摘要：This paper presented novel Carrier Frequency Offset (CFO) acquisition together with multi-antenna channel estimation algorithms for burst-mode Multiple-Input Multiple-Output combined with Orthogonal Frequency Division Multiplexing (MIMO-OFDM) systems. The new algorithms take full advantage of some known channel information (for example, the length of channel impulse response L which can be estimated easily and explicitly by systems) to reduce computational complexity, the impact of noise and interference. Special time-domain training sequences were designed to accomplish these algorithms in simple and efficient ways. Firstly, the paper derived a precise mathematical model and the coarse CFO acquisition in which the influence of unknown multipath Channel Impulse Response (CIR) can be removed. Then mathematics seal solution Maximum Likelihood (ML) detection for coarse CFO is formulated instead of EM iteration detection. Secondly, time-domain CIR estimation algorithm is formulated with the advantage of noise and residual CFO reduction. Moreover, the new CIR estimator adopts matrix dimension reduction algorithm to reduce computational complexity. Thirdly, to improve the CFO acquisition performance, improved Expectation Maximization (EM) algorithm is employed as fine CFO acquisition. Simulation results indicate that compared with classical EM algorithm, the performance and convergent speed of EM iteration are both improved.
