Research On CFO And DCO Estimator In OFDM System

Orthogonal frequency division multiplexing (OFDM) holds great potentials for the next generation wireless communication demands for its spectral bandwidth efficiency, robustness to frequency selective fading channels, etc. However, one of the major drawbacks is its sensitivity to carrier frequency offset (CFO). A CFO causes the loss of subcarrier orthogonality and hence severe performance degradation. So estimations and compensation of CFO are essential. Besides, the recent demand for low-cost mobile terminals has led to the application of direct-conversion architecture based receiver (DCR), which, however, introduces DC offset (DCO) which in turn severely degrades the transmission performance. So far, several CFO estimation approaches have been proposed. In this thesis, we firstly introduce these algorithms, including ME algorithm, ME-TDA algorithm, NBE algorithm and ML algorithm, then we verify and analyze the performance of these estimators with simulation. Among them, ME can only achieve excellent performance in the absence of DCO, while the performance of ME-TDA is not completely satisfying. On the other hand, by NBE and ML algorithms, very accurate CFO estimation results can be obtained, but the involved high computational complexity may pose a major hindrance to its practical deployment. In this thesis, we show that it is possible to construct a null space to both the CFO and DCO terms. Our particular construction makes use of the Householder transform. As a consequence, the null space approach to CFO estimation in the presence of DCO can be done just as easily as in the absence of DCO, without having to resort to SVD. Experimental results show that the proposed method can obtain the same estimation accuracy as the state of the art methods, yet with much lower complexity and more than10times the speed-up gain.