光纤陀螺的动态性能研究

发布时间：2018-06-17 00:14

本文选题：光纤陀螺 + 动态性能　；参考：《哈尔滨工程大学》2012年博士论文

【摘要】：光纤陀螺广泛应用于捷联惯性导航系统中,它的工作性能在很大程度上决定着整个惯性系统的精度。目前国内外对处于静止或匀速运动状态下的光纤陀螺性能即光纤陀螺的静态性能进行了大量的研究报道,而对处于角加速运动状态的光纤陀螺性能即光纤陀螺的动态性能则研究的较少。本文系统研究了光纤陀螺动态条件下的误差理论、测试方法、测试结果的分析方法和性能的改进方法。捷联系统直接与运动载体固连,在实际运动中经常处于角加速运动状态。因此,角加速运动下光纤陀螺性能的优劣决定着捷联惯导系统的导航精度。在分析光纤陀螺系统模型和角加速运动响应特性的基础上,建立了光纤陀螺动态条件下的误差模型。详细剖析了光纤陀螺的稳定性条件,结构参数、信号处理周期以及外界环境对光纤陀螺动态性能的影响。从理论上阐述了光源平均波长和反馈通道增益的变化作用于结构参数的机理,台阶高度信号和输出信号的滤波作用于信号处理周期的机理。仿真结果证明所建立的光纤陀螺动态条件下的误差模型准确可靠。为了对光纤陀螺的动态性能进行合理评价,采用三轴惯导测试转台来模拟光纤陀螺的角加速运动状态,给光纤陀螺施加瞬时变化的角速度输入,对其进行动态测试。对测试原理和测试步骤进行详细阐述。采用位置差商法、分周期离散理论曲线法和公式变换法确定动态测试中的基准值。分析结果表明分周期离散理论曲线法和公式变换法较适合。三轴转台的精度是制约光纤陀螺测试结果精确性的瓶颈因素。根据转台在摇摆状态下输出信号的特点,采用皮萨伦科谱分解法对其进行辨识。由此得到转台信号的幅度失真、频率失真以及夹杂在其中的分量谐波的幅度和频率。并通过比较辨识前后转台误差的标准差,证明了辨识结果的可靠性。同时,分析了不同转台框在相同摇摆条件下的失真规律和同一转台框在不同摇摆条件下的失真规律。实验结果表明,皮萨伦科谱分解法应用于转台正弦信号的失真辨识中是非常适用的。为了避免转台摇摆运动轨迹中夹杂的噪声传递到动态测试的基准信号中,结合小波多分辨分析和空间局部化的性质,选用db5小波对转台的运动轨迹进行降噪。结果表明,小波滤波法的噪声滤除效果理想,提高了对光纤陀螺动态测试结果评定的准确性。采用时延估计理论分析和补偿动态测试中基准信号与实际信号之间的时间延迟。利用测试中信号与噪声、噪声与噪声互不相关的特点,对两个信号进行相关运算,估计出两信号之间的时间延迟,并进一步对其补偿。结果表明,时间延迟对测试结果有较大影响,经过补偿后的动态误差更符合实际情况。为了更好的分析光纤陀螺的动态性能,采用动态Allan方差法对光纤陀螺的动态误差进行分析。根据动态Allan方差法中窗函数的原理,讨论了不同窗口长度对动态误差分析结果的影响。并且,提供了单一摇摆运动和两种复合摇摆运动的分析结果。从分析图中方差的起伏变化可以看出,动态Allan方差法可以准确地反映动态误差里的突变和周期性变化等非稳定性因素,能够清晰地辨识出隐藏在动态误差里的不同摇摆状态。由理论分析和实验结果可知,动态Allan方差法对光纤陀螺的动态性能分析是非常适用的。动态Allan方差法是一种分析非平稳性信号的有效工具,但在辨识噪声时存在功率泄漏和定量表示单一的缺陷。为此,提出窗函数组合法和噪声量值的二维表示法对其改进,并将其用于分析和定量描述光纤陀螺动态误差中的各种噪声项。窗函数组合法在光纤陀螺动态误差分解的基础上采用矩形窗和汉宁窗对其中的低频噪声和高频噪声分别进行分析。噪声量值的二维表示法根据动态Allan方差法原理得到噪声量值随采样点数目增加的变化规律。实验结果表明窗函数组合法可以满足不同频段噪声的辨识要求,减小功率泄漏；噪声量值的二维表示法可以准确地反映出动态误差中噪声项的变化特征。为了改善光纤陀螺的动态性能,将普通PID控制器的设计思想应用到光纤陀螺的动态误差控制器中,并进一步基于光纤陀螺动态条件下的误差特性设计出一种新型的动态误差控制器结构。前者主要通过普通PID控制器的设计思想与光纤陀螺的控制方式相结合,即基于光纤陀螺内部控制方式实现。而后者将微分环节设置在输出信号之前,使输出信号和反馈信号同时具有提前预测功能,并在微分环节后面加一个低通滤波器以抑制微分环节带来的高频干扰。这一新控制器结构既可以使控制量减小,各个时刻的控制误差不累积,还可以使输出跟踪输入,反映输入的变化。两种动态误差控制器均通过在光纤陀螺数字信号处理芯片FPGA的各模块中编写VHDL语言实现。实验结果表明两种动态误差控制器都可以明显地改善光纤陀螺的动态性能,且基于光纤陀螺动态条件下误差特性的动态误差控制器具有更优的控制效果。
[Abstract]:Fiber optic gyroscope is widely used in strapdown inertial navigation system. Its performance determines the precision of the whole inertial system to a large extent. At present, a lot of research reports on the static performance of fiber optic gyroscope in the state of static or uniform motion are carried out at home and abroad, and it is in the state of angular acceleration. The performance of fiber optic gyroscope is less studied. This paper systematically studies the error theory, test method, analysis method and performance improvement method of the fiber optic gyroscope under the dynamic condition.
The strapdown system is directly connected with the motion carrier and is often in angular acceleration in actual motion. Therefore, the performance of the fiber optic gyroscope under angular acceleration is determined by the navigation precision of the sins. Based on the analysis of the model of the optical fiber gyroscope and the response characteristics of the angular acceleration motion, the dynamic condition of the fiber optic gyroscope is established. The stability conditions, the structure parameters, the signal processing period and the influence of the external environment on the dynamic performance of the fiber optic gyroscope are analyzed in detail. The mechanism of the change of the average wavelength of the light source and the gain of the feedback channel on the structure parameters is expounded in theory, and the filtering of the step height signal and the output signal is acted on the effect of the filtering of the step height signal and the output signal. The simulation results show that the error model established under the dynamic condition of fog is accurate and reliable.
In order to reasonably evaluate the dynamic performance of fiber optic gyroscope, three axis inertial navigation test turntable is used to simulate the angular acceleration motion state of fiber optic gyroscope. The instantaneous variation of angular velocity input is applied to the fiber optic gyroscope, and the dynamic test is carried out. The test principle and test steps are detailed and detailed. The position difference commercial method is used and the periodic discrete principle is adopted. The curve method and the formula transformation method are used to determine the reference values in dynamic testing. The analysis results show that the discrete periodic curve method and the formula transformation method are more suitable.
The accuracy of the three axis turntable is the bottleneck factor that restricts the accuracy of the test results of the fiber optic gyroscope. According to the characteristics of the output signal of the turntable in the swing state, the spectrum decomposition method is used to identify it. Thus the amplitude distortion, frequency distortion and the amplitude and frequency of the component harmonics in the turntable signal are obtained. Compared with the standard deviation of the turntable error before and after identification, the reliability of the identification results is proved. At the same time, the distortion law of different turntable frames under the same swing condition and the distortion law of the same turntable under different swing conditions are analyzed. The experimental results show that the spectral decomposition method of pzararlco should be used in the distortion identification of the sinusoidal signal of the turntable. It is applicable.
