MPOS联邦实时组合算法

MPOS federated real-time integration method

微小型位置姿态测量系统(micro position and orientation system,MPOS)是为成像载荷提供实时高精度运动补偿信息的关键装置,其测量精度严重制约成像精度的提升。针对MPOS集中滤波实时性差的问题,在基于双ARM(advanced reduced instruction set computing machines)硬件平台的基础上,采用联邦滤波实时组合算法对多组传感器数据进行最优信息融合。以微惯性测量单元为公共参考系统,双天线全球导航卫星系统(global navigation satellite system,GNSS)和磁强计作为子系统提供量测信息,将各个子系统得到的局部误差协方差及状态估计值在主滤波器中进行信息融合得到最优估计值。通过动态实验验证了基于联邦滤波实时组合方法的MPOS,其位置、速度、航向角及水平姿态角精度可达0.6 m,0.03 m/s,0.15°,0.04°。

Micro position and orientation system(MPOS)is a key device to provide real-time and high-precision motion compensation information for imaging loads,and its measurement accuracy severely restricts the improvement of imaging accuracy.Aiming at MPOS real-time integrated filtering algorithm,based on dual-ARM(advanced RISC machines)hardware platform,this paper adopts federated filtering real-time integrated method to achieve optimal information fusion of multiple groups of sensor data.Micro inertial measurement unit is the common reference system,double antenna global navigation satellite system(GNSS)and magnetometer are the measuring information of subsystem.The local error covariance and state estimation obtained by each subsystem are fused in the main filter to obtain the optimal estimation value.The MPOS system based on real-time federated filtering integration method is verified by dynamic experiments.The accuracy of position,speed,heading and horizontal attitude can reach 0.6 m,0.03 m/s,0.15°and 0.04°.