基于卡尔曼滤波的交互式多模型GPS定位方法研究

Interacting Multiple Model Based on Kalman Filter Applied in GPS Point Positioning

针对GPS定位中,由于模型单一而不能适应环境影响或机动过程变化的问题,提出将交互式多模型算法引入到定位方法中。文中详细阐述了基于卡尔曼(Kalman)滤波的交互式多模型(IMMKF)算法原理及其在GPS定位中的应用。根据静态单点定位实测数据的试验分析,验证了变噪声模型的IMMKF能很好地适应环境噪声的变化;进一步通过动态定位仿真,说明了IMMKF算法在动态过程中能对多模型进行有效的融合,在很大程度上弥补了单模型的不足。

The problem of GPS positioning using single model is that,single model can＇t adapt to the changes of operating conditions or maneuvering process.In order to solve this problem,Interacting Multiple Model（IMM） is employed here.The algorithm of IMM based on Kalman filter is explained in detail and the application in GPS positioning is introduced.With the data of GPS static point positioning,it is proved that IMMKF of varying noise models can adapt to the changes of operating condition well.And the simulation of dynamic positioning shows that,multiple models can be fused effectively with IMMKF,and the disadvantage of single model can be overcome in a great degree.