Parametric Channel Estimation And Statistics Analysis For OFDM Systems In Time-Varying Multipath Fading Channels

In this thesis, we addressed several problems about the parametric channel estimation and statistics analysis for orthogonal frequency division multiplexing (OFDM) Systems in time-varying multipath fading channels.First of all, we proposed a parametric channel estimation algorithm applicable to uplinks of OFDMA systems whose subcarriers are pseudo-randomly allocated. By exploiting pilot hopping, estimation of signal parameters via rotational invariance technique (ESPRIT) is employed to estimate the sparse path delays. From the delay information, a global channel interpolator is given to estimate the channel state information on the desired tones. Simulation results demonstrate that the proposed algorithm outperforms the local linear channel interpolator within a wide range of signal-to-noise ratios (SNRs) and Dopplers.Then, we proposed a maximum Doppler spread (MDS) estimation algorithm for OFDM systems with the comb-type pilot pattern in doubly selective fading channels (DSFC). With two auto-correlation matrices with different lags, a non-linear high-order polynomial equation is builded, from which MDS is readily solved by using the Newton's method. The delay-subspace is utilized to reduce the noise bias. And the subspace tracking algorithm is adopted to automatically update the delay-subspace. Simulation results demonstrate that the proposed algorithm is robust and converges over a wide range of SNRs and MDSs.After that, we derived and analyzed the analytical expression of the mathematic expectation of the sample frequency correlation matrices (SFCMs) in DSFC for OFDM systems. Assuming the equiprobable centro-symmetric constellation for modulation, SFCM was proved to be biased due not only to the noise, but also to the irreducible inter-carrier interference aroused by the Doppler spread. Further investigations revealed that the eigen-structures of SFCMs suffer various impact for different types of modulation. Numerical results demonstrate this bias-property and the related analysis.Afterwards, the Cramer-Rao lower bound (CRLB) of FCM is derived in DSFC for OFDM systems. By assuming a random but fixed pilot sequence, SFCM is complex Wishart distributed. Then, the maximum likelihood estimator (MLE) and the analytical expression of CRLB are obtained. From CRLB, the average mean squared error (AvgMSE) is deduced, which reveal that the amount of samples is the dominant factor affecting AvgMSE while SNR and MDS have negligible effect. Numerical simulations demonstrate the analytic results.At last, we derived the CRLB for spatial correlation matrices (SCM) in DSFC for multiple-input multiple-output (MIMO) OFDM systems. Adopting an orthogonal pilot pattern for multiple transmitting antennas, SSCM is complex Wishart distributed. Then, the MLE and CRLB are derived by assuming that temporal and frequency correlations as well as noise power are known. Furthermore, a lower bound of AvgMSE is deduced from CRLB, according to which the amount of samples and the order of frequency selectivity show dominant impact on the accuracy of estimation. Besides, the number of pilot tones, the power delay profile, SNR and normalized MDS together influence the effective order frequency selectivity. Numerical simulations demonstrate the analytic results.