OFDM系统中载波同步算法的研究

发布时间：2018-05-22 20:09

本文选题：正交频分复用 + 符号间干扰　；参考：《西安电子科技大学》2015年硕士论文

【摘要】：正交频分复用(OFDM)在多载波传输技术里算是比较特殊的一种,并且它是第四代移动通信系统(4G)的核心技术之一。近年来OFDM技术不断受到大家的关注,主要是因为它存在很多优点,其中最主要的一个优点就是能够很好地对抗窄带干扰以及频率选择性衰落。在OFDM系统中,因为各子载波上的功率谱之间互相重叠,所以对子载波之间的正交性要求比较严格。载波频率偏差(CFO)会破坏子载波之间的正交性,引入子载波间干扰(ICI)和符号间干扰(ISI),从而导致OFDM的优势丧失,因此必须采取有效的措施对CFO进行补偿。本文主要研究了OFDM系统中的同步技术,尤其是系统中的载波同步算法。这些算法可以分为不借助辅助数据类和借助辅助数据类两大类,不借助辅助数据类主要是借助于OFDM系统本身的结构特点来完成载波同步。这类算法因为不需要借助于辅助信号,因此频谱利用率比较高。而数据辅助型算法大都要在数据信息中加入循环序列,这类算法是通过牺牲传输量来获得同步的,虽然会降低系统的数据传输效率,但捕获速度相对较快,估计精度也比较高并且适合分组数据通信。本文针对这两类算法都给出了几种具体的算法,包括ML算法、LS算法、Moose算法和SC算法。本文首先对这些算法的基本思想进行了研究,然后通过软件对算法进行了仿真实现,最后对仿真结果和算法性能进行了深入分析。本文在后半部分对给出的载波同步算法进行了深入研究,并给出了相应的改进算法。首先对ML算法的缺点进行了分析,给出了基于多符号并且只利用部分循环前缀(CP)的改进算法。其次,针对LS算法,给出了与自抵消相结合的最小二乘载波同步算法。新算法引入了自抵消技术,通过自抵消技术减弱了子载波干扰,使得接收信号更加纯净,从而使得新算法具有更好的性能。然后,通过对Moose算法的研究给出了一种余弦函数算法,该算法跟Moose算法类似都是在发送端的有效数据信息之前发送两个完全相同的OFDM符号作为训练序列,但该算法通过两个符号之间的相关矩阵的性质来进行载波同步。通过仿真分析,可以看出该算法具有比较小的误差。最后,针对SC算法给出了一种改进的算法。新算法在进行符号定时同步和整数倍频偏估计时,从一个新的角度出发,使得估计结果都更加的精准,并且在进行符号定时同步时不会出现地板效应。而在进行小数倍频偏估计时,由于引进了迭代算法,因此具有更好地性能。
[Abstract]:Orthogonal Frequency Division Multiplexing (OFDM) is a special kind of multicarrier transmission technology, and it is one of the core technologies of the fourth generation mobile communication system. In recent years, OFDM technology has attracted more and more attention, mainly because it has many advantages, one of which is its ability to resist narrowband interference and frequency selective fading. In OFDM systems, the orthogonality of subcarriers is strict because the power spectrum on each subcarrier overlaps with each other. The carrier frequency deviation (CFO) will destroy the orthogonality between subcarriers, and introduce inter-carrier interference (ICI) and inter-symbol interference (ISI), which will lead to the loss of the advantages of OFDM. Therefore, effective measures must be taken to compensate CFO. This paper mainly studies the synchronization technology in OFDM system, especially the carrier synchronization algorithm in the system. These algorithms can be divided into two categories: the auxiliary data class and the auxiliary data class. The carrier synchronization is accomplished by the structure of the OFDM system without the aid of the auxiliary data class. This kind of algorithm has high spectral efficiency because it does not need auxiliary signal. However, most of the data-aided algorithms need to add cyclic sequences to the data information. This kind of algorithm is synchronized by sacrificing the amount of transmission, although it will reduce the efficiency of data transmission, but the capture speed is relatively fast. The estimation accuracy is also high and suitable for packet data communication. In this paper, we give several algorithms for both kinds of algorithms, including ML algorithm, LS algorithm, moose algorithm and SC algorithm. In this paper, the basic ideas of these algorithms are studied, and then the algorithm is simulated by software. Finally, the simulation results and the performance of the algorithm are analyzed in depth. In the second half of this paper, the proposed carrier synchronization algorithm is deeply studied, and the corresponding improved algorithm is given. Firstly, the shortcomings of ML algorithm are analyzed, and an improved algorithm based on multi-symbol and using only partial cyclic prefix (CPP) is presented. Secondly, the least square carrier synchronization algorithm combined with self-cancellation is proposed for LS algorithm. The new algorithm introduces the self-cancellation technology, which weakens the subcarrier interference and makes the received signal purer, which makes the new algorithm have better performance. Then, a cosine function algorithm is proposed by studying the Moose algorithm. The algorithm is similar to the Moose algorithm, which sends two identical OFDM symbols as the training sequence before the valid data information of the sender. But the algorithm performs carrier synchronization by the property of the correlation matrix between two symbols. Through simulation analysis, we can see that the algorithm has a relatively small error. Finally, an improved SC algorithm is presented. When the symbol timing synchronization and integer frequency doubling offset are estimated, the new algorithm makes the estimation results more accurate from a new point of view, and there is no floor effect in symbol timing synchronization. In the estimation of decimal-octave frequency offset, the iterative algorithm is introduced, so it has better performance.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN929.53

【相似文献】