A Coarse Fingerprint-Assisted Multiple Target Indoor Device-Free Localization With Visible Light Sensing
A Coarse Fingerprint-Assisted Multiple Target Indoor Device-Free Localization With Visible Light Sensing
复制标题
具有可见光传感功能的粗指纹辅助多目标室内无设备定位
作者：
Shuai Zhang;Kaihua Liu;Yunlei Zhang;Ya Wang
可见光传感技术在建筑物内人体定位方面显示出巨大的潜力。对于指纹室内定位来说，如何在保持准确位置估计的同时降低线下人工成本至关重要。本文提出了一种新颖的粗指纹辅助无设备定位（DFL）方法，用于通过可见光传感定位多个目标，不需要太多的校准工作来建立高精度指纹数据库。此外，针对多目标DFL中的伪目标定位问题，设计了一种基于粗单目标指纹数据库的逐次消除算法。由于阴影链接的交点不能保证对称地位于人体质心周围，因此估计的位置将偏离目标的真实位置。为了解决这个问题，引入了基于凸包的定位算法来提高定位精度。大量的仿真结果表明，我们提出的方法在目标计数和定位精度方面表现出优越的性能。
Visible light sensing technique has shown great potential in locating human in the building. For indoor positioning using fingerprints, how to reduce the offline labor cost while maintaining accurate location estimation is of vital importance. In this paper, a novel coarse fingerprint-assisted device-free localization (DFL) approach is proposed for locating multiple targets via visible light sensing, which does not need too much calibration effort to build high accurate fingerprint database. Furthermore, a successive cancellation algorithm based on the coarse single target fingerprint database is designed to address the problem of pseudo target positioning in the multi-target DFL. Since the intersection points of shadowed links could not be guaranteed to be symmetrically located around the center of mass of the human body, the estimated location would be biased from the true location of the target. To solve this problem, a convex hull-based localization algorithm is introduced to improve positioning accuracy. Extensive simulation results show that our proposed approach exhibits superior performance in terms of target counting and localization accuracy.