一种室内位置指纹数据库构建方法

A method of constructing indoor fingerprinting database for pedestrian navigation

针对室内定位导航服务很大程度上依赖于数据库的支撑,高精度参考点位置指纹数据库的构建与更新需要耗费大量的人力劳动的问题,该文提出一种快速构建室内位置指纹数据库的方法,该方法通过智能手机采集惯性传感器数据和视频数据,使用改进的图像匹配方法连续匹配视频图像,结合运动恢复结构方法和惯性航位推算方法进行行人运动轨迹的几何重建与位置估计,进而实现室内位置指纹数据库的快速构建。实验结果表明,该方法得到的位置标注结果平均误差为0.53 m,构建得到的位置指纹数据库可直接用于室内定位。

Aiming at the problem that indoor positioning and navigation service relies heavily on the support of the fingerprinting database or radio map,and the construction or update the indoor fingerprinting database is labor cost and time consuming,this paper proposed a method which can fast construct the indoor fingerprinting database by using smartphone. By integrating the inertial sensors and video images collected from the smartphone,this method can recover the walking trajectory of participant. The sampling points of trajectory can be used for constructing fingerprints and radio map. This method collects the accelerator data,gyroscope data and video data from smartphone. An improved image matching method was proposed for calculating the heading of trajectory. The experiments result demonstrated that average error of recovering trajectory was 0.53 m,the constructed fingerprinting database could directly use in indoor localization.