基于众包校正的多源融合室内定位算法

Multi-sensor fusion indoor positioning algorithm based on crowdsourcing correction

为了提高室内定位算法的精度,本文融合Wi-Fi和蓝牙两种信号源,提出了一种结合粗定位和众包校正的适用于多用户环境的室内定位算法。该定位算法分为离线和在线两个阶段,离线阶段的主要任务是构建多个指纹库,在线阶段的主要任务是进行粗定位和众包校正。粗定位分别利用Wi-Fi接收信号强度指示和蓝牙接收信号强度指示计算用户的粗略位置和用户间的距离;众包校正包括聚类校正和虚拟空间校正两个部分,它利用用户间的距离和用户组的位置分布提高定位精度。在UJIIndoorLoc和IPIN2017-CAR数据集上进行验证,实验结果表明,提出的定位算法将平均定位误差分别降至4.96 m和4.35 m。

To improve the accuracy of the indoor positioning algorithm, an indoor positioning algorithm for multi-user environments by fusing Wi-Fi and Bluetooth data is proposed.The positioning algorithm combines coarse positioning and crowdsourcing correction and is divided into offline and online phases. The main task of the offline phase is to build multiple fingerprint databases, while the online phase is to perform coarse positioning and crowdsourcing correction. Specifically, coarse positioning uses Wi-Fi received signal strength indicator and bluetooth received signal strength indicator to calculate the cursory position of users and the distance between users. Crowdsourcing correction includes clustering correction and virtual space correction, which uses the distance between users and the position distribution of user groups to improve positioning accuracy. Lots of experiments are conducted on UJIIndoorLoc and IPIN2017-CAR datasets.The experimental results show that the average positioning error of our proposed positioning algorithm can be reduced to 4.96 m and 4.35 m respectively.