Research on Synchronization Technologies in Distributed MIMO OFDM Systems
|School||University of Electronic Science and Technology|
|Course||Communication and Information System|
|Keywords||OFDM MIMO distributed system time synchronization frequency synchronization parameter estimation|
Mobile communications always develop to provide higher data rate and better transmission quality. Orthogonal frequency division multiplexing (OFDM), which has been attracting much attention, can achieve higher frequency spectrum efficiency and overcome frequency-selective fading or narrow-band interference. Multiple input multiple output (MIMO), which has become one of the focuses in the area of wireless communications, can increase channel capacity and frequency spectrum efficiency remarkably. Hence MIMO-OFDM which is the combination of MIMO and OFDM has become the core technology of the next generation of mobile communications.Synchronization is the primary mission of communication systems, which influences the performance of the whole systems directly. OFDM is sensitive to synchronization errors. The performance of OFDM system will be deteriorated severely only by small carrier frequency offset. Correct time synchronization and precise frequency synchronization are the necessary precondition for OFDM systems. Furthermore, synchronization errors would bring severe inter-antenna interference to distributed MIMO systems, especially to distributed MIMO-OFDM systems. Hence synchronization has become the key technology of distributed MIMO-OFDM systems.This thesis focuses on the synchronization problems in OFDM, distributed MIMO and distributed MIMO-OFDM systems. The main contributions of this thesis can be described as follows:Firstly, the influences of timing error and carrier frequency offset on OFDM and MIMO-OFDM systems are mathematically modeled and analyzed. For the time synchronization method using cross-correlation, several different methods of setting detection threshold are compared. The method using the power of the received sequence not only adapts to the change of channel fading, but also works well with large change of signal-to-noise ratio.Secondly, based on the research on const amplitude zero auto-corelation (CAZAC) sequences, two synchronization methods for OFDM systems are proposed. One is the time synchronization method which works under narrowband interference. In this method, the interference detection and suppression procedure and the timing procedure are both processed in the frequency domain. Another method is the time and frequency synchronization method which works well with large frequency offset. By specially designed training sequences, this method can achieve sharp timing correlation peak without any additional operations in timing detection. And the cross-correlation also ensures the estimation performance for the large frequency offset.Thirdly, considering the multiple timing and frequency offsets in distributed MIMO systems, a maximum likelihood (ML) estimation model is set up. An iterative space-alternating generalized expectation-maximization (SAGE) based algorithm is proposed to resolve the problem of high computational complexity in ML estimation. Furthermore, the Cramér-Rao bound (CRB) for the multi-parameter estimation is also derived as a benchmark. The simulation results show the performance of the proposed method can reach the CRB.Finally, considering the multiple timing offsets in distributed MIMO-OFDM systems, two time tracking methods using pilots in the frequency domain are proposed. One is the method based on asynchronous coherence. This method is with simple structure and easy to perform. The other method is a ML estimation method with considering channel fading. This method can achieve better time tracking performance both with channel estimation.The research in this thesis extends the existing researches on OFDM synchronization, and explores the problem of synchronization in distributed MIMO systems and distributed MIMO-OFDM systems. The achievements have certain theoretical and practical values.