Research On The Error Analysis And Compensation Technology Based On Micro Inertial Navigation System

The research purpose of this dissertation is to analyze the error mechanism and to build the error model based on the character of the MIMU system. The main content is to discuss and explore the real time compensation method for the random error of the MIMU system which being used in dynamic environment. The background project of the dissertation is to design and develop a practical MIMU/GNSS integrated navigation system.The precision of the silicon micromechanical inertial sensor are lower in general because of its manufacture technology. If the MIMU error is analyzed and compensated, the MIMU precision will be improved without changing hardware system. It is an important method to improve precision of navigation system based on the error analysis and compensation of the MIMU. It has significant meaning to analyze and build error model of MIMU, and to determine practical compensation method according to its used environment.At the different stage of the project research, we do a lot of experiments for BEI Company’s MMQ50 system and the MIMU system produced by Tsinghua University, so we get plenty of experiment data which is the fundamental of this dissertation.In chapter 1, the reference material and the new information of the MIMU application domestic and overseas are collected; In chapter 2, the influence caused by MIMU error to navigation precision is analyzed systematically, the error broadcasting equation of strapdown inertial system is deduced theoretically; In chapter 2, it also deduces the strapdown inertial system navigation algorithm and MIMU/GPS integrated navigation algorithm, the algorithm is realized and tested true by experiment; In chapter 4, the error character is analyzed and the error model is built based on the silicon micromechanical inertial sensor structure and principle; The measurement and calibration method is determined, and the calibration result is given.The chapter 5,6,7 are the kernel parts of the author’s study. First, it uses Time Series Analysis theory to pre-processing the static measurement data of MIMU, builds ARMA model, predicts and filters the random error based on the model recurrence formula and Kalman filter. The predicting and filtering effect is tested by experimental data. Then, it uses Wavelet Analysis theory to separate real motion signal and random error from the dynamic measurement data effectively. The separation effect is also tested by experimental data. This method is the fundamental to random error real time compensation in dynamic environment. In chapter 7, the experiment research and practical application about the MIMU error compensation technology are discussed. It uses experimental data and graph to supply a practical method for the MIMU’s engineering application. At last, it summarizes the research content and creative point systematically.