基于蜂窝网格粒子滤波的行人导航航向估计方法

发布时间：2018-07-04 08:58

本文选题：行人导航 + 惯性导航　；参考：《中国惯性技术学报》2014年05期

【摘要】：为了提高低成本行人惯性导航系统的实时导航精度,针对行人惯性导航系统航向角发散问题,提出一种适用于实时行人导航系统的蜂窝网格粒子滤波算法(CSPF,cellular structure particle filter)。该方法将二维平面划分为紧密相连的蜂窝六边形区域,记录行人运动的历史轨迹信息,采用粒子滤波方法对航向误差补偿量进行估计。蜂窝网格粒子滤波算法在行人惯性导航系统的实验结果表明,引入了蜂窝网格粒子滤波的行人惯性导航方法与常规方法相比能有效改善实时行人惯性导航精度。在二维行走实验中,水平位置误差百分比由5.3%改进到1.7%;三维行走实验中,水平误差百分比由4.2%改进到2.9%。
[Abstract]:In order to improve the real-time navigation accuracy of low cost pedestrian inertial navigation system, aiming at the divergence of heading angle of pedestrian inertial navigation system, a cellular particle filter algorithm (CSPF / cellular structure particle filter).) is proposed for real-time pedestrian navigation system. In this method, the two-dimensional plane is divided into closely connected honeycomb hexagonal regions, and the historical track information of pedestrian movement is recorded, and the course error compensation is estimated by particle filter method. The experimental results of cellular mesh particle filter algorithm in pedestrian inertial navigation system show that the proposed method can improve the accuracy of real-time pedestrian inertial navigation effectively compared with the conventional method. In the two-dimensional walking experiment, the percentage of horizontal position error was improved from 5.3% to 1.7, and the percentage of horizontal error was improved from 4.2% to 2.9% in the three-dimensional walking experiment.
【作者单位】：南京航空航天大学自动化学院;中航工业陕西宝成航空仪表有限责任公司;
【基金】：国家自然科学基金(61374115,61328301,61104188,61273057,61174197) 江苏高校优势学科建设工程项目
【分类号】：TN96

【参考文献】