(A) channel estimation method for OFDM-CDMA systems in fading channels

Abstract

In this thesis we propose and analyze a method of estimating the channel transfer function for orthogonal frequency division multiplexing-code division multiple access (OFDM-CDMA) systems under slowly time-variant radio channel conditions. The proposed channel estimation method is based on pilot symbols inserted both in time and frequency domains and a truncated DFT-zero padding interpolation algorithm. Owing to the low-pass filtering in a transform domain, additive white Gaussian noise in the received pilot signals is reduced considerably. The optimal cut-off frequency of the transform domain filter is selected according to the number of multipaths. After the truncated DFT-zero padding interpolation in the frequency domain, we also apply it to the time domain to obtain the time-variant channel response in each subchannel. We present analytical expressions of mean-squared error (MSE) for various interpolation methods. The proposed channel estimator not only qualifies the MSE performance comparable to that of linear minimum mean-squared error (LMMSE) estimator without knowing channel correlation but also has much lower computational complexity. Computer simulations over Rayleigh fading channels demonstrate an improvement in $E_b$/$N_o$ for same bit error rate (BER).