多源信息融合的室内定位方法

Indoor positioning method based on multi-source information fusion

为实现在无法接收到卫星信号的复杂室内环境下行人精确定位与导航,提出一种多源信息融合的室内定位方法,通过融合智能手机中方向传感器数据进行室内行人航位推算(PDR),采集室内环境中的WiFi与蓝牙信号信息进行信号强度匹配及修正PDR累计误差;针对室内复杂环境下WiFi指纹定位精度低的情况,提出一种将随机森林(RF)分类与改进的K最近邻(KNN)算法相结合的WiFi指纹定位算法,降低计算复杂度,提高定位精度;在改进的WiFi指纹定位算法及多源信息融合分析的基础上,提出使用粒子滤波及地图约束的方法,进一步提高定位精度。实验结果表明:结合粒子滤波及地图约束的多源信息融合定位结果方法估计的路径与真实路径最为接近,证明了该方法的有效性。

In order to realize precise positioning and navigation of people in complex indoor environment where satellite signals cannot be received,an indoor positioning method based on multi-source information fusion is proposed.The indoor pedestrian dead reckoning(PDR)is collected by integrating the direction sensor data of smart phones.WiFi and Bluetooth signal information in indoor environment for signal strength matching and correction of PDR cumulative error.For the case of low precision of WiFi fingerprint positioning in complex indoor environments,a WiFi fingerprint localization algorithm combines random forest(RF)classification and improved K nearest neighbor(KNN)algorithm,reduces computational complexity and improves positioning precision.Based on the improved WiFi fingerprint localization algorithm and multi-source information fusion analysis,the method of particle filtering and map constraint is proposed to further improve the positioning precision.The experimental results show that the path estimated by the multi-source information fusion localization method combined with particle filtering and map constraint is the closest to the real path,which proves the effectiveness of the proposed method.