测绘通报 ›› 2017, Vol. 0 ›› Issue (7): 5-9.doi: 10.13474/j.cnki.11-2246.2017.0213

• 学术研究 • 上一篇    下一篇

一种基于改进UKF滤波的GPS+PDR组合定位方法

牛欢, 廉保旺   

  1. 西北工业大学电子信息学院, 陕西 西安 710072
  • 收稿日期:2017-03-13 出版日期:2017-07-25 发布日期:2017-08-07
  • 作者简介:牛欢(1993-),男,硕士生,主要研究方向为多传感器组合导航。E-mail:nih.a@163.com
  • 基金资助:
    国家自然科学基金(61301094);西北工业大学研究生创意创新种子基金(Z2017027)

An Integrated Positioning Method for GPS+PDR Based on Improved UKF Filtering

NIU Huan, LIAN Baowang   

  1. School of Electronics and Information, Northwestern Polytechnical University, Xi'an 710072, China
  • Received:2017-03-13 Online:2017-07-25 Published:2017-08-07

摘要: 针对接收信号质量恶化的环境,提出了一种适用于信号遮蔽环境的改进GPS+PDR组合定位算法。该方法用短时间内的陀螺仪积分数据校正数字罗盘的航向偏差,在一定程度上消除了数字罗盘受到的偶发干扰。采用约束残差的无迹卡尔曼滤波(UKF)算法对GPS和行人航迹推算(PDR)定位信息进行融合处理,有效克服了PDR定位中累积航向误差产生的位置漂移问题,提高了算法的定位精度和稳定性。试验结果表明,改进算法能有效抑制数字罗盘的漂移误差,航向相对误差平均降低56%;行人步行时,GPS定位标准误差为2.67 m,单纯PDR定位标准误差为6.83 m;随机给予若干点GPS数据辅助定位,标准误差降至3.12 m;全程融合GPS与PDR定位,标准误差可降至1.94 m。

关键词: GPS, 行人航迹推算, 组合导航, 多传感器数据融合, 精度, 无迹卡尔曼滤波

Abstract: In view of the deterioration of received signal quality, an improved GPS+PDR fusion positioning system applicable for signal shielding environment has been proposed. This method uses gyro integrating data within short time to correct heading deviation of digital compass, which has eliminated the incidental interference for digital compass to some extent significantly. UKF (unscented Kalman filter) algorithm used for constraining residual has been adopted to make fusion treatment for GPS and PDR positioning information, which has got over the position drift problem produced by cumulative heading error in PDR positioning effectively and improved the positioning accuracy and stability of algorithm. Experimental results show that the improved algorithm can restrain the drift error of digital compass effectively and the relative heading error has decreased by 56% on average. During walking, the standard error of GPS positioning is 2.67 m and the standard error of pure PDR positioning is 6.83 m. GPS data in several points have been offered randomly to assist positioning, the standard error decreased to 3.12 m. With GPS and PDR positioning integrated in the whole process, the standard error can be decreased to 1.94 m.

Key words: GPS, PDR, integrated navigation, multi-sensor data fusion, accuracy, UKF

中图分类号: