Study On Synchronization Techniques For Software Radio And OFDM Systems

With recent research and advancements in B3G/4G, OFDM, MIMO and software radio are increasingly important topics in wireless mobile communications. Synchronization is regarded as the primary object for communication signal processing, which influences on the following baseband signal processing. In this dissertation, the carrier frequency offset estimation for OFDM and synchronization of single carrier signal for software radio are studied. The main research production of this dissertation is as follows:1. A novel method for integer frequency offset estimation of OFDM systems is derived, which is based on maximum likelihood (ML) technique and exploits the differential information between two consecutive blocks of OFDM data symbols in the frequency domain. The reason why the proposed ML method has better performance than the conventional method is analyzed.2. Symbol timing and frequency offset estimation are key techniques in OFDM systems. Maximum likelihood algorithms based on correlation operation are widely used. Optimal correlation length plays an important role in correlation operation. This dissertation proves that the performance of maximum likelihood estimation method based on cyclic prefix can be improved by decreasing correlation length. The optimal correlation length under multipath propagation environment equals the length of cyclic prefix minus maximum multipath length. The partial correlation algorithm tracks maximum multipath length by subspace decomposition and calculates optimal correlation length subsequently for symbol timing and frequency offset estimation. Simulation results show that the algorithm can efficiently suppress affection of multipath on synchronization estimation.3. The MV analytical model of OFDM system and signal is presented in this dissertation. Each process of the OFDM signal from transmitter to receiver is defined by the strict mathematical formulation in the MV model. Some factors such as virtual carriers, carrier frequency offset, multipath channel and channel order more than cyclic prefix are synthetically considered in this model. Based on the expression of the MV model, the received signal model in the OFDM frequency domain helps to analyze the ICI and ISI of the OFDM system. The MV derived expression of the received signalmodel in the OFDM time domain distinctly illustrates the similarity between OFDM signal model and uniform linear array signal model, which enlightens us on the OFDM frequency offset estimation used by DOA estimation method.4. Analysis indicates that the model of OFDM frequency offset estimation based on virtual-carrier can be equivalent to the model of DOA estoimation of uniform linear array among array signal processing. This important conclusion provides many heuristic methods for carrier frequency offset estimation of OFDM systems. This is because many methods of DOA estimation can be applied to OFDM frequency offset estimation without any modification.5. Combined with the special structure of the OFDM equivalent steering matrix and based on the DML method for DOA estimation, four OFDM frequency offset estimation methods are proposed. These methods are provided with either the high computational efficiency or the satisfied estimation performance.6. A burst detection method for burst-mode GMSK signal is presented, and corresponding burst frame and strategy of detection window slip is designed. It consists of burst acquisition and acknowledgement. It is based on digital frequency discriminator and Fast Fourier Transform (FFT), and suitable for software radios. A combined efficient method of burst detection, symbol timing and frequency offset estimation for burst-mode GMSK based on discriminator and FFT is also proposed. The two-level decision method based on an integrate-and-dump filter suitable for discriminator detection is employed.7. An efficient combined method of symbol timing recovery, frequency offset estimation and correction is proposed for burst-mode DQPSK in software radios. Due to its low calculation requirement and easy realization in DSP, the method is readily implemented in a software radios receiver. The method has a feedforward and preambleless structure that is particularly suited in burst-mode transmissions.