基于小波降噪的MEMS惯导系统对准方法研究∗

摘要

针对当前MEMS器件受加工工艺限制导致器件输出信噪比低，进而影响MEMS惯导系统初始化的问题，提出一种采用小波降噪的“重力+地磁”信息粗对准方法。分析MEMS器件输出信息结构特征，探索不同小波基函数在不同分解层数下的降噪效果并设计小波分解方案并实现惯性器件噪声的有效滤除；根据重力矢量垂直当地地理坐标系水平面的特征，利用加速度计敏感重力矢量水平分量实现载体初始水平姿态角的确定；结合地球表面任意位置处的地磁信息具有唯一指向性的特征，通过测量载体所在位置处的地磁信息并结合磁偏角修正实现载体初始方位角的求取。转台实验结果表明，经过小波降噪处理后得到的重力与地磁信息可实现MEMS惯导系统初始对准。%According to the low signal noise ratio of present MEMS devices caused by its processing technology, which influence the initial problem of MEMS inertial systems. The coarse alignment by gravity plus geomagnetic in-formation based on wavelet de-noise was proposed in this paper. The structure of MEMS output information was ana-lyzed. The de-noising with different wavelet function and decomposition levels were compared and the wavelet de-composition scheme was designed and used to filter out the inertial devices’ noise effectively. Considering the fact that gravity vector perpendicular to the level of local geographical coordinate system,the vehicle’ss initial horizontal attitudes could be confirmed by the gravity vector’ss level component sensed by accelerometers. Combined with the sole direction feature of geomagnetic information,the initial azimuth angle could be calculated by measuring geomag-netic information and the magnetic declaration angle. Turntable experiment results show that,the gravity and geo-magnetic information processed by wavelet de-noise could be used to realize the initial alignment of MEMS.