基于IMU设备的室内行人航位推算

Pedestrian Dead Reckoning Based on IMU Equipment

摘要:

在室内环境中,GPS 信号很难获取,以至于室内行人的航位推算(PDR)成了一个比较大的问题。现如今,常用的室内定位以及PDR的方法是结合外部设备如基于RFID、基于WiFi的定位系统。但是,这些设备在室内部署的时候需要消耗较大的物力以及人力资源。在这篇论文中,设计了一个较成熟的基于惯性测量单元(IMU)设备完成行人航位推算的计算方案。整个计算流程分为五个部分,载体的姿态解算、基于滑动窗口以及限定阈值的方法下的行人的步数检测、基于BP神经网络的行人的步长估计、结合磁力计与陀螺仪进行方向融合的航向推算。最后,根据上述所说的算法流程架构,实现了相应的安卓应用。经测试,在100米左右的回路环境下行走,总体距离误差不超过5米,航向角度偏差不超过5%,能够满足实际的需求

Abstract:

Pedestrian dead reckoning of indoor environment and indoor positioning is always a tough problem because it is hard to get the Global Positioning System’s signal when people stay in indoor environment. Nowadays, some practical methods about indoor positioning and PDR are mostly based on peripheral equipment, such as RFID, WIFI position systems. Unfortunately, Setting up these peripheral equipment needs a large number of human resources and money. This paper presents a new and mature calculating framework which is based on inertial measurement unit devices to calculate and estimate pedestrians’ movement. This framework include five parts, carrier attitude algorithm, based on sliding window and the method of limit threshold to estimate steps of pedestrian, the step length estimation of pedestrians based on BP neural network, the estimation of course based on the magnetometer and gyroscope fusion. Finally, according the above framework have developed an Android application. After several experiments, the total distance error is not more than 5 meters, and the deviation of the course Angle is no more than 5%, which can meet the actual demand.