Accurate performance analysis of maximum likelihood frequency estimator for OFDM systems
Kai-Yi Fang, Wern‐Ho Sheen, Jeng‐Shin Sheu
Abstract
Kai-Yi Fang, Wern‐Ho Sheen, Jeng‐Shin Sheu
Abstract
In orthogonal frequency-division multiplexing (OFDM) technique, system performance is degraded greatly by the intercarrier interference (ICI) caused by the loss of orthogonality between subcarriers. One of such causes is due to the mismatch of the carrier frequency between the transmitter and the receiver. Various maximum likelihood (ML) methods have been proposed in the literature for carrier frequency estimation in OFDM systems. In theory, the performance of the ML estimators is asymptotically close to the Cramer-Rao lower bound (CRLB). However, it might be significantly different from it at the SNRs of practical interest. In this paper, a very accurate method is proposed for analyzing the performance of the ML estimators in terms of the mean and variance of the estimators. Simulation results show that the proposed method is much more accurate than the previous ones at low SNRs.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
In orthogonal frequency-division multiplexing (OFDM) technique, system performance is degraded greatly by the intercarrier interference (ICI) caused by the loss of orthogonality between subcarriers. One of such causes is due to the mismatch of the carrier frequency between the transmitter and the receiver. Various maximum likelihood (ML) methods have been proposed in the literature for carrier frequency estimation in OFDM systems. In theory, the performance of the ML estimators is asymptotically close to the Cramer-Rao lower bound (CRLB). However, it might be significantly different from it at the SNRs of practical interest. In this paper, a very accurate method is proposed for analyzing the performance of the ML estimators in terms of the mean and variance of the estimators. Simulation results show that the proposed method is much more accurate than the previous ones at low SNRs.
Key concepts: Orthogonal frequency-division multiplexing, Orthogonality, Estimator, Cramér–Rao bound, Upper and lower bounds, Transmitter, Algorithm, Interference (communication)