Pilot-aided multipath delay spread approximation for MMSE channel estimation in OFDM systems

Minimum mean square error (MMSE) channel estimation in OFDM systems requires knowledge on channel correlation function. For systems with a small number of pilots operating in time-varying channels, however, information on channel correlation is difficult to acquire. In this paper, instead of estimating power-delay profile (PDP) with one of the three delay spread parameters: mean delay, root-mean-square (RMS) delay, and maximum delay spread, we propose an efficient method to approximate the multipath delay spread by the curvature of channel's amplitude, which is then used to generate the Wiener filter coefficients for data subcarriers channel estimation. The proposed technique could be applied to many prevailing wireless communication standards, and the performance of our proposal is investigated via simulations based on IEEE 802.16m packet structure.