【作者】 张庆；

【导师】 谈振藩；

【作者基本信息】 哈尔滨工程大学 ， 导航、制导与控制， 2011， 博士

【摘要】 自主式航姿基准系统是水下无人运载器执行各种功能任务的保证。本文基于地磁导航和捷联惯性导航两种方法开展有关用于小型水下无人运载器的全自主组合航姿基准系统研究。采用自研的磁通门罗盘为组合系统提供地磁航向信息。设计输出灵敏度与次级线圈的匝数成反比的电流输出型的双铁芯磁通门传感器作为地磁测量传感器,与传统电压输出型磁通门传感器相比,电流输出型传感器有利于减小磁通门罗盘的体积和降低工艺难度。针对传统二次谐波检测法不能充分利用信号有用分量的不足,提出了复合偶次谐波检测法通过增加磁通门信号中被检测偶次谐波的数量提高传感器的灵敏度。为减小模拟电路的不稳定对磁通门罗盘航向解算精度造成的不利影响,设计了数字方法实现的信号处理和跟踪式航向解算单元。为改善磁通门罗盘的动态性能,构建了地磁航向信号数字随动系统。为了补偿由于水下无人运载器上铁磁性物质导致的磁通门罗盘自差,从地磁力、硬铁磁力和软铁磁力三方面对磁通门罗盘的自差成因进行了力学分析,推导了自差的大小和正负与水下无人运载器所在的磁纬度、航向以及是否倾斜的关系。提出了利用电磁方式代替传统的校正磁棒补偿磁通门罗盘的倾斜和硬磁自差的方法。提出了基于最小二乘的磁通门罗盘软磁自差补偿方法。实验结果表明,本文提出的整套磁通门罗盘的自差补偿方法正确,操作难度和复杂度低,可操作性强。由于光纤陀螺的随机误差是捷联惯导系统的主要误差源,本文采用时间序列分析法对光纤陀螺的随机误差进行建模研究。采用Allan方差法分离了光纤陀螺的各主要随机误差源,精确确定了各项误差系数的大小。对随机误差数据进行了趋势项、周期项、自相关、偏自相关的分析。研究了自回归和滑动平均(Autoregressive and Moving Average,简称ARMA)模型阶次搜索算法,将模型适用性检验法中的贝叶斯信息准则(BIC)用于模型阶次搜索,把二维搜索转化为一维搜索,得到了模型阶次的一致性估计。提出一种改进U-C算法用于长自回归模型计算残差法的模型参数估计过程,避免了递推估计法引入的累积误差,通过正序和逆序二次使用漂移序列的信息,提高了参数估计精度。根据随机误差建模结果,设计了卡尔曼滤波器对光纤陀螺数据进行滤波。滤波估计误差的统计特性和Allan方差分析结果表明,本文的建模方法能够指导滤波器设计有效抑制光纤陀螺的随机误差。为了提高系泊状态下捷联惯导系统的初始对准速度和精度,在粗对准过程中引入基座惯性坐标系以有效地分离载体摇摆运动干扰对系统对准精度的影响,通过组合磁通门罗盘航向信息提高了粗对准的快速性。在精对准参数辨识过程中引入可变遗忘因子,通过适时改变遗忘因子的大小和步长有效地跟踪参数的变化以提高对准的速度和精度。为提高组合系统的容错能力,利用递推和惯导系统辅助判断的方法对磁通门罗盘的航向信息进行滤波处理。为了提高航姿解算的效率,对组合航姿系统模型进行了合理近似。由于难以获得系统噪声的准确统计特性,采用H∞滤波技术提高系统的鲁棒性。实验结果表明,本文提出的组合航姿基准系统方案具有较好的工程参考价值。更多还原

【Abstract】 High-precision attitude heading reference system (AHRS) is a crucial instrument for Unmanned Underwater Vehicle (UUV) to implement complex tasks. Geomagnetic navigation and strapdown inertial navigation are chosen to compose a independent integrated AHRS for the UUV.A fluxgate compass is designed to provide reliable geomagnetic heading information for the UUV. The current ouput dual iron cores fluxgate sensor whose output sensitivity is inversely proportional to the number of the secondary coils is designed for measuring the geomagnetic field. Compared with the traditional voltage ouput fluxgate senor, the current ouput sensor could minish the volume of the fluxgate compass and reduce the difficulty of production. Aiming at the deficiency that the traditional second harmonic method couldn’t make full use of information of the fluxgate signal, the multiple even harmonics method is proposed to improve the sensitivity of the sensor by increasing the number of measured harmonics. In order to weaken the adverse effects produced by the analog circuits’instability and asymmetry which could decrease the heading accuracy of the fluxgate compass, a digital heading tracking algorithm is presented. A digital controller is introduced to improve the dynamic performance of the fluxgate compass.The impact of the geomagnetic force, hard and soft iron magnetic force on fluxgate compass are analyzed and the relationship between deviation compensation and UUV’s magnetic latitude, heading direction and heeling is given. A electromagnetic method to a compensate the heeling and hard magnetic deviation is proposed instead of traditional mesures. A method based on least-square algorithm is put forward to compensate the soft deviation. As the experiment results shown, the whole set of deviation compensation method is correct and easy to implement.Because the random errors of the fiber-optic gyro(FOG) are the main error sources in a strapdown navigation system, the model of random error is studied based on time series analysis method. Allan variance method is used to separate the main random error source and to determine the error coefficient accurately. The tendency item, periodic item, autocorrelation and partial autocorrelation fuctions of the data of FOG are analyzed. An order search algorithm for Autoregressive and Moving Average(ARMA) model is proposed, which employs the Bayesian Rules in model selection to search the model’s order. The consistent estimation of the model’s order is obtained by transforming a two-dimension searching mode into a one-dimension searching mode. In order to avoid the accumulated error from recursive estimation method, a improved U-C algorithm is introduced in the process of parameters’ estimation of the method that get the parameters of ARMA model by calculating the residual of the long auto-regression models.The accuracy of the estimation of the parameters is improved by using the data forward and backward.According to the ARMA model of the random errors of FOG, a Kalman filter is degined for eliminateing the random errors. According to the statistical properties of filtering estimation error and Allan variance, the model algorithm could help the filter eliminateing the random errors efficiently.In order to realize a rapid and accurate alignment for the strapdown inertial navigation system under mooring conditions, the body inertial coordinate system is introduced in the process of coarse alignment to separate the influence caused by swing motion of the board, and the heading of fluxgate compass is combined with the attitude of the inertial navigation system to improve the speed of coarse alignment, and the variable forgetting factor (VFF) is introduced in the process of fine alignment to improve the speed and precision of alignment through changing forgetting factor’s size and step length timely. In order to enhance the anti-interference ability of the geomagnetic heading subsystem, the heading information of the fluxgate compass is filtered through a recursive method aided by the informations of the inertial measurement unit (IMU). For the sake of the improving the efficiency of the computation, the order of the integrated attitude heading reference system model is reduced to a proper level. Because it is difficult to obtain the precise statistic characteristics of the system noise, H∞, filter is adopted to improve the robustness of the integrated system. As Experimental results shown, the integrated attitude heading reference system scheme presented in this thesis is a referencable application.更多还原