State-Space Estimation for OFDM and SC-FDE Schemes with Strongly Varying Carrier Frequency Offset

Abstract

Parametric algorithms for the estimation of rapidly-varying Carrier Frequency Offset (CFO) usually employ pilot symbols multiplexed with the data transmission. As the CFO variation rate increases so has to increase the density of pilot symbols transmitted, thus impairing the bandwidth efficiency. In order to reduce the number of pilot symbols used in the estimation of rapidly-varying CFO it was proposed to use a truncated Taylors series to predict the CFO, where the derivatives up to order d-1 are recursively estimated with a d-order Kalman filter (KF).We propose to compare the performance of a fourth-order KF predictor in the most popular block transmission systems: Orthogonal Frequency Division Multiplexing (OFDM) and Single Carrier Frequency-Domain Equalization (SC-FDE). Simulating different transmission scenarios, e.g., channel coding and spatial diversity, our results show that for static multipath fading channels the proposed receiver for the SC-FDE scheme exhibits better Bit Error Rate (BER) performance than that of OFDM.