Research Of Multi-Carrier Frequency Offsets Estimation Based On Distributed MIMO-OFDM

With the development of science and technology, the people's requirement for mobile communication services becomes larger and larger. The problems which how to transmit data with a high-speed in a more and more complex electromagnetic environment and reduce frequency resources have become the most important in wireless communication systems. It is proposed distributed MIMO-OFDM technology to solve these problems. It has the advantages of MIMO-OFDM technology:increase system channel capacity and data transfer rates significantly, against frequency selectivity fading and multi path fading. At the same times, it can provide a wider range of communications than MIMO-OFDM with less transmission power.Many technologies of MIMO and OFDM have been researched; even some have been used in the third generation wireless communication systems. However, there are still lots of problems in the distributed MIMO-OFDM technology need to be worked on and provide better solutions. Multi-carrier frequency offsets estimation for distributed MIMO-OFDM system is one of the most difficult problem need to be resolved. In this paper, we study deeply the algorithms for multi-carrier frequency offset estimation, the contents as follows:Firstly, in order to solve estimation of multi-CFOs in distributed MIMO-OFDM system, an algorithm which based on eigenvalue decomposition is proposed in the paper. It is the first time that eigenvalue decomposition has used to estimate CFO. This algorithm has a simple structure of pilot symbols, high precision of CFO estimation and fast speed of synchronous. Feasibility of the algorithm is verified by simulation results.Secondly, two measures for improvement the algorithm are proposed in this paper: spectral peak searching by large/small step and arithmetic mean for spectral function. Spectral peak searching by large/small step can reduce computation of spectral peak searching greatly. Mean of spectral function against the "false peak" which can lead mistakes in multi-CFOs estimation. Using simulated data illustrate the performance of algorithm.Finally, ML algorithm and eigenvalue decomposition algorithm will be used together. Firstly eigenvalue decomposition estimate multi-CFOs, then ML algorithm match the multi-CFOs to find out the correspondence between multi-CFOs and transmitting antenna. Improved method of ML algorithm in used for matching multi-CFOs, in order to make the algorithm be viable at low signal noise ratio(SNR). The simulation results show the performance of new multi-CFOs estimation algorithm.