运动分类步频调节的微机电惯性测量单元室内行人航迹推算  

Micro-Electro-Mechanical System-Inertial Measurement Unit Indoor Pedestrian Dead Reckoning Based on Motion Classification and Step Frequency Adjustment

作　　者：李若涵[1] 张金艺[1,2,3] 徐德政[2] 陈兴秀[1] 徐秦乐 

机构地区：[1]上海大学特种光纤与光接入网省部共建重点实验室,上海200072 [2]上海大学微电子研究与开发中心,上海200072 [3]上海大学新型显示与系统应用重点实验室,上海200072

出　　处：《上海大学学报（自然科学版）》2014年第5期612-623,共12页Journal of Shanghai University:Natural Science Edition

基　　金：上海市教委重点学科建设资助项目(J50104);上海市科委基金资助项目(08706201000;08700741000)

摘　　要：行人航迹推算(pedestrian dead reckoning,PDR)作为一种新兴的导航定位方法,因其不易受外界环境因素影响而受到广泛关注.针对室内行人航迹推算,采集并分析了微机电惯性测量单元(micro-electro-mechanical system-inertial measurement unit,MEMS-IMU)数据,设计了运动分类的区间对称步频检测,并建立了步频调节的步长估计模型,最后提出了运动分类步频调节的MEMS-IMU室内行人航迹推算,从而实现较精准的定位.针对不同个体,对步频调节的步长估计模型进行个性化标定,以进一步提高室内行人航迹推算性能.验证结果表明:与传统峰值非线性方法相比,运动分类步频调节的MEMS-IMU室内行人航迹推算的定位误差降低了32.6%,使短距离室内行人航迹推算在无其他定位技术支持的情况下具有较高精度.As a new navigation method, pedestrian dead reckoning （PDR） has attracted much attention because it is less susceptible to environmental factors. To solve the indoor PDR problem, data of a micro-electro-mechanical system-inertial measurement unit （MEMS-IMU） are collected and analyzed. A step detection algorithm is developed for motion classification and interval symmetry, and step length estimation model is established for step frequency adjustment. Thus a MEMS-IMU indoor PDR based on the motion classification and step frequency adjustment is constructed to realize accurate positioning. For different individuals, personalized step estimation model parameters are used to improve the positioning performance. Experimental results show that, the indoor PDR based on motion classification and step frequency adjustment reduces positioning error by 32.6% as compared to a traditional method using peak detection and a nonlinear model, achieving high positioning accuracy without resorting to any other positioning techniques.

关 键 词：室内行人航迹推算 微机电惯性测量单元 步频 步长 运动分类 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]