新型GPS动态定位自适应卡尔曼滤波方法

为了获取GPS卫星的信号及定位数据的真实值,减小信号传播中因各种因素混入的随机误差对定位精度的影响,通过应用运动载体"当前"统计模型,取速度和位置为观测量建立GPS动态定位模型,将观测量维数增大1倍,从而增加了系统的可观测性和定位测算精度.此外,针对传统标准卡尔曼滤波法在动态滤波方面的不足进行了分析,提出了改进型Sage自适应卡尔曼滤波法.该方法在递推和滤波过程中不断地修正模型参数,始终保持噪声模型接近于真实模型,从而避免了标准卡尔曼滤波法中因建模不准确可能导致的滤波发散等问题,较好地解决了GPS动态定位中状态变量维数与滤波快速性之间的矛盾,以及状态噪声和观测噪声建模不准确和时变的问题.

A new self-adaptive Kalman filtering method for GPS kinematic positioning

In order to obtain the real value of positioning data of GPS satellite and to reduce the bad effect on position accuracy, Kalman filter is used to process raw data. First, by means of "current" statistic model and choosing location and speed as observation variable , the paper puts forward a mathematical model for GPS kinematic position. In this case, the dimension of measurement is twice of its predecessor, the observability and positioning accuracy are also increased. The deficiency of traditional and standard Kalman filtering method used in dynamic filter is analyzed, and an improved Sage adaptive Kalman filtering method is proposed. With this method, the conflict between the number of state variable and the speed of filtering calculation is successfully solved and the difficulty to model state noise and observation noise accurately is overcome as well. Performance of GPS kinematic positioning filter is improved for moving vehicle.