Efficient Scheme for Channel Estimation in OFDM System

The choice of channel estimation is of fundamental importance in OFDM receiver designs, which further involves a trade-off between complexity and estimation accuracy. In pilot-symbol-aided OFDM system, the Least Square (LS) estimator suffers from inherent additive Gaussian noise and Inter Carrier Interference (ICI), although the estimator exhibits lower complexity and requires implicit knowledge of the channel. In comparison, the Minimum Mean-Square-Error (MMSE) estimator shows much better performance than the LS. However, a major drawback of the MMSE is its higher computational overhead, which further grows with increasing the number of pilots. Accordingly, the optimal design of channel estimators has remained an area of ongoing research. This study performs modifications to both existing LS and MMSE algorithms, followed by proposing two new channel estimators, namely Simplified Least Square (SLS) and Simplified Minimum Mean Square Error (SMMSE). Mathematical analyses and simulation results show that the proposed SLS method is more robust against the additive Gaussian noise and outperforms the original LS estimator. Moreover, this method can perform almost similar to the MMSE for a range of SNRs. Meanwhile, in our study the proposed SMMSE method requires only a minimum knowledge of the Channel Impulse Response (CIR) to efficiently estimate the channel. Additionally, the SMMSE exhibits computational complexity to be significantly lower than that of the original MMSE estimator.