Study On Synchronization Algorithm Of OFDM System

Orthogonal frequency-division multiplexing is a technology of multi-carrier modulation. OFDM has been widely applied in high data transmission wireless systems,such as digital video broadcast (DVB), digital television (DTV), wireless local areanetwork (WLAN), due to its high frequency utilization efficiency and its robustness to multipath delay spread. OFDM will play a key role in the forth generation of mobile communication.OFDM systems are much sensitive to synchronization errors, especially frequency offset. The synchronization errors destroy the orthogonality between sub-carries of system,which lead to interfere between sub-carries and degrade the system performance. So it is important to have a good synchronization algorithm for OFDM system.The paper gives a brief survey of the development of OFDM technology and introduces the disadvantage and advantage of OFDM technology and the key technology of OFDM system. After presenting the characteristics of the wireless channel, this paper analyzes the impact of timing error and frequency offset on OFDM system in theory. Then the paper studies the synchronization algorithms,particularly the synchronization algorithms using training symbol and synchronization algorithms based on cyclic prefix. Thereinto, the relativity of cyclic prefix is used for synchronization estimation in ML method;training sequence with speciall structure is used for synchronization estimation in Schmidl&Cox,Minn and Park method. Moreover, this paper detailedly analyzes a joint synchronization estimation scheme which combines cyclic prefix(CP) with pilot carrier.Although traditional algorithm which combines cyclic prefix(CP) with pilot carrier is very effective to frequency offset estimation, its computational complexity is high. This paper proposes an improved frequency synchronization algorithm on the basis of traditional algorithm. Estimation veracity is sacrificed to get low computational complexity.By simplifying the maximum likelihood function, the proposed algorithm reduces the computational complexity of frequency offset estimation. Balance is achieved between complexity and veracity of the proposed algorithm. The simulation results show that the performance of this method is closer to traditional algorithm in high SNR.