节点文献

用于车载自主导航激光多普勒测速仪的初步研究

Preliminary Study on Laser Doppler Velocimeter for Vehicle Self-contained Navigation

【作者】 周健

【导师】 龙兴武;

【作者基本信息】 国防科学技术大学 , 光学工程, 2010, 博士

【摘要】 随着现代工程技术的飞速发展,特别是航空、航天、机械制造等技术的发展,需要对物体的运动速度进行高精度测量。在导航系统中,已有的测速方式(加速度、全球定位系统等)存在误差累积、动态响应能力差等缺点,已经不能满足导航日益发展的要求。激光多普勒测速具有精度高、空间分辨率好、动态响应快、测量范围大、非接触测量等特点,已成为当今速度测试技术的重要发展方向。本文从实际应用的角度出发,对激光多普勒固体表面运动速度测量的原理和信号的检测、处理及分析进行了理论和实验研究,以推动该项技术在更广泛领域中的应用。论文指出了导航系统中已有测速方式的缺点,比较并分析了几种常见的光学非接触速度测试方法的特点,提出了将激光多普勒测速仪应用于车载导航系统中;回顾了激光多普勒测速技术的发展历程并介绍了国内外的研究现状,阐明了激光多普勒技术用于固体表面运动速度的测量是可行的。系统地分析了波源与探测器之间不同相对运动情况下的多普勒效应,阐述了激光多普勒测量车式载体运动速度的基本原理;通过比较直接光谱技术和光学外差技术的特点和适用条件,阐明了运用光学外差技术探测多普勒频率的巨大优势。对比分析了双光束差动模式、参考光模式以及自混合模式三种常见光路结构的优缺点,针对双光束差动系统不能进行离焦测量的问题,提出将多点分层技术用于双光束差动模式,并设计了多点分层差动激光多普勒测速系统;为了解决参考光模式中载体颠簸摇摆严重影响测量精度的难题,将毫米波、声波计程仪中的Janus配置技术用于参考光模式,并设计了基于Janus配置的参考光束型激光多普勒测速系统;采用声光调制器,实现方向辨别和低速测量。根据散斑场光强变化规律,阐明了由固态散射体运动导致多普勒信号产生的机理――两个散斑场的相干叠加;将散斑理论与随机过程相结合,推导出固态散射面激光多普勒信号强度的表示方式,并分析待测物体上的光斑直径与信号强度及信噪比的关系。根据条纹模型和散射机理建立了固体激光多普勒测速的数学模型,模拟固体表面运动多普勒测速原理,并通过数值模拟分析了固体表面特性对多普勒信号强度和测量精度的影响;通过对系统中光场分布的研究,给出了信号光与参考光在最佳匹配以及各种失配情况下外差效率的计算公式及数值模拟的结果,为光学系统的设计和光学参数的选择提供了理论依据。针对多普勒信号的特点,基于频率反馈技术设计了跟踪滤波器实时跟踪多普勒信号,消除系统工频干扰、高斯基底及其它干扰(如声光调制器)引入的噪声;为了更准确地从微弱信号中提取多普勒频率,运用数字相关技术抑制了跟踪滤波后信号中的噪声,进一步提高多普勒信号的信噪比。针对实时多普勒信号处理中直接进行FFT点数有限和精度不高的问题,提出了一种应用于激光多普勒测速的高精度信号处理技术,即首先利用FFT得到信号的频谱,搜索其谱峰值频率,接着利用Goertzel频谱细化算法对搜索的谱峰进行细化分析,再引入比值频谱校正算法对细化后的谱峰进行校正分析,从而将离散频谱分析算法、频谱细化算法和频谱校正算法三者有机结合起来,充分发挥各自的优点,仿真结果表明,该技术使频率估计精度优于0.004%;提出了频域内设定门限阈值与加速度门限阈值相结合的方法来判别多普勒信号的有效性,并采用最小二乘多项式拟合的方法对丢失的速度信息进行实时补偿。系统地分析了车载激光多普勒测速仪的测量误差,并针对各种情况提出了误差控制措施;在引入加速度分量的基础上,对高斯包络型多普勒信号参数估计方差的克莱姆-拉奥下限(Cramer-Rao Lower Bound, CRLB)进行了深入研究,明确了提高系统测量精度可能性的大小。搭建了激光多普勒测速实验系统,开展了固体表面运动速度测量的实验研究。建立双光束差动激光多普勒测速系统测量斩波片上任一点的切向运动速度,与光计数法测量结果相比较,双光束系统速度测量的平均相对精度为0.47%;建立参考光束型激光多普勒测速系统进行转台转速测量实验,和转台本身设定的速度相比较,参考光系统速度测量的相对精度优于0.35%;分别设计了多点分层差动LDV和基于Janus配置的参考光束型LDV测量小车相对于地面的真实速度,验证了两种激光多普勒测速系统用于车载自主导航系统的可行性。

