车载GPS/DR组合导航系统自适应信息融合算法研究

Adaptive Data Fusion for Vehicle Integrated GPS/DR Navigation System

结合全球定位系统(GPS)和航位推算(DR)两种定位方式的优点,构建了基于卡尔曼滤波的自适应联邦滤波算法,实现陆地GPS/DR组合定位系统的数据融合;针对DR子系统的强非线性和扩展卡尔曼滤波算法带来的较大线性化损失,并结合机动加速度均值自适应算法,设计了一种基于U-D分解的自适应迭代卡尔曼滤波算法,更有效的减少DR子系统线性化带来的误差损失,提高定位精度;与同仿真环境下,DR子系统采用扩展卡尔曼滤波方法作了比较,结果表明该信息融合算法能更有效解决DR子系统的线性化误差问题,整个系统数据融合精度更高。

An adaptive federative filtering method based on Kalman Filter is set up for data fusion of Vehicle integrated GPS/DR navigation system by using the advantages of GPS and DR. The adaptive iterative Kalman filter based on the technique of U-D decomposition, which adapts to the mechanical acceleration mean adaptive algorithm, is designed to solve the high non-linearity of the subsystem of DR and decrease the loss of linearization yielded by Extend Kalman Filter. This new data fusion algorithm not only solves the linearization of DR subsystem more efficiently than EKF, but also makes the whole system fusion data more precise. At last the simulation resultes prove these conclusions.