基于手机的室内数字平面图构建方法

Construction Method of Indoor Digital Floor Plan Based on Smartphone

针对目前室内地图绘制和更新困难的问题,提出一套室内平面图自动构建方法。采集智能手机内置传感器数据,包括加速度、角速度、磁场强度和海拔,进行行人航位推算,获得行走轨迹。通过对大量轨迹的分析,结合WiFi信号变化,识别出轨迹中的转弯、门位置和楼梯位置,并将轨迹划分为房间类型和走廊类型。采用聚类算法和α-shape算法对房间类型轨迹进行分析,绘制出房间形状。根据走廊类型轨迹,利用聚类算法和主成分分析算法确定走廊的长和宽。实验结果表明,该算法能够较准确地自动构建出包括房间和走廊的室内数字平面图。

Aiming at the difficulty of drawing and updating indoor floor plan,this paper proposes an automatic construction method of indoor floor plan based on smartphones.The method first collects smartphone sensor data,including acceleration,angular velocity,magnetic field strength and altitude,to estimate walking traces by pedestrian dead reckoning.Afterwards,the method analyzes the walking traces and recognizes turnings,door locations and stairs,together with WiFi signals.Clustering and theα-shape algorithm are applied to construct the rooms.Clustering and the principal component analysis algorithm are applied to construct the corridors.Evaluations show that this algorithm is able to construct accurate indoor floor plans with rooms and corridors.