【Abstract】 With the rapid development of modern engineering and technology, especially in the area of aviation, spaceflight and manufacture, it is necessary to measure the speed with high accuracy. In the navigation system, the existing speed measuring methods (accelerometer, global positioning system and so on) have the disadvantages of error accumulation and slow dynamic response. As a result, they cannot meet the requirement of the development of the navigation system any more. Laser Doppler technique has many advantages such as high accuracy, good spatial resolution, rapid dynamic response, large measuring range, and non-contact, so it represents the developing direction of speed measurement. This dissertation starts from the practical application, does some theoretical and experimental studies in terms of laser Doppler measurement, signal detecting, processing and analysis to broaden the application areas of this technology.The dissertation points out the defects of the existing speed measuring methods in the navigation system and compares and analyzes the characteristics of several ordinary optical and noncontact speed measurement methods. As a result, the idea of using laser Doppler velocimeter (LDV) to offer velocity for the vehicle navigation system is proposed. The feasibility of measuring the speed of certain solid surface using LDV is discussed after reviewing the development history and current status of the laser Doppler technology.The Doppler effects with different relative motion between a source and a detector are analyzed systematically and the measurement principle for vehicle’s velocity is expatiated. The superiority of optical heterodyne method for detecting the Doppler frequency is demonstrated based on the comparison of the features and the conditions between using direct light spectrum and optical heterodyne technique.The advantages and disadvantages of three optical model: dual-beam, reference beam and self-mixing, are discussed and analyzed. Multipoint layer technique is put forward for dual-beam model to solve the problem that it can not work in the status of out of focus and a multipoint layer-type LDV is designed. In order to reduce the effects of sway and toss of a vehicle on the measurement accuracy, Janus configuration used in millimeter wave and acoustic wave logs is proposed for reference beam model and reference beam LDV based on Janus configuration is designed. Especially, acousto-optic modulator is used for two kinds of LDVs to realize the direction discrimination and low speed measurement.The essence of laser Doppler signal, produced by the motion of solid surface that is superposed between two relevant speckle fields, is explained according to the order of the speckle field intensity variation. The expression of the solid-state scattering signal intensity of the laser Doppler is derived based on the speckle theory combined with the stochastic process. In addition, the relationship among light spot size on objects to be measured, the signal intensity and the signal-to-noise ratio is analyzed.A mathematic model of LDV, which is based on a theoretical model of fringes and the principle of light scattering, is developed for solid-state surface for analyzing the formation of the Doppler signal. The relationship among the characteristics of the solid-state surface, the signal intensity and the measurement accuracy is given by numerical simulations. By studying the distribution of the optical field, the expression and simulation results of heterodyne efficiency are given on the conditions of both exact match and mismatch. The results are of great importance to the design of the optical structure and the optimization of the optical parameters of the measurement system.A tracking filter is designed using frequency feedback technique to track the Doppler signal and to eliminate the noises of 50Hz, Gaussian basement, and other frequency interference (for example acousto-optic modulator) based on the characteristics of the Doppler signal. In order to extract Doppler frequency more exactly from the weak signal, the digital correlation technique is used to suppress the noise after filtering so that the signal-to-noise ratio is improved greatly.A high precision signal processing technology is introduced into LDV owing to the shortages of the real-time fast Fourier transform (FFT) algorithm such as limited data and low measurement precision. Firstly, using FFT algorithm to obtain the frequency spectrum of the Doppler signal and searching its peak value frequency; secondly, making zooming analysis using Goertzel algorithm; finally, using ratio correction algorithm to obtain a precise Doppler frequency. This technology combines three kinds of algorithms and makes good use of their advantages. The simulation results show that the precision of the frequency estimation is better than 0.004%. Moreover, a new method which combines frequency domain threshold and acceleration threshold is put forward to distinguish whether the Doppler signal is valid or not. The lost velocity information is compensated by the method of least square polynomial fitting.Measurement error of the LDV is analyzed systematically and methods to reduce the error in various situations are put forward. In order to clarify the potential of improving the measurement accuracy, the estimation of parameters and their Cramer-Rao Lower Bounds (CRLBs) of the Doppler signal with Gaussian envelope are investigated based on the introduction of the acceleration.Experimental systems of the LDV are built to do some research on speed measurement of solid-state surface. Firstly, a differential laser Doppler system is set up for measuring the tangential velocity of a certain point on the chopper. The mean relative accuracy of the laser Doppler system is 0.47% compared with the optics counting method. Secondly, a reference beam laser Doppler system is built to measure the rotating speed of a rotary table and the relative measurement accuracy is better than 0.35%. Finally, a multipoint layer-type LDV and a LDV based on Janus configuration are designed respectively for measuring the speed of a pushcart, which verified the feasibility of using LDV for the vehicle self-contained navigation systems.

节点文献中: 

本文链接的文献网络图示:

本文的引文网络