Abstract:

In fast time-varying OFDM（orthogonal frequency division multiplexing）system, in order to lessen the effects of the CIR（channel impulse response）leakage and ICI（inter-carrier interference）, a new channel estimation algorithm was proposed, which was adapted to non-sample-spaced channels. The algorithm combined complex exponential basis expansion model and the feedback of DFT（discrete fourier transform）, which was built according to the fraction taps channel approximation.The channel parameters were firstly estimated with the complex exponential basis expansion model based on the fraction taps channel approximation.Then the ICI in fast time-varying OFDM system could be calculated according to the estimated channel parameters. Then the feedback was obtained by the elimination of ICI, and noises could be eliminated by DFT. The algorithm could suppress the CIR leakage, eliminate ICI and noises and effectively approximate the actual channels. The simulation results show that the bit error rate and channel mean square error of the proposed algorithm are both effectively improved.

Key words: non-sample-spaced channel, fast time-varying, OFDM, channel estimation, basis expansion model, fraction taps channel approximation, DFT