一种12维零速状态更新的智能行人航位推算  

12-dimensional Zero Velocity State Updating Intelligent Algorithm for Pedestrian Dead Reckoning

作　　者：刘恒志 李擎[1] Liu Hengzhi;Li Qing(Beijing Information Science Technology University Beijing Key Laboratory of High Dynamic Navigation Technology,Beijing 100192,China)

机构地区：[1]北京信息科技大学高动态导航技术北京市重点实验室,北京100192

出　　处：《系统仿真学报》2018年第11期4387-4394,共8页Journal of System Simulation

基　　金：国家自然科学基金(61471046);北京市教委市属高校创新能力提升计划(TJSHG201510772017);高动态导航技术北京市重点实验室开放课题

摘　　要：针对目前不同的可穿戴式行人导航的捷联惯导装置需要不同的PDR(PedestrianDead Reckoning)参数,并且非最优值参数影响定位精度的问题,提出一种12维零速状态更新智能算法:通过引入三维速度、姿态、位置、磁力计误差作为系统观测量,并将支持向量机与卡尔曼滤波器构成智能估计器估计系统状态误差,从而提升系统精度。利用自研的IMU(Inertial Measurement Unit)传感器进行实验验证,结果证明该方法有效观测系统状态,最优估计系统误差,提升了系统的精度:水平和空间误差平均缩减40%和45%,对行人导航技术的研究具有实际应用价值。For wearable pedestrian strapdown inertial navigation and location devices, the different devices need different pedestrian dead reckoning （PDR） parameters, and the parameters are not such the optimal value that it affects the accuracy. A self-pedestrian navigation and location method based on the 12-dimensional zero-velocity state update intelligent algorithm is proposed, in which three dimensional errors of velocity, angular speed, location and geomagnetism are introduced as the system observations and an intelligent estimator which is formed by the support vector machine （SVM） and Kalman filter is established to estimate the system state error, and therefore the system accuracy is improved. By the experimental verification with using the self-developed IMU sensor, the results prove that this method observes system status effectively and estimates system errors intelligently. Comparing with the traditional ZUPT, the proposed method can reduce the horizontal error by an average of 40% and the spatial error by an average of 45%.

关 键 词：捷联惯导 可穿戴式 行人自主导航 IMU PDR 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]