Doubly selective channel estimation in OFDM using discrete prolate spheroidal sequences

Time varying channel breaks the orthogonality between the subcarriers of Orthogonal Frequency Domain Multiplexing (OFDM). This introduces Intercarrier Interference (ICI). One of the ways to mitigate ICI is to first estimate the channel and then equalize it. In this paper we propose to estimate the channel using Discrete Proplate Spheroidal Sequences (DPS) using pilot carriers interleaved with data carriers in a single OFDM symbol. We show that the Normalized Mean Square Error (NMSE) using DPS is lower at high SNR compared to the one obtained using Complex Exponential basis. We also show how iterative approach can be used to estimate data and calculate average Bit Error Rate (BER).