视觉辅助惯性定位定姿技术研究

Inertial Position and Attitude Determining Technology Aided by Visual System

在利用视觉系统辅助捷联惯性导航系统(SINS)进行定位定姿时,由于两者输出的位置表示方法不同,在信息融合时会存在匹配问题。以相对地理坐标系作为定位定姿系统的导航坐标系,重新对SINS进行了力学编排,分析了其误差传递特性,建立了相对地理坐标系下的状态方程模型,并利用Kalman滤波器实现了视觉辅助惯性定位定姿算法。仿真结果表明,方法避免了信息融合时位置匹配问题,同时降低了系统的计算量,满足了定位定姿系统的精度要求。

A matching problem in information fusion of the strapdown inertial navigation system（ SINS） aided with visual system should be solved,because of the completely different position expressions of the two individual system. This paper describes a special frame,the relative geographic frame,which is selected as the navigation reference frame. On this basis,the SINS mechanization equations and error equations which use the relative geographic frame as the navigation reference frame are derived. A state model in the relative geographic frame is established,and an algorithm of inertial position and attitude determining system aided by visual system is implemented by Kalman filter. The simulation results demonstrate that method avoids matching problem in information fusion,reduces the burden of calculation and satisfies the accuracy requirements of the position and attitude determining system.