高动态低载噪比信号的频率估计算法研究
发布时间：2018-10-26 19:27
【摘要】：为了从高动态环境中估计出载噪比极低的目标信号的频率,国内外开展了大量的算法研究,但这些算法的关注点大都集中在对恒频信号频率的估计,应用环境的载噪比一般也在30dBHz以上,只有极少数的文献研究了载噪比为17~18dBHz的情况。本文针对极低载噪比的高动态信号,且信号频率并不单一,提出了一种捕获和拟合相结合的频率估计算法,并且保证估计的精度在5mHz内。首先,本文给出了频率估计的数学模型并且推导了参数估计的理论界,给出了其闭合表达式,为以后的研究提供了理论基础。然后,本文介绍了恒频估计和变频估计的算法,针对恒频估计,主要介绍了基于FFT的频率估计算法,并且给出了不同算法的仿真性能对比;针对变频估计主要介绍了时域分析方法、基于极大似然估计和含预处理的FFT算法,仿真分析发现时域匹配周期图算法在实际中性能较优;最后,通过前面对恒频估计算法的分析,总结出了最优的恒频估计算法,并将其应用于对变频信号的估计,仿真验证了其在高载噪比时具有较好的性能;对于低载噪比高动态信号的频率估计,本文提出了新的高精度估计算法,由捕获模块和拟合模块两部分组成,捕获模块中包括两次捕获操作,主要是根据时域匹配周期图算法改进而来,通过设置递减的变化率步长,确定信号变化率的范围,最终可将其变化率降到0.1Hz/s以下,实现对信号的抓捕,同时给出了最优捕获的参数设置;拟合模块是对频率估计精度的进一步提升,将捕获得到的信号看作是恒频信号,然后采用傅里叶插值的频率估计算法进行估计和线性拟合,其中也包括两次拟合操作,最终该算法在载噪比为11dBHz以上,并且频率变化率范围为[-3kHz/s,3kHz/s]的信号中,估计精度达到5mHz的要求。
[Abstract]:In order to estimate the frequency of target signal with very low carrier to noise ratio from high dynamic environment, a large number of algorithms have been studied at home and abroad, but the focus of these algorithms is mostly on the estimation of frequency of constant frequency signal. The carrier noise ratio of the applied environment is generally above 30dBHz, and only a few literatures have studied the case where the carrier noise ratio is 17~18dBHz. In this paper, a frequency estimation algorithm combining acquisition and fitting is proposed for high dynamic signals with very low carrier / noise ratio and the signal frequency is not single, and the accuracy of the estimation is guaranteed in 5mHz. First of all, the mathematical model of frequency estimation is given, and the theoretical bound of parameter estimation is derived, and the closed expression is given, which provides a theoretical basis for further research. Then, this paper introduces the algorithms of constant frequency estimation and frequency conversion estimation. For constant frequency estimation, the frequency estimation algorithm based on FFT is mainly introduced, and the simulation performance of different algorithms is compared. Based on the maximum likelihood estimation (MLE) and the FFT algorithm with preprocessing, the simulation results show that the time-domain matching periodic graph algorithm has better performance in practice. Finally, through the analysis of the previous constant frequency estimation algorithm, the optimal constant frequency estimation algorithm is summarized, and it is applied to the frequency conversion signal estimation. The simulation results show that the algorithm has better performance when the load / noise ratio is high. For the frequency estimation of low load to noise ratio and high dynamic signal, a new high precision estimation algorithm is proposed in this paper, which consists of two parts: acquisition module and fitting module. The capture module includes two acquisition operations. It is mainly improved according to time-domain matching period diagram algorithm. By setting the decreasing rate of change step size, the range of signal change rate can be determined, and finally its change rate can be reduced below 0.1Hz/s to realize the capture of the signal. At the same time, the parameter setting of the optimal capture is given. The fitting module is a further improvement of the accuracy of frequency estimation. The captured signal is regarded as a constant frequency signal, and then the frequency estimation algorithm based on Fourier interpolation is used to estimate and linear fit, which also includes two fitting operations. Finally, the estimation accuracy of the proposed algorithm is up to the requirement of 5mHz when the carrier noise ratio is above 11dBHz and the frequency change rate is in the range of-3 kHz / s ~ 3 kHz / s.
【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】：2014
【分类号】：TN911.23
本文编号：2296721
[Abstract]:In order to estimate the frequency of target signal with very low carrier to noise ratio from high dynamic environment, a large number of algorithms have been studied at home and abroad, but the focus of these algorithms is mostly on the estimation of frequency of constant frequency signal. The carrier noise ratio of the applied environment is generally above 30dBHz, and only a few literatures have studied the case where the carrier noise ratio is 17~18dBHz. In this paper, a frequency estimation algorithm combining acquisition and fitting is proposed for high dynamic signals with very low carrier / noise ratio and the signal frequency is not single, and the accuracy of the estimation is guaranteed in 5mHz. First of all, the mathematical model of frequency estimation is given, and the theoretical bound of parameter estimation is derived, and the closed expression is given, which provides a theoretical basis for further research. Then, this paper introduces the algorithms of constant frequency estimation and frequency conversion estimation. For constant frequency estimation, the frequency estimation algorithm based on FFT is mainly introduced, and the simulation performance of different algorithms is compared. Based on the maximum likelihood estimation (MLE) and the FFT algorithm with preprocessing, the simulation results show that the time-domain matching periodic graph algorithm has better performance in practice. Finally, through the analysis of the previous constant frequency estimation algorithm, the optimal constant frequency estimation algorithm is summarized, and it is applied to the frequency conversion signal estimation. The simulation results show that the algorithm has better performance when the load / noise ratio is high. For the frequency estimation of low load to noise ratio and high dynamic signal, a new high precision estimation algorithm is proposed in this paper, which consists of two parts: acquisition module and fitting module. The capture module includes two acquisition operations. It is mainly improved according to time-domain matching period diagram algorithm. By setting the decreasing rate of change step size, the range of signal change rate can be determined, and finally its change rate can be reduced below 0.1Hz/s to realize the capture of the signal. At the same time, the parameter setting of the optimal capture is given. The fitting module is a further improvement of the accuracy of frequency estimation. The captured signal is regarded as a constant frequency signal, and then the frequency estimation algorithm based on Fourier interpolation is used to estimate and linear fit, which also includes two fitting operations. Finally, the estimation accuracy of the proposed algorithm is up to the requirement of 5mHz when the carrier noise ratio is above 11dBHz and the frequency change rate is in the range of-3 kHz / s ~ 3 kHz / s.
【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】：2014
【分类号】：TN911.23
【参考文献】
相关期刊论文 前1条
1 段瑞枫;刘荣科;周游;王闰昕;侯毅;;一种低复杂度的极低信噪比高动态信号载波粗捕获算法[J];航空学报;2013年03期
，本文编号：2296721
本文链接：https://www.wllwen.com/kejilunwen/wltx/2296721.html
