Research On Pilot-Aided OFDM Channel Estimation Methods

OFDM (Orthogonal Frequency Division Multiplexing) is a multi-carriers modulation technique. It transmits information through orthogonal multiple sub-carriers and has good capability of countering multi-Path interference, high utilization of frequency spectrum. OFDM is concerned by the field of communication, being regarded as an important component in the fourth generation. However, due to the transmission characteristics of wireless channel and high-bandwidth nature of OFDM system itself channel estimation becomes a very important part in OFDM system. OFDM channel estimation should meet the requirements of faster transfer rate and estimate to be accurate.This paper mainly researches on the channel estimation based on pilots. Firstly analyzes the transmission properties of general wireless channel and presents its mathematics model. Then analysis the basic principle and realization of OFDM systems, describes the system model, discusses some key problems in OFDM system. Based on these analyses, we introduce block-type and comb-type and pilot arrangement, then three channel estimation criterions of LS, MMSE and SVD and three interpolation methods of linear interpolation, Gauss interpolation and interpolation of zero-padding in transform domain based on DFT are respectively discussed, in which the interpolation method of zero-padding in transform domain based on DFT is especially considered.An improved algorithm for zero-padding in transform domain based on DFT is proposed in this paper to overcome the disadvantages of traditional transform domain algorithm. Compared with the traditional transform domain based on DFT algorithm, the improved algorithm by using window function to minimize the spectrum leakage, using dynamic filtering to decrease the effects of noise on OFDM system and introducing base compensation to make up for timing error in system. The simulation results demonstrate that the improved algorithm based on DFT has better performance and higher stability than traditional transform domain based on DFT algorithm. However, if the original sequence is disconnected, additional higher-order harmonics will produced when data are transformed by DFT. Those higher-order harmonics is likely to cause spectrum aliasing. There proposed an improved transform domain algorithm based DCT to minimize higher-order harmonics as much as possible. This improved transform uses DCT instead of DFT., DCT has higher energy compressing property than DFT., can availably reduce estimation error caused by higher-order harmonics. The simulation results show that the improved algorithm based on DCT has better estimation precision and stability.