基于多传感器数据融合的室内定位算法

Multi-sensors data and information fusion algorithm for indoor localization

基于惯性测量单元(IMU)的定位方法是一种全自主定位方法,该方法通常是基于单个IMU(Single-IMU)实施定位,其具有较大的漂移误差和累积误差.因此,提出了一种基于多个可穿戴式IMU(Multi-IMUs)与室内无线传感器网络(IWSN)的多传感器数据融合的室内定位算法,根据佩戴于不同部位的Multi-IMUs信息协同,提高人体姿态检测的有效性,并且利用模糊投票机制(Fuzzy Voting Scheme)融合Multi-IMUs位置信息;此外,结合IWSN,采用卡尔曼滤波算法(Kalman Filter Algorithm)融合IWSN解算出的位置信息与Multi-IMUs计算出的位置信息降低基于IMU的累积误差.实验结果表明,所提出的基于多传感器数据融合的室内定位算法能够识别出行走的姿态,与基于Single-IMU的定位算法相比,该算法有效地降低了累积误差和漂移误差,提高了室内定位的有效性和可靠性.

The localization algorithm of based on IMU is one of autonomous localization methods while it possesses the disadvantage of drift error and accumulated error,so this paper proposes a multi-sensors including wearable multi-IMUs and IWSN data and information fusion algorithm for indoor localization. On the one hand,almost all indoor localization algorithms based on IMU use only one IMU while this single-IMU-based algorithm can＇ t judge the posture of person precisely,one appropriate solution is that we can utilize multi-IMUs to cooperate in localization process,besides,we can fuse position information of multi-IMUs by fuzzy voting scheme. On the other hand,in order to overcome the disadvantage of drift error and accumulate error,combining with IWSN in indoor and fusing the position information calculated by IWSN and multi-IMUs via Kalman Filter algorithm. Experiment results show that the proposed indoor localization algorithm possesses good property in judging posture of person and decreasing drift error and accumulate error comparing with traditional IMU-based indoor localization algorithm.