Research And Implementation Of Channel Estimation Technology For OFDM System

OFDM technology is a high-rate multi-carrier transmission system, it has highspectrum efficiency because of orthogonality subband, and inserting cyclic prefix canreduce the ISI and ICI caused by multipath delay. The capability to resist interferencehas been improved. It is convenient to use IDFT/DFT realization subband’s modulationand demodulation in OFDM system. OFDM technology has been widely used inwireless communication system. However, OFDM system is very sensitive to frequencyoffset, which causes distortion of the transmission signals during propagating inwireless mutipath channel. In order to recovery the sending signals in receiver，it isnecessary to equalize distortion of the transmission signals by channel estimation.In this thesis, we studies the transmission characteristics and the main parametersof wireless mutipath fading channel, introduces the principle of modulation anddemodulation in OFDM system and block diagram of OFDM base-band transmissionsystem, studies a lot of channel estimation algorithms deeply. It expounds the key factorto design pilot pattern by introducing various pilot pattern. We have studied severalclassic channel estimation algorithms which include LS algorithm, MMSE algorithmand LMMSE algorithm based on SVD. We analyze various interpolation algorithms. Westudy a channel estimation algorithm with PN-based iterative correlation detection whenconsidering the characteristics of pilot sequence. This paper improves an iterativechannel estimation algorithm based on superimposed pilot because of the tension ofspectrum resources. According to influence of transmission signals caused by fasttime-varying fading channel, this paper proposes an improved combination channelestimation algorithm based on preamble training sequences and pilots decision feedbackequalization. It solves channel estimation used preamble just suitable for lowtime-varying fading channels and channel estimation based on pilot interpolation reducespectrum efficiency by inserting a lot of pilots. Through the performance comparisonafter simulation analysis with MATLAB，we can know the performance of thecombination channel estimation algorithm is similar to the channel estimation usedpreamble in low time-varying fading channels, and the performance of channelestimation based on pilot interpolation is not good because of error caused by linearinterpolation. The channel estimation algorithm based on preamble training sequencescan not reflect the change of the channel effectively in fast time-varying fading channels, but channel estimation algorithm proposed in this thesis can resist influence whichcaused by multipath delay and Doppler frequency shift effectively。Finally, according to system parameters of the project, the LS channel estimationalgorithm is realized on FPGA by pilot insertion and extraction. From the signalTapⅡsimulation waveform of the receiving data of modulation after equalized，we can knowthat the result meet the system parameters requirements of the project.