基于证据理论的车辆组合导航系统的信息融合

由于车载GPS(G lobal Position ing System)接收机的定位精度通常受卫星状况和道路环境的影响,同时DR(Dead Reckon ing)系统的距离和航向传感器随着行驶距离的增加,会产生误差积累,这些误差使传感器的定位轨迹偏离车辆所在的实际位置。为此,在信息融合理论的基础上,利用高精度的电子地图对导航系统提供的车辆位置进行修正,以车辆位置到匹配道路的距离、车辆行驶方向与道路方向的夹角作为评判可信度的证据,提出了基于D-S(Dempster-Shafer)证据推理的信息融合方法。实际跑车实验表明,该地图匹配融合方法能够使车辆导航系统的定位精度达到5 m左右。对于改善低成本的GPS/DR车辆导航系统的性能是个行之有效的方案。

Information Fusion Methods Based on Evidential Theory for Vehicle Integrated Navigation System

Most car navigation systems estimate the car positions from the GPS(Global Positioning System) and the DR(Dead Reckoning).However,the unknown GPS noise characteristic and the unbounded DR accumulation of errors over time make the position information with undesirable position errors.The accurate digital road map is used to correct the position trajectory,the perpendicular distance between the vehicle position and matched road as well as the angle between the vehicle running heading and matched road direction are used to assess the probability which candidate roads are the matched roads,and an information fusion method based on(D-S(Dempster-Shafer)) evidence reasoning for GPS integrated navigation system is proposed,The field test demonstrates the positioning accuracy of the CNS(Car-integrated Navigation System)can reach about 5 m using the Map-Matching method.It is a useful scheme to improve the performance of the low cost CNS.