Study Of The Channel Estimation Algorithm Based On MIMO-OFDM Systems

MIMO-OFDM technology is the core technology of next generation of mobile communications. To realize MIMO-OFDM technology, channel estimation plays an important part, as well as the most difficult and hottest one. At the receiver of the wireless communication system, the ultimate goal is to restore communication signal sent by the transmitter. To achieve a balanced recovery, parameters of the original signal channel is needed. Therefore, more reliable and stable channel estimation algorithms are required to ensure communication quality in a communication system.To study the estimation algorithm based on MIMO-OFDM channel is the main content of this thesis, comparative analysis of the advantages and disadvantages of various algorithms, and figure out algorithms selection rules. In this thesis, BER and MSE are treated as references index. Firstly, this thesis describes the basic characteristics of MIMO-OFDM systems, and puts forward the MIMO-OFDM channel model. Comparing the algorithm simulation results of the LS(Least Square) and LMMSE(Linear Minimum Mean Square Error) based on pilot frequency, using FFT(Fast Fourier Transform) transformation and windowing technology in view of the LS algorithm not taking into account the noise impact, to propose the improved LS algorithm and reduce the influence of noise to estimate the channel matrix. For the larger computational problem caused by LMMSE’s inverse algorithm, this thesis proposes an improvement program based on LMMSE algorithm. To decompose the channel matrix some mathematical methods is used, with time-domain windowing to achieve frequency domain filtering,to reduce the algorithm’s computation without decreasing the performance of the algorithm. The algorithms simulation for data subcarriers are compared, such as linear interpolation algorithm and temporal interpolation algorithm, for the problem of temporal interpolation algorithm this thesis use mathematical methods of filtering technology to solve it. Linear interpolation algorithm and temporal interpolation algorithm data-carrier is simulated in comparison. In contrast to the blind and semi-blind estimates simulation based on subspace decomposition, based on QR decomposition and linear pre-coding, combined with the semi-blind estimation method based on subspace decomposition and joint estimation method based on linear pre-coding, put forward an improved semi-blind estimation algorithm for joint estimation based on linear pre-coding. In order to obtain an initial value of the correlation matrix, a small amount of pilot symbols into OFDM symbol before pre-coding is inserted, constructing an iterative formula to eliminate the phase ambiguity in the system, speeding up the convergence of the system.Through extensive simulations, it can be proved that the algorithm has excellent performance. With the increase of SNR(signal to noise ratio), the BER and MSE both shrink. Finally, the LS algorithm based on pilot frequency and first-order linear interpolation algorithm are implemented in the DSP, and compared with the results of simulation, to the conclusion that determine the application feasibility of the algorithm.