行人自主导航定位的IEKF-SWCS方法

IEKF-SWCS Method for Pedestrian Self-navigation and Location

针对利用可穿戴式IMU对行人进行导航定位过程中,惯性器件产生累积漂移误差影响导航定位精度的问题,提出了一种基于改进扩展卡尔曼滤波(Improved Extended Kalman Filter,IEKF)的行人自主导航定位方法。该方法建立了融合人体运动特征的18维滤波模型,在IEKF中设计分段闭环平滑(Step Wise Closed loop Smoothing,SWCS)算法,消除跳变的修正采样点,提高了轨迹平滑度。利用自研的IMU传感器进行试验验证,结果表明该方法能够有效抑制惯性器件的发散,进一步提高了行人自主导航定位精度,并且不增加任何额外的硬件成本,对行人导航的研究具有实际应用价值。

In the process of using wearable inertial measurement unit to realize pedestrian navigation, accumulated drift errors are increasing with pedestrian moving, which has serious effects on the navigation accuracy. To solve this problem, a pedestrian self-navigation and location method was proposed based on improved extended kalman filter(IEKF). An 18 dimensional filter model fused with human motion characteristics was built. Meanwhile, a step wise closed loop smoothing(SWCS) algorithm was designed in IEKF, which could eliminate the sharp correction at some sample points and improve the smoothness of the trajectory. A self-developed IMU sensor was used to make tests. The results demonstrate that the proposed method can significantly restrain the divergence of MEMS IMU, and effectively improve the location accuracy. In the process, no extra hardware cost has produced. So this method has practical application value for pedestrian navigation.