Research On Channel Estimation Algorithm Based On FFT In OFDM System

OFDM technique is a kind of multi-carrier transmission technique. With higher spectral efficiency and better anti-interference ability, OFDM has been a hot research area and also widely adopted by a variety of wireless local area network standards in practical use. Since the wireless channel is complex and the receiving terminal in OFDM system needs coherent demodulation, channel estimation of OFDM system has been a necessary part. Channel estimation is a key technology of OFDM system. Among all the estimation algorithms, the channel estimation based on pilots is a perfect compromise between algorithm complexity and performance that can be applied to practical use. This paper mainly focuses on the research on the channel estimation algorithm based on pilots.Firstly, the technology of OFDM and its principle are introduced, and then the characters of the wireless channel are analyzed. Secondly, the channel estimation algorithm based on pilots is given introduction in detail, including the study on the design of pilot structure, the criterion of pilot channel estimation and interpolation algorithm, and the performance comparisons of channel estimation based on different criterion are made through computer simulation as well as interpolation algorithm. Considering the lower algorithm complexity and better performance of channel estimation based on FFT, this paper mainly researches the FFT-based channel estimation and its improved algorithm.(1) The research on one-dimensional channel estimation based on FFTThis paper proposes two improved algorithms to the one-dimensional channel estimation based on FFT. For the first improvement, we mainly research the improved algorithm in domain, including the improved filter and the improved interpolation. Of the improved filter, we get the time domain channel response through IFFT after obtaining the whole frequency domain channel response, to the results of which performing linear transformation before through an estimator A, and then we can get the frequency domain channel response after FFT. Of the improved interpolation, we perform linear transformation befor zero-padding in time domain, and then get the FFT results.Simulation results prove the better permormance of the two improved algorithms in time domain. For the second improvement, since the noise has been existed, we implement noise reduction operations after LS channel estimation, then getting the frequency domain channel estimation of noise deduction through FFT. Simulation results illustrate the efficiency of the algorithm.(2) The research on two-dimensional channel estimation based on FFTFor the higher complexity of the traditional FFT-based united algorithm and its inconvenience in practical use, some scholar proposed the improved algorithm which interpolation is operated once respectively in time domain and frequency domain, the results of which bring the lower algorithm complexity and just a little loss in performance.On the basis of above, we study the FFT-based channel estimation by interpolating twice respectively in time domain and frequency domain, and the results show the better performance. In view of the effect of noise that the interpolation and the wireless channel bring, the improved algorithm is proposed. By designing a filter in time, the interpolating result is processed. The results prove better performance through the simulation of the improved algorithm and the consideration of the effect of different parameters.