show that the improved method not only inherits the advantages of the original one, such as not rely on the channel order (only an upper bound for all the channel order is required), valids for multiuser system, but also improves the estimated accuracy and performance, especially in middle SNR.