| Multiple-input multiple-output (MIMO) systems have greater capacities than the conventional single-input and single-output (SISO) systems. This capacity increases linearly with the increase of antenna pair number. Diversity and multiplexing techniques are employed to improve reliability and efficiency of communication systems, respectively. Orthogonal frequency-division multiplexing (OFDM) splits a wide frequency band into a bank of narrow sub-bands. Although the whole bandwidth may subject to frequency selective fading, each sub-band undergoes flat fading. After the insertion of cyclic prefix (CP) before each data symbol, OFDM can overcome Inter Symbol Interference (ISI) caused by multipath channels. MIMO and OFDM are seen as the essential techniques for future communication systems. As the combination of multiuser technique and OFDM, Orthogonal frequency-division multiple access (OFDMA) technique has the same advantages of OFDM. Moreover, frequency diversity and multiuser diversity can be obtained, and multiple access interference can be eliminated very well in OFDMA. Therefore, it will be a very promising research and application topic.This paper addresses the frequency synchronization problem for two OFDM based systems that is MIMO-OFDM and OFDMA uplink systems, respectively. As for the carrier frequency offset (CFO) estimation in MIMO-OFDM systems. Two Maximum Likelihood Estimators (MLEs) and four best linear unbiased Estimators (BLUEs) are proposed. Periodic preambles are used for synchronization. In the MLE, signal from multiple antennas are superimposed at the receiver and the estimation is based on weighted correlation calculations. The first MLE has an estimation range of the entire signal bandwidth. The second MLE reduces the computational load with the loss of estimation range. Phase or its differentiation obtained from the autocorrelation of periodic sequences with different lags can be used to estimate CFO. The way of dealing with noise-noise product items in covariance affects the accuracy of CFO estimator. Based on these signal processing methods, four best linear unbiased estimators (BLUEs) with low complexity are proposed. The weighting coefficients of all BLUEs are closed-form and uncorrelated to signal-to-noise ratio (SNR). Four BLUEs have asymptotically consistent variances. All estimators can asymptotically approach the theoretical Cramer-Rao Lower bound (CRLB) at different SNRs.Then we deal with the multiple CFOs estimation issues for interleaved OFDMA uplink system. Interleaved carrier assignment scheme (CAS) based uplink system has obtained much attention because this CAS brings the maximized frequency diversity effect when communication link undergoes severe CFO problem. The signal structure for multiple-CFO estimation problem in interleaved system has a similar structure as that of Direction of Arrival (DOA) estimation for array signal processing. Therefore, multiple signal classification (MUSIC) algorithm based on subspace decomposition can be utilized to estimate multiple CFOs. Two types of CFO estimator are proposed in this paper, unitary MUSIC and unitary root-MUSIC estimators. Due to the use of real value covariance matrix, the proposed algorithms have lower computational complexity when compared with the conventional complex value covariance matrix based MUSIC algorithm. However, the unitary estimators have comparable performance, which is verified by the computer simulation.At last, based on software defined radio testbed, we finish the verification of key physical layer techniques in IEEE802.1In. In the verification, the proposed best linear unbiased estimator is utilized as the frequency synchronization algorithm. The result shows that the synchronization and other signal processing methods are efficient and the performance is excellent. |
PDF Full Text Request