【作者】 陈璐；

【导师】 刘兴钊； 高叶盛；

【作者基本信息】 上海交通大学 ， 电子与通信工程（专业学位）， 2020， 硕士

【摘要】 信号参数测量是电子侦察领域中的一个重要课题。随着信号频率的提高和信号带宽的增大,如何快速有效地对高频率,大带宽信号进行参数测量成为一个重要的问题。传统信号参数测量方法基于数字信号处理。但对于信号逐渐趋向高频率、高密度、大带宽的特性,数字信号处理的处理能力受到现有的高速采样的能力和数字处理器的处理能力的限制,无法完全满足未来的发展趋势。本文采用的参数测量是基于光电协同技术的处理方式,在光域对信号进行频域计算后利用数字处理方式进行信号的参数测量。光电协同技术结合了微波光子学高速运算和数字信号处理灵活控制的优势,能够对高频率、大带宽信号进行快速准确的参数测量。本文主要讨论的是在基于具有光学傅里叶变换功能的系统基础上进行的参数测量算法。与传统的信号处理方式不同的是,文中设计的算法中的傅里叶变换均是由光学系统处理得到。由于光学系统在计算上高速运行、实时处理的特点,本文针对光学系统处理后的结果开展信号参数测量技术的研究。论文的主要工作和创新点如下:(1)利用自相关法,短时傅里叶变换法,插值法等瞬时测频方法估计信号的瞬时频率,从而估计三类信号的频率参数。针对正弦信号,提出了一种新的插值算法,该算法在较低信噪比下即接近CRLB。(2)针对线性调频信号频谱的宽带特性,提出了两种基于MDCFT的线性调频信号参数估计算法,提高了算法测量精度。(3)结合光域二维傅里叶变换的特点,通过光域变换结果得到信号的时频变换结果。并借助PCA-SVM方法,通过时频图像进行信号类型识别和时间参数的测量。更多还原

【Abstract】 Signal parameter measurement is an important part of electronic reconnaissance.Along with the increase of the signal frequency and bandwidth in the radar and communication systems,measuring the parameters of high-frequency and large-bandwidth signals quickly and effectively has been regarded as major research concerns.Traditional signal parameter measurement approaches are generally based on digital signal processing,however,the reliability and efficiency would be sharply reduced when the high-frequency,high-density and large-bandwidth signals are applied due to the limitation of sampling and processing capability of existing digital processors,which cannot satisfy the future development.In the interest of solving these problems,the approach based on the optronic processing is proposed to accomplish the frequency domain calculation in optical domain combined with the digital processing to estimate the signal parameters.The novel approach takes advantages of high computational efficiency of microwave photonics and flexibility of digital processing,which enables fast and accurate parameter measurement in the case of high-frequency and large-bandwidth signals.Based on the discussion above,this dissertation mainly introduces the parameter measurement approach with optical Fourier transform.Compared with traditional signal processing approached,the Fourier transform designed in this dissertation is estimated from the optical system which make the usage of more complex and effective parameter measurement approaches possible.The main research contents of this dissertation are listed as follows.First,this dissertation takes advantage of the auto-correlation,short-time Fourier transform,interpolation and other instantaneous frequency measurement approaches to estimate the instantaneous frequency of the signals,thus estimating the frequency parameters of the three types of signals.Furthermore,a novel interpolation algorithm for sinusoidal signals is proposed which is closed to CRLB at a lower SNR.Second,in the cases of wideband characteristics of the chirp signal spectrum,two algorithms for estimating the chirp signal parameters based on MDCFT are proposed,which improves the measurement accuracy of the algorithm.Third,this dissertation analyzes the two-dimensional Fourier transform in the optical domain to estimate the signals time-frequency transform results through the optical domain transform results.With the help of PCA-SVM method and time-frequency images,the signal type and time parameter recognition are accomplished as well.更多还原