Distributed Mimo System Frequency Synchronization Technology

Multiple-Input Multiple-Output (MIMO) will be the key technology for the next wireless communications system. Recently, distributed antennas for wireless communication is a hot topic. Distributed MIMO system attracts increase interest because this system is capable of achieving a higher diversity gain, which is a promising solution for future MIMO system. No matter in which system, the received signal is corrupted by frequency offset due to the Doppler shift and local oscillators drift. A lot of research work has been devoted to solving the problem of estimation and compensation of frequency offset, in conventional SISO and MIMO systems. However, in distributed MIMO systems, due to multiple frequency offsets, conventional estimation and compensation techniques are not effective. According to the best of our knowledge, few previous researches were done to solve the problem of multiple frequency offsets estimation and compenstiaon. In this dissertation, we pay attention to the problem of frequency synchronization including frequency offsets estimation and correction in distributed MIMO system.In this paper, we first introduce distributed MIMO system which is attracting much attention in the field of MIMO wireless communications and state detailedly the fundamental problems: multiple frequency offsets estimation and correction which is the research interest of this paper.Chapter 2 covers various categories of existing techniques on frequency offset estiamation. The frequency offset estimation algorithms are described and analyzed in single input single output (SISO) orthogonal frequency division multiplx (OFDM) system, collected MIMO system and distributed MIMO system separately. Problems are proposed when existing frequency synchronization algorithms are applied to distributed MIMO system: multiple antenna interference is ignored which make the mean square error (MSE) far beyond the CRLB when frequency offsets are estimated; prior information of frequency offsets are ignored when frequency offsets are estimated; due to multiple frequency offsets, conventional compensation techniques are not effective. Even if we got the knowledge of frequency offsets, it is difficult to compensate them in the receiver side with traditional method.Above questions are analysised in Chapter 3, 4 and 5. Novel algorithms are proposed to solve these problems.In Chapter 3, default of exsiting multiple frequency offsets estimation algorithm problems are found out: for the second-order estimator based on correlation, an error floor in the mean square error (MSE) performance due to the existence of multiple antenna interference in a multi-antenna system and the estimation range is reciprocal ratio to the length of the correlator pilot. For the expectation conditional maximization (ECM) which can overcome the error floor and performance of MSE achieves the CRB, the estimation range is limited. Chaper 3 models the ML multi-dimensional optimization problems as a multiple CFOs evolutionary game optimization problem, where CFOs are considered as different player in the evolutionary game theory. Simulations show that the proposed algorithm improves the MSE performance and extends the frequency offset estimation range to the upper bound of ML estimator.Frequency offsets to be estimated in existing algorithms are all based on the point that the frequency offsets are fixed and unchanged ignoring the oscillor is affected by environment factors like the temperature. Chapter 4 proposes frequency offsets estimation based on maximum a priority (MAP). Priory distribution is explored based on maximum entropy. Cramer-Rao bounds for multiple frequency offsets estimation using MAP are also derived. These simulations verify that this proposed algorithm outperforms existing ML estimation for the use of prior distribution of frequency offset..Chapter 5 considers the problem of multiple frequency offsets pre-correction in distributed multi-input multi-output (MIMO) systems. The received signal at each receiver is the summation of all the transmitted signals. Multiple frequency offsets make frequency compensation difficult if not impossible. In this chapter, taking advantage of equivalent channel matrix with frequency offsets information, an enhanced limited feedback precoding algorithm is proposed to correct the effect of frequency offsets. Simulation results show that the proposed scheme works as well as the system without frequency offsets.In a word, distributed MIMO system improves the capacity of converntional MIMO system. It will definitely influence our daily life. Frequency offset estimation and correction are a vital part for distributed MIMO to become reality, so deep investigation into this area is presented in this dissertation: multiple frequency offsets estimation algorithms based on evolutionary game theory and MAP estimation improve the estimation performance well and multiple frequency offsets correction algorithm based on limited feedback precoding can cancel the effect of frequency offsets well. As will be seen, these have great potential for application in the future.