| With the rapid development of the modern digital system,more broad bandwidth is needed matching people's requirement, the simple voice communication and network video communication have not meet the higher volume of data on the full-service communications desire, however, there is contradiction between the limited spectrum resource and the demand for bandwidth. OFDM has the advantages of high spectrum efficiency, anti-multi-path fading, simple channel equalization, easy hardware realization ,making it become one of the most important technology of the fourth-generation mobile communications , and it has been used widely. Synchronization is the most important part in the receiver, and directly impact on the performance of the entire system. OFDM system is more sensitively in synchronization comparing with that of single-carrier system, symbol synchronization error, may not only introduce inter-symbol interference (ISI), but also make a sharp decline in the performance of the system. Frequency error of the system brought about by inter-carrier interferrnce (ICI), quantitatively, the normalized frequency offset of 0.2, the carrier to interference has more than loss of 12dB.So frequency offset estimation and compensation become particularly important.The methods of OFDM synchronization can be divided into two classes in accordance with whether or not use the auxiliary data: data-aided method and non-data-aided method. General system model can also be divided into two classes in accordance with the continuity of sending data: continuous model and burst model,this paper mainly research on two classes method applied into corresponding model:1. The method of non-data-aided is more suitable for continuous model, this article used the equivalent low-pass model of OFDM to instead of the continuous model, a novel time synchronization algorithm suited multi-path fading channel is proposed. First, the received signals are multiplied by these conjugate to acquire the signal energy, then the correlation function is acquired by carrying out the energy differential operation with the interval of the length of OFDM data symbol part minus the length of cyclic prefix, finally timing synchronization is estimated by uniting the multiple symbols sampling. Because of taking advantage on differential operation of the signal energy, the synchronous peak is clearer, and the estimation of multiple symbols strengths the stability of synchronization. The simulation shows that the proposed algorithm can acquire a better performance of symbol timing, and remove the impact of frequency offset.2. In burst model, frame synchronization detection is get more focus, and the IEEE 802.11a simulation platform is selected to instead of burst model , An improved frame synchronization scheme based on CAZAC (Constant Amplitude Zero Auto Correlation) sequences is presented. Firstly, the CAZAC sequence is used as a short training sequence to change the standard preamble structure of IEEE 802.11a, and the number of the short training sequence for coarse frequency estimation is increased. Secondly, the proposed algorithm employs the cross-correlation of the short training sequence with a threshold to implement the frame synchronization. Because of the good correlation properties of CAZAC sequences and the synchronization information of its cross-correlation, higher accurate frame synchronization can be achieved, and the coarse frequency estimation is considerably improved. Furthermore, the stable performance of frame synchronization detection is also acquired even at the large frequency offset. Simulation results show that the proposed algorithm can get 2dB gain compared with that of traditional algorithm when multi-path delay is 100ns and frequency offset is 200 kHz. |
PDF Full Text Request