Channel and frequency offset estimation for OFDM-based systems.

摘要

Orthogonal frequency-division multiplexing (OFDM) has been employed in several current and future 4-th generation (4G) wireless standards. Frequency offsets in OFDM introduce intercarrier interference (ICI). Channel estimations are also required.;For multiple-input multiple-output (MIMO) OFDM systems, channel and frequency offset estimation errors are investigated. The signal-to-interference-and-noise ratio (SINR) is first analyzed given channel and frequency offset estimation errors. The bit error rate (BER) is then approximated for multiple-antenna reception with maximal ratio combing (MRC) and equal gain combining (EGC).;For orthogonal frequency-division multiplexing access (OFDMA) systems, the variance of the frequency offset estimation is derived as a function of SINR and signal-to-noise ratio (SNR). This variance information is exploited to improve the accuracy of frequency offset estimators. A successive interference cancelation (SIC)-based frequency offset estimator is also developed.;The accuracy of frequency offset estimation of the OFDMA uplink can also be improved by using the cooperative relaying. Both conventional amplify-and-forward (AF) relays and new decode-and-compensate-and-forward (DcF) relays are studied. The frequency offset estimate is derived from combining different link estimates. In addition, when CSI is available, a scheme is proposed to adaptively switch between the cooperative and conventional (no relaying) transmissions to optimize the frequency offset estimation.;This thesis focuses on the channel and frequency offset estimation for OFDM-based systems. For cooperative-relay OFDM with frequency offsets, where inter-relay interference (IRI) exists, channel estimation is developed. Optimal pilot designs are proposed by minimizing the IRI in the mean square error (MSE) of the least square (LS) channel estimation. The impact of frequency offset on the channel estimation accuracy is derived. The pairwise error probability (PEP) with both the frequency offset and channel estimation errors is evaluated. The power allocation is discussed.