Improved channel estimation using noise reduction for OFDM systems

A new channel estimation algorithm for OFDM systems is proposed in this paper based on a noise reduction method. The performance of a least squares (LS) channel estimator is improved by using the colored noise characteristics in the LS estimation of the channel impulse response (CIR). The noise, which is in the CIR (signal) subspace, is estimated and suppressed based on the correlation between two parts of noise, which are within and outside the signal subspace. Knowledge of the CIR statistical parameters is not needed in the proposed estimation algorithm as long as the CIR length is not longer than the cyclic prefix interval.