The Research On Signal Detection And Parameter Estimation For OFDM System With Antenna Array

Orthogonal Frequency Division Multiplexing (OFDM) is the key technique of the 4G mobile communication. Array antenna is an effective way to solve the problem of lacking frequency resources and it also can improve system capacity and communication quality. In recent years, more and more research is focused on its applications on OFDM systems. It can achieve beam-forming and adaptive signal processing, improve signal to interference plus noise power ratio significantly, reduce the delay spread and multipath fading, attain angle diversity and spatial division multiplexing access, eliminate co-channel interference of system effectively, reduce the system error ratio significantly, and improve spectrum utilization and system capacity.This thesis is focused on the research of symbol detection and parameter estimation for OFDM system with antenna array. The main contents are as follows:Firstly, the thesis summarizes the developments, applications, basic principle and system model of OFDM and smart antenna technique. An analysis on wireless channel and establish data model. Secondly, the thesis analyses the impact of OFDM system performance caused by carrier frequency offset (CFO) and gives an introduction on the conventional CFO algorithms, especially the maximum likelihood based on cyclic prefix (CP-ML) algorithm and estimation of signal parameters via rotational invariance technique (ESPRIT) method. According to the analysis and reconstruction of the received OFDM signal with array antennas, we can find it has multi-invariant characteristic. So a blind CFO estimation algorithm based on MI-ESPRIT is investigated. The simulation results reveal that it has fine CFO performance. The lager the number of symbol blocks and antennas is, the better the CFO performance of MI-ESPEIT algorithm will be. The algorithm has a better performance than both CP-ML algorithm and ESPRIT method. And the MI-ESPRIT algorithm does not need any training symbol and cyclic prefix, so it has blind characteristics.Finally, the thesis summarizes the conventional beamforming algorithms of OFDM system with array antennas and gives an introduction on the basic mathematic knowledge of the parallel factor. An analysis of the received signal of OFDM system with array antennas shows that the received signal has trilinear model characteristics. So Trilinear decomposition-based joint symbol detection and DOA estimation (Trilinear-JSDE) algorithm is investigated. The simulation results reveal that the symbol detection performance of the proposed algorithm is very close to non-blind least squares method. Joint symbol detection and DOA estimation performance of our proposed algorithm improves when the number of snapshots increases. This algorithm even supports small sizes. The Trilinear-JSDE algorithm does not need the channel fading information, DOA and pilot or training information, so it has blind characteristics.