Research For Initial Alignment Technique Of Inertial Navigation Based On MEMS

MEMS Strapdown Inertial Navigation Systems (MEMS-SINS) is a system which combines micro-mechanical, microelectronics, semiconductor and other advanced technologies in one.It has a lot of advantages, such as wide application,low cost,simple structure,small volume,etc., and it has great potential for development. One of the key technologies of inertial navigation system is the initial alignment technique. In this paper, the study aims at the initialalignment technology based on MEMS.In order to improve the performance of initial navigation, the inertial systems technology plays a significant role, and directly affects the flowing algorithm accuracy. Firstly, the paper discusses the basic principles and system error propagation equation of inertial navigation systems and inertial MEMS devices used. Then the simulation experiments under real data is done. The features and characteristics of automatic alignment,attitude algorithm and error propagation are observed and recorded concluding that: the error of self-aligned heading, attitude algorithm and propagation equation is large and accumulates quickly. Because of the problem of MEMS gyroscope can't accurate sensitive the earth angular rate, this paper mainly studies the method of magnetometer aided initial alignment based on MEMS-IMU. Firstly in the coarse alignment, the method of double-vector attitude determination is used which combines the information of the MEMS accelerometers and magnetometers. Study three different auxiliary matrix construction methods, compare its alignment accuracy and the amount of calculation and choose the best one. In the fine alignment, Kalman filter and extended Kalman filter are studied on non-magnetometer aid condition. Due to low precision MEMS devices are relatively strong nonlinear, the alignment accuracy is not high. Firstly, the tratidional method of KF was given, then the method of heading error to aid alignment was studied, and a new method of traxial magnetic measurement error to aid was put out. The simulation result showed that the method of heading error performed better on the orientation accuracy, the method of magnetic measurement error performed better on convergence time.Finally, in order to the algorithm in actual system is feasible, software based on DSP /FPGA and PC is developed to complete a inertial system. In the FPGA ,the function of data transmission ,reception and caching , interrupt and timing management is completed.The DSP completes the coarse alignment, fine alignment, attitude algorithm. The PC completes the sending of commands, the receving of data, the displaying of attitudes and other functions. The software system achieves satisfactory results in practical application and proves that the algorithm is effective .