Research On Estimation And Equalization Of Fast-changing Channels In OFDM Communication Systems

The coherent demodulation in the OFDM system requires the current channel frequency response obtained through channel estimation to recover the data transmitted by the transmitting end,so the channel estimation is closely related to the performance of the OFDM system.The performance of channel equalization is directly related to the degree of suppression of inter-carrier interference in OFDM systems.Therefore,channel estimation and channel equalization techniques are the key to OFDM implementation.This dissertation focuses on channel estimation and equalization in fast-varying channels.The main tasks are as follows:(1)This paper first introduces the basic principles,modulation and demodulation methods and corresponding key technologies of OFDM systems.It analyzes the causes of multipath and Doppler effects in wireless channels,and discusses common basis expansion model.The advantages of the model and the basis expansion model in fast-varying channel estimation are studied and discussed.(2)With the increase of the channel time delay spread and the multipath number in mobile communication system,which leads to a large number of BEM coefficients need to estimate in the time-variant channel estimation.so consequently more pilot symbols are needed,which will seriously reduce the spectrum utilization.In order to effectively decrease the insertion of pilot symbols,we inserted discrete pilot symbols in the frequency domain direction.Then a cyclic iterative algorithm based on the Kalman filter is used in this paper,the data of the symbol corresponding to the position are estimated to be the virtual pilot.This algorithm can effectively reduce the use of the pilot symbols and greatly improve the spectrum utilization rate compared with the continuous comb-type estimation method.(3)The channel equalization technology mainly considers two aspects of equalization performance and computational complexity.The MMSE and other equalization criteria are required to execute the matrix inverse operation.In order to effectively avoid the inversion of the matrix and reduce the complexity of the equalization operation,we proposed a channel equalization algorithm based on QR decomposition,and for different normalized Doppler frequency shifts,the performance of the equalization algorithm based on the LS criterion and equalization algorithm base on the QR decomposition is compared.The simulation results show that the QR decomposition-based equalization algorithm has superior performance in the channel environment where the Doppler shift is severe.