【作者】 温永智；

【导师】 吴杰；

【作者基本信息】 国防科学技术大学 ， 航空宇航科学与技术， 2007， 硕士

【摘要】 系统测量误差模型验证是惯导系统研制过程中十分重要的一环。随着精确打击概念的提出和不断深化,如何提高对误差模型的验证试验能力成为惯导系统研制和试验部门关注的焦点。车载验证试验正是在这种需求下产生和发展起来的地面试验技术,是惯性测量系统误差模型验证、精度及可靠性验证过程中必须通过的重要试验环节。本论文针对33系数捷联式惯性测量系统误差模型开展车载验证试验方法研究。论文首先应用环境函数矩阵法建立了惯导系统工具误差系数分离模型,介绍了环境函数矩阵的计算方法。在长时间(约1小时)试验条件下,惯导工具误差系数分离模型的线性化误差随时间快速增加。为此,本论文提出了减小线性化误差影响的“消去法”和“补偿迭代法”。论文设计了GPS外测方案,建立了试验车的6自由度运动模型,完成了三维空间“0”字型、“1”字型、“8”字型等典型轨迹的车载试验计算机仿真。研究了捷联惯导系统相对于车体坐标系的调姿策略,以充分激励各项惯导工具误差并降低环境函数矩阵的复共线性。论文介绍了几种常用的参数估计方法和验证试验结果的评价指标,并就惯导系统初始对准误差对误差系数分离结果的影响进行了分析和处理。计算机仿真结果表明:在一次试验中,使用典型轨迹,试验车采用不断加速和减速的速度控制策略,惯导系统姿态角按一定规律变化,使用最小二乘辨识法可以有效地分离出21个参数;剩下的12个安装误差系数与惯导系统初始对准误差具有很强的相关性,此相关性只与惯导系统相对于当地北天东坐标系的初始姿态有关;当进行多次试验并知道惯导系统初始对准误差统计特性时,则可以比较准确地估计出这12个安装误差系数。更多还原

【Abstract】 INS-error model validation test plays an important role in the manufacture of INS. Along with the conception of Point-Target-Clear, the study of INS error model validation test technique becomes more and more popular, and vehicle test technique is brought up in this background. Vehicle test is the vital segment in the validation of error model, precision and reliability.This thesis does intensive rsearches on the vehicle validation test method of SINS instrument error model which has 33 coefficients.Firstly, the separation model of guidance instrumentation systematic error is built using environment function matrix method, and the calculation of environment function matrix is introduced. The linearization error of the separation model increases very rapidly, when the time of test becomes long (about 1 hour). Taking account of this characteristic, two methods are brought forward to decrease the linearization error, they are respectively named "Expurgation-method" and "Repair-iterative-method" in this thesis. Then the outside measure scheme is designed using GPS; the 6-Ds dynamic model of the vehicle motion is built and the vehicle tracks of "0" type, "1" type, and "8" type have been simulated. In order to inspirit all errors of the guidance instrumentation and reduce the multicollinearity of environment function matrix, the postures of SINS are analyzed, which should be kept against the reference frame of vehicle system in the test. In this thesis, several estimate methods and criterions for the appraisal of results are also introduced. At last, the influence of initial aiming error on test results are analyzed and dealt with.The results of simulation indicate that: 21 parameters can be separated effectively in one test by least square method, through the experiment vehicle speed control strategy and the rules which the posture angles of SINS should change according to. There are 12 installation error parameters can not be separated effectively, as they have strong pertinence with the initial error. And the simulation results show that the correlativity only has relations with the initial posture of SINS in NRE reference frame. The 12 installation error parameters can be estimated more accurately when the tests are done for many times and the initial error characteristic is known.更多还原