The Timing Synchronization Algorithm For OFDM Based On Compressed Sensing

Orthogonal Frequency Division Multiplex(OFDM) is the core technology of LTE. Since this technology is transfer data in parallel via a plurality of sub-carrier, which are orthogonal to each other, the orthogonality between subcarriers will be lost with the time deviation caused by the communication channel. The OFDM system should limit the time deviation extremely strict. Timing synchronization performance affects the stability of the communication system and the transmission efficiency, a good estimate of the timing synchronization communication system has an extremely important role. Timing synchronization is one of the most critical technical issues. It is the most common used method to utilize preamble to estimate the timing synchronization. While the timing synchronization algorithms applies preamble exist several non-negligible disadvantages:(1) the preamble, which is an auxiliary sequence and carry no useful information, will consume much spectrum resources.(2)The traditional timing synchronization algorithms applies preamble is not robust enough to against noise. The peak of timing metric is not stable due to the interference and noise existed in the channel.Compressed Sensing, which combine sampling with compression and reconstruct the signal at a lower sampling rate, is a revolutionary information processing theory that was first proposed by Donoho, Candes and Tao. In the research area of wireless communication, compressed sensing is widely used to improve the method of channel estimation for its low sampling rate and strong noise immunity.This paper take the lead to attempt to apply compressed sensing into traditional timing synchronization method for OFDM that uses preamble. First, for the timing sequence is not sparse, a sparse representation method based on the autocorrelation properties of timing sequence is proposed. Then, a timing metric based on the symmetry of synchronization sequence is constructed also. Utilizing the sparse representation method and timing metric, a timing synchronization algorithm that applies compressed sensing or CST for short is proposed. This paper analyses the shortcoming of the traditional algorithms and gives the system model and the concrete procedure of CST. The main procedure of CST includes the following steps: sparse representation of timing sequence to satisfy the premise of applying compressed sensing and compress the sequence by measurement matrix to shrink the length of timing sequence at the transmitting side, reconstruct the timing sequence by reconstruction algorithm and estimate timing point by the proposed timing metric with the reconstructed timing sequence at the receiving terminal. In this paper, a large number of simulation experiments that compare the new algorithm with the traditional algorithms for OFDM system applies preamble is preformed to verify the accuracy and reliability of the CST. The result of simulation experiments shows that the CST proposed in this paper can not only shrink the length of preamble effectively and save the spectrum resource, but also enhance the robustness of the timing synchronization estimation.