The use of Basis Expansion Model (BEM) for channel estimation is proposed for Orthogonal Frequency Division Multiple Access (OFDMA) system uplink. After formulating a BEM-based signal model for OFDMA Uplink transmission over time and frequency selective fading channels, we develop a pilot-aided channel estimation scheme for user channel information tracking in the time domain. Specifically, a Linear Minimum Mean Square Error (LMMSE) and a Least Square (LS) estimator have been investigated for channel estimation under the help of time-domain pilots. In particular, the proposed pilot based estimation scheme does not require a whole OFDMA block of pilot to estimate channel.
Cramer-Rao Lower Bound (CRLB) for LS (LS-T) and the MSE lower bound for LMMSE (LMMSE-T) estimator have been derived.It is shown that in very high Doppler spread scenario, LS-T estimator performs as the optimal unbiased estimator and chases the CRLB, while LMMSE-T minimizes the overall MSE and gains more improvement over unbiased estimator. Simulation results are presented to show the performance improvement that can be achieved by the proposed estimation algorithm over current notable schemes.