基于RSS空间线性相关的WLAN位置指纹定位算法

摘要

Due to RSS time-varying and difference of signal receiving ability of different terminals,the performance of RSS-based technologies is usually instability.In order to solve such problem,a novel fingerprint localization algorithm based on linear spatial dependence of RSS was proposed.Multiple sets of RSS samples were collected at each reference point to form a feature matrix and an offline location fingerprint database was conducted.When the real-time RSS matrix was used to calculate the correlation between the real-time RSS matrix and the reference point of the fingerprint library,the k-reference points were obtained,and the final position of the user was calculated by the quadratic weighted centroid algorithm.In order to effectively reduce the influence of signal time-varying,the sampling and sorting process were carried out,and the number of sampling times increased as much as possible when constructing the offline fingerprint database,but the samples needed to be aggregated to fit the positioning correlation calculation.Experiment results show that the proposed algorithm can guarantee the high positioning accuracy and also achieve the better stability for different terminal.%针对RSS(接收信号强度)时变性以及不同终端信号接收能力的差异性,导致WLAN位置指纹定位不稳定的问题,基于RSS空间线性相关性提出一种新颖的位置指纹定位算法.在每个参考点分别采集多组RSS样本形成特征矩阵,并构建离线位置指纹数据库.定位时,通过计算实时RSS矩阵与指纹库参考点相关性,得到最相关的k个参考点,利用二次加权质心算法计算用户的最终位置.为了有效降低信号时变性的影响,采样时进行了滤波、排序等处理,构建离线指纹数据库时尽量增加采样次数,但需要对样本进行聚合处理以适应定位相关性计算.实验结果表明,该算法在保证较高定位准确度的同时,针对不同终端有更好的定位稳定性.