融合GPS/SINS的容积卡尔曼滤波智能车位置姿态估计方法

Position and attitude estimation method for intelligent vehicle using GPS/SINS and cubature Kalman filter

针对智能车在城市密集区域其全球定位系统(Global Positioning System,GPS)系统易受遮挡、干扰与多路径反射等因素影响,导致定位失灵和定位精度较低的问题,以及智能车位置姿态估计模型的非线性问题,提出了1种融合GPS/SINS的容积卡尔曼滤波智能车位置姿态估计方法。该方法将GPS和捷联惯导系统(Strap-down Inertial Navigation System,SINS)优势互补,构建以姿态误差、速度误差和位置误差等15维的系统状态方程,以GPS的位置/速度与SINS的位置/速度差值的6维系统观测方程,并采用容积卡尔曼滤波器对GPS和SINS的观测矢量进行有效关联与融合,估计并解算出车辆运动情况下的最优位置、速度、姿态参数。通过与GPS系统、SINS系统和基于扩展卡尔曼滤波的位姿估计方法仿真对比。结果表明,本文方法能给智能车提供精确可靠的车辆位姿参数。

Since the Global Positioning System（GPS）fails to location when the antenna is obstructed and lower positioning accuracy of the intelligent vehicle in crowded urban environment,a cubature Kalman filter-based vehicle position and attitude estimation using fusion of GPS/SINS is proposed.The GPS and Strap-down Inertial Navigation System（SINS）integrated navigation system using the observation of position,velocity and attitude are described,and the dynamic equations for the integrated system are established.The optimal estimation fusion structure of the cubature Kalman Filter is built.Experimental results proved that the proposed method can provide real-time,stable and reliable vehicle location parameters compared with the GPS subsystem,the SINS subsystem and the location method using the extended Kalman filter.