基于自相关系数插值的迭代符号定时同步
Iterative Symbol Timing Synchronization by Interpolation on Autocorrelation Coefficients
作者：游行远(哈尔滨工程大学 信息与通信工程学院, 黑龙江 哈尔滨 150001;武汉船舶通信研究所, 湖北 武汉 430079)；杨平(武汉船舶通信研究所, 湖北 武汉 430079)；彭军伟(哈尔滨工程大学 信息与通信工程学院, 黑龙江 哈尔滨 150001;武汉船舶通信研究所, 湖北 武汉 430079)；彭开志(武汉船舶通信研究所, 湖北 武汉 430079)
Author：YOU Xingyuan(College of Info. and Communication Eng., Harbin Eng. Univ., Harbin 150001, China;Wuhan Maritime Communication Research Inst., Wuhan 430079, China)；YANG Ping(Wuhan Maritime Communication Research Inst., Wuhan 430079, China)；PENG Junwei(College of Info. and Communication Eng., Harbin Eng. Univ., Harbin 150001, China;Wuhan Maritime Communication Research Inst., Wuhan 430079, China)；PENG Kaizhi(Wuhan Maritime Communication Research Inst., Wuhan 430079, China)
收稿日期：2016-03-11 年卷（期）页码：2017,49(3):117-122
期刊名称：工程科学与技术
Journal Name：Advanced Engineering Sciences
关键字：符号定时估计;匹配滤波;自相关系数;迭代;高斯插值
Key words：synchronization;matched filters;autocorrelation;iterative methods;Gauss interpolation
基金项目：船舶工业国防科技预研基金资助项目（11J3.4.2）
中文摘要
在无线数字通信系统中，全数字化接收机以固定的时钟进行采样，容易出现符号采样偏差，需要在接收端进行估计并补偿。针对符号定时估计的问题，提出了一种基于自相关系数插值的迭代符号定时同步算法。该算法分为两个步骤：粗同步与精同步。粗同步先利用零相滤波器或三相滤波器组对接收信号进行匹配滤波，再利用滑动自相关的方法对同步头序列进行相关运算，当超过同步门限时，开始进行同步跟踪对自相关系数进行峰值搜索，以确定精同步搜索区间。精同步直接在搜索区间内进行迭代搜索，迭代过程中需要先根据估计的时延值生成群时延的匹配滤波器以对接收信号进行补偿，再利用最小二乘逼近函数度量，同时，采用高斯插值的方法提高搜索速度以获取精确的估计结果。与传统算法相比，该算法降低了同步搜索过程中的计算复杂度，同时利用迭代搜索易于实现高精度的符号定时同步。仿真结果表明：对于长度为64的同步头序列，在高斯白噪声信道条件下，信噪比-2 dB时，该算法利用两次迭代即可逼近克拉美罗界；在信噪比为10 dB，针对±1/2符号周期内的采样偏差，本文算法的MSE性能均优于升余弦插值、MMSE插值与多相滤波器组算法；在瑞利信道、CCIR中等信道以及CCIR坏信道下也能进行有效的估计，适合无线数字通信系统的信道环境。
英文摘要
In a wireless digital communication system,a digital receiver was sampled with a fixed system clock.The asynchronous digital inputs may cause symbol time offset.To tackle the issue of symbol timing synchronization,this paper proposed an iterative symbol timing synchronization method by interpolation on autocorrelation coefficients.The proposed method was divided into two steps:coarse timing synchronization and fine timing synchronization.In coarse timing synchronization,the received samples were filtered by a zero-phase filter or a three-phase filter-banks.Then,the autocorrelation coefficients was calculated by using the sliding window autocorrelation of the preamble.When the synchronization threshold was reached,the maximum bin search on autocorrelation coefficients was used to obtain the searching interval for fine timing synchronization.In fine timing synchronization,an iterative method was used in the searching interval.It is necessary to generate the group delay matching filter according to the estimated time offset to compensate the received signal,and the likelihood ratio was calculated by using the least squares approximation function.Finally,the Gaussian interpolation was used to obtain an accurate result.Compared with the traditional methods,the proposed method had a lower computational load in coarse synchronization step.Meanwhile,the iterative searching method was utilized to conveniently provide the precise symbol timing synchronization.Simulation results showed that the proposed method can achieve Cramer-Rao bound at -2 dB with a preamble of 64 symbols in additive white Gaussian noise channel (AWGN);when the SNR was 10 dB and symbol time offset belong to [-1/2~1/2] symbol timing interval,the MSE performance exceed the cosine interpolation method,MMSE interpolation method and polyphase filter-banks method.Also,it worked well in Rayleigh,CCIR moderate channel as well as CCIR bad channel.In conclusion,the proposed method was suitable for wireless digital communication system.The proposed method was suitable for wireless digital communication system.
【关闭】