In order to avoid the noise in the swing movement track of the turntable to the datum signal of dynamic testing, combined with the properties of wavelet multi-resolution analysis and spatial localization, db5 wavelet is used to denoise the motion trajectory of the turntable. The results show that the noise filtering of the wavelet filtering method is ideal and improves the dynamic test result of the fiber optic gyroscope. The accuracy of the assessment.
Using the time delay estimation theory to analyze and compensate the time delay between the reference signal and the actual signal in dynamic testing. Using the characteristics of the signal and noise, noise and noise in the test, the correlation operation between the two signals is performed, the time delay between the two signals is estimated, and the compensation is further compensated. The result shows that the time delay is the same. The test result has great influence, and the dynamic error after compensation is more consistent with the actual situation.
In order to better analyze the dynamic performance of fiber optic gyroscope, dynamic Allan variance method is used to analyze the dynamic error of fiber optic gyroscope. According to the principle of window function in the dynamic Allan variance method, the influence of the length of different window on the result of dynamic error analysis is discussed. And the single swing motion and two kinds of compound swaying motion are provided. It can be seen from the fluctuation and fluctuation of the analysis diagram that the dynamic Allan variance method can accurately reflect the unstable factors such as the sudden change and the periodic variation in the dynamic error, and can clearly identify the different swaying states hidden in the dynamic error. The dynamic Allan variance method is known to the light by the theoretical analysis and the experimental results. The dynamic performance analysis of the fiber gyroscope is very suitable.
The dynamic Allan variance method is an effective tool for analyzing non-stationary signals, but there is a single defect in the power leakage and quantitative representation of the noise. Therefore, a two dimensional representation method of window function combination and the value of noise is proposed, and it is used to analyze and quantify all kinds of noise items in the dynamic error of fiber optic gyroscope. The window function combination method uses the rectangular window and Hanning window to analyze the low frequency noise and the high frequency noise respectively on the basis of the dynamic error decomposition of the fiber optic gyroscope. The two-dimensional representation method of the noise value is based on the principle of the dynamic Allan variance method to obtain the variation of the noise value with the number of the sampling points. The experimental results show the window function group. The method can satisfy the identification requirements of different frequency bands and reduce the power leakage, and the two-dimensional representation of the value of the noise can accurately reflect the change characteristics of the noise term in the dynamic error.
In order to improve the dynamic performance of fiber optic gyroscope, the design idea of ordinary PID controller is applied to the dynamic error controller of fiber optic gyroscope, and a new type of dynamic error controller structure is designed based on the error characteristics of the dynamic condition of fiber optic gyroscope. The former is mainly based on the design idea of ordinary PID controller and optical fiber. The control mode of the gyroscope is combined, which is based on the internal control mode of the fiber optic gyroscope. The latter sets the differential link before the output signal, so that the output signal and the feedback signal have the pre prediction function simultaneously, and a low pass filter is added to the differential link to suppress the high frequency interference caused by the differential link. The structure can not only reduce the control amount, not accumulate the control error at all time, but also can make the output tracking input and reflect the change of input. The two kinds of dynamic error controllers are realized by VHDL language in each module of the fiber optic gyroscope digital signal processing chip FPGA. The experimental results show that the two kinds of dynamic error controllers are all clear. The dynamic performance of fog is obviously improved, and the dynamic error controller based on the error characteristics of fog has better control effect.
【学位授予单位】：哈尔滨工程大学
【学位级别】：博士
【学位授予年份】：2012
【分类号】：TN966;TP212

【参考文献】