Carrier frequency offset estimation for multicarrier communications

by Li, Cheng

(Uncorrected OCR) Abstract of thesis entitled Carrier Frequency Offset Estimation for Multicarrier Communications Submitted by Li Cheng For the degree of Master of Philosophy at the University of Hong Kong in Feb., 2004 Multicarrier (MC) transmission is a promising technology in broadband wireless access involving high-throughput data transmission. MC communications are advantageous over single-carrier transmission in dealing with inter-symbol interference (ISI) due to the channel delay spread. However, MC communications also have certain drawbacks, including the vulnerability to carrier frequency offset (CFO) arising from the difference of oscillator frequencies between the transmitter and the receiver or from the Doppler shift. The performance of MC systems degrades severely even if the CFO is a small fraction of the subcarrier spacing. Therefore, correction of CFO is very important for MC communications. This thesis considers two variants of MC technologies, OFDM and MC-CDMA, and investigates CFO estimation for both systems. For OFDM systems, a semi-blind CFO estimation algorithm utilizing the embedded pilot tones, which are primarily used for channel estimation, is proposed for slowly-varying multipath channels. The algorithm is bandwidth efficient because it requires no training symbols. It is also demonstrated that the proposed algorithm can be regarded as the general form of the virtual-carrier-based CFO estimation algorithm previously proposed in the literature. For the purpose of practical implementation, a reduced-complexity estimator is also developed. It offers a tradeoff between estimation accuracy and implementation complexity. It is shown by simulation results that the proposed algorithm outperforms the cyclic-prefix-based algorithm in terms of the estimated CFO mean-square error performance. For MC-CDMA systems, a blind CFO estimator is proposed for the downlink transmission, where all users?signals are synchronously transmitted. The estimation scheme consists of two stages: feed-forward acquisition and error-feedback data-directed tracking. The acquisition stage is implemented utilizing the orthogonality among spreading codes and it is capable of bringing the CFO as large as the signal bandwidth to a small value. The residual CFO is refined by the tracking stage, which estimates the CFO and users?data simultaneously based on the maximum likelihood criterion. A reduced-complexity estimator based on the zero-padding FFT technique is also developed to reduce the implementation complexity. The performance of the proposed algorithm is investigated by extensive computer simulations.