基于惯性/地磁的弹体组合测姿方法

摘要

针对单一传感器测姿精度有限和容错性不佳的问题,提出了基于惯性/地磁传感器的弹体姿态组合测量方法.该方法使用矢量因子联邦滤波算法对测量数据进行融合,并针对系统中存在的模型误差和噪声不严格符合高斯分布的问题,使用了改进的强跟踪无迹卡尔曼滤波算法作为联邦滤波的子滤波器.仿真结果表明,本文方法可以获得比任何单一传感器都高的估计精度,并比传统联邦滤波算法的融合精度更高.%Aiming at the problem of attitude detection accuracy is limited and the fault tolerance is low when u-sing single sensor ,a inertial sensor and magnetometer based attitude estimation approach was proposed .Meas-ure data was integrated via the vector formed information share federated filter .As there exist the problem of system model bias and the Non-Gaussian noise ,proposed a novel strong tracking unscented kalman filter algo-rithm as the local filter .Simulation result shows ,this method could get higher estimation accuracy than using any single sensor ,and which was higher than traditional integration approach .