Research On Pilot Signal Design And Channel Estimation Technique For MIMO-OFDM Systems

Data transmission in wireless channels with high speed and reliability is required in future wireless systems. With the parallel modulation, orthogonal frequency-division multiplexing (OFDM) can change a frequency selective fading channel into multiple flat fading channels. Multiple-input multiple-output (MIMO) system which utilizes multiple transmit/receive antennas can increase the channel capacity significantly and therefore greatly improve the spectrum effectively. As a combination of OFDM and MIMO techniques, MIMO-OFDM technology which can not only enhance the transmission rate and capacity of the wireless communication system significantly, but also combat multi-path fading and interfere effectively, has recently received considerable attentions and considered as the core technique of the next generation mobile communication systems. Channel estimation technique is regarded as a key technology due to that the channel state information is necessary in MIMO-OFDM systems for effective detection. In this thesis, the channel estimation technique in MIMO-OFDM systems is investigated, with the focus on the pilot-based channel estimation algorithms.First, the background and development of OFDM and MIMO technologies is presented. Then the characteristics of communication channel and the principle of OFDM, MIMO, MIMO-OFDM system are discussed.Second, the channel estimation techniques for SISO-OFDM systems are investigated. The mode of pilot are studied, and the performance of LS algorithm and MMSE algorithm is compared and analyzed, then some interpolation algorithms are discussed. The shortcoming of the second order interpolation is analyzed and a modified algorithm is presented to improve the second order interpolation. The analysis and comparison results obtained through computer simulation show that the modified interpolation algorithm provides better performance.Finally, the channel estimation techniques with 2 transmit and 2 received antennas for MIMO-OFDM systems are studied. In view of the contradiction of the time-domain least squares (LS) channel estimation performance and the practical realization complexity, a novel reduced complexity channel estimation method for MIMO-OFDM based on pilot is proposed. This approach can transform the complexity of MIMO-OFDM channel estimation problem into SISO-OFDM channel estimation problem and therefore there is no need for large matrix pseudo-inverse, which greatly reduces the complexity of algorithms. Simulation results show that the bit error rate (BER) performance of the proposed method with time orthogonal training sequences and linear minimum mean square error (LMMSE) criteria is better than that of time-domain LS estimator and nearly optimal performance. What's more, the channel estimation techniques with 4 transmit and 4 received antennas for MIMO-OFDM systems are studied. Under specific pilot design, a novel algorithm for 4 transmit and 4 received antennas is proposed. The simulation results show that the proposed algorithm has better performance.