Cramer-Rao Bound for SAGE-based Residual Frequency Offset Estimation Algorithm in OFDM Systems

摘要

The Cramer-Rao bound (CRB) is a powerful tool in estimation theory as it gives a performance lower bound for parameter estimation problems. In this paper, a much tighter CRB for Lee's residual frequency offset (RFO) estimation method (IEEE Transactions on Communications 54:765, 2006) is first given. The tighter low bound is obtained by considering the ICI that affects the performance of space-alternating generalized expectation-maximization (SAGE) based RFO estimator. It can be concluded that the performance of SAGE based RFO estimation method decreases as the normalized RFO increases and increases with the increasing of signal-to-noise (SNR). Simulation results show that the proposed CRB of SAGE based RFO estimator is extremely tight. It approximates closely the MSE performance obtained by Monte Carlo simulation.