A survey on OFDM channel estimation techniques based on denoising strategies

Channel estimation forms the heart of any orthogonal frequency division multiplexing (OFDM) based wireless communication receiver. Frequency domain pilot aided channel estimation techniques are either least squares (LS) based or minimum mean square error (MMSE) based. LS based techniques are computationally less complex. Unlike MMSE ones, they do not require a priori knowledge of channel statistics (KCS). However, the mean square error (MSE) performance of the channel estimator incorporating MMSE based techniques is better compared to that obtained with the incorporation of LS based techniques. To enhance the MSE performance using LS based techniques, a variety of denoising strategies have been developed in the literature, which are applied on the LS estimated channel impulse response (CIR). The advantage of denoising threshold based LS techniques is that, they do not require KCS but still render near optimal MMSE performance similar to MMSE based techniques. In this paper, a detailed survey on various existing denoising strategies, with a comparative discussion of these strategies is presented.