基于卡尔曼滤波的视觉辅助惯导定位算法研究

Research on Visual Assisted Inertial Navigation Positioning Algorithm Based on Kalman Filter

针对惯性器件长期工作出现累计误差的问题,该文提出了一种基于卡尔曼滤波(Kalman Filtering)算法的视觉辅助惯导定位算法,该算法利用视觉定位和惯导定位的互补性,以松耦合的方式将视觉数据和惯导数据在卡尔曼滤波器中融合,有效地降低了惯导的累计误差。经实验验证:该算法可以有效地抑制惯导的累计误差,为惯导能够在长时间工作中进行准确定位提供了保障。

For the problem of cumulative errors in long-term operation of inertial devices,a KF-based visionassisted inertial navigation positioning algorithm is proposed.This algorithm uses the complementarity of visual positioning and inertial navigation positioning to combine visual data and inertial navigation.The derivative data is fused in a Kalman filter,which effectively reduces the cumulative error of the inertial navigation.It is verified by experiments that the algorithm can effectively suppress the cumulative error of the inertial navigation system,which provides a guarantee for the inertial navigation system to perform accurate positioning during long-term work